Debian Codegraph @

ChangeLog

raw
1 *******************************************************************************
2 Version 1.14.31
3 *******************************************************************************
4
5 - Fix for CVE-2026-41682
6
7 *******************************************************************************
8 Version 1.14.30
9 *******************************************************************************
10
11 - Revert 25d4bd0b which introduced an API change.
12
13 *******************************************************************************
14 Version 1.14.29
15 *******************************************************************************
16
17 - CMake builds from the tarball now work.
18
19 *******************************************************************************
20 Version 1.14.28
21 *******************************************************************************
22
23 - Hopefully we have SONAMEs right for CMake builds now.
24
25
26 *******************************************************************************
27 Version 1.14.27
28 *******************************************************************************
29
30 - Fix OmniOS build when using POSIX compliant asctime_r by @mill1000 in #522
31 - remove asctime_r and gmtime_r by @neheb in #523
32 - cmake: link to libnsl and libsocket for SunOS by @neheb in #524
33
34 *******************************************************************************
35 Version 1.14.26
36 *******************************************************************************
37
38 - bugfix-skip-init-on-VPN-adaptors by @ss1969 in #500
39 - readFromSSDPSocket: Don't return error when sockets receive zero length packet
40 by @mill1000 in #501
41 - Fix warning on Mac compilation by @mrjimenez in #503
42 - shutdown sockets in finish routine by @ss1969 in #504
43 - Fix segfault on ixml, thanks to Harrison Green (hgarrereyn) #506
44 - Crash in ixml element set attribute node ns by @mrjimenez in #507
45 - Add another test to ixml by @mrjimenez in #509
46 - fix and add omnios CI by @neheb in #511
47 - cmake: add missing file in Windows compilation by @robUx4 in #518
48
49 *******************************************************************************
50 Version 1.14.25
51 *******************************************************************************
52
53 - Handle multiple results of getaddrinfo by @KarlStraussberger in #496
54 - Tweak ssdp_device by @KarlStraussberger in #497
55
56 *******************************************************************************
57 Version 1.14.24
58 *******************************************************************************
59
60 2025-06-11 Marcelo Roberto Jimenez <mrjimenez(at)github.com>
61
62 Fix possible close() on non-initialized socket.
63
64 2025-06-11 Benjamin Gilbert <bgilbert(at)backtick.net>
65
66 ssdp_device.c: Fix missing casts in setsockopt()
67
68 Windows thinks argument 4 of setsockopt() is a const char *. Add casts
69 to avoid build errors with GCC.
70
71 Fixes: 52cb6b52d2 ("ssdp_device.c: Bind SSDP UDP response to 1900")
72
73 *******************************************************************************
74 Version 1.14.23
75 *******************************************************************************
76
77 2025-06-07 Ryan C. Underwood <runderwo(at)github.com>
78
79 Improve SSDP initialization and error handling (#486)
80
81 It was possible for some uninitialized garbage to cause getaddrinfo() to
82 fail on glibc 2.41, which would crash the client program. Instead,
83 zero the whole struct, and gracefully handle getaddrinfo() failures.
84
85 *******************************************************************************
86 Version 1.14.22
87 *******************************************************************************
88 2025-06-02 KarlStraussberger <KarlStraussberger(at)github.com>
89
90 Fix broken compilation on Alpine 3.21 / gcc-14.2
91
92 *******************************************************************************
93 Version 1.14.21
94 *******************************************************************************
95
96 2025-01-23 LutzenH <LutzenH(at)github.com>
97
98 Fix subs not working on Windows due to missing netmask
99
100 2024-11-21 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
101
102 - Lots of IXML fuzzer fixes
103 - ssdp_device.c: Bind SSDP UDP response to 1900
104
105 *******************************************************************************
106 Version 1.14.20
107 *******************************************************************************
108
109 2024-10-03 Vollstrecker <werner(at)vollstreckernet.de>
110
111 - ixml has its own version
112
113 2024-07-09 Karlchen <k_straussberger(at)netzland.net>
114
115 - Add CORS header to avoid browser warnings
116 - Add new API call
117 - Add internal function
118 - Implement setting allow-origin in header
119 - Implement new API call
120 - Headers do not have to be quoted
121 - Allow setting MaxJobsTotal for ThreadPools
122
123 2024-06-29 Rosen Penev <rosenp(at)gmail.com>
124
125 - fix formats under newer newer MSVC
126 - fix 32-bit format warnings
127
128 2024-04-25 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
129
130 - Fixes broken cmake regex
131 - Fixes cmake redefinition warnings
132
133
134 *******************************************************************************
135 Version 1.14.19
136 *******************************************************************************
137
138 2024-04-22 Steve Lhomme <robux4(at)ycbcr.xyz>
139
140 - disable fseeko usage on 32-bit Android older than API 24
141
142 feesko is not supported on 32-bit Android older than API 24 [1].
143 Older SDKs would allow the call but the 64-bit off_t was not accurate.
144 With the NDK26 the call is hidden in that case and calling fseeko() results in a compilation error.
145
146 [1] https://android.googlesource.com/platform/bionic/+/main/docs/32-bit-abi.md
147
148 2023-09-16 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
149
150 - Fixed memory allocation in `tv_device.c::TvDeviceSetContrast()`.
151
152 2023-08-31 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
153
154 - Updated some old AX macro files to fix bootstrap messages.
155
156 *******************************************************************************
157 Version 1.14.18
158 *******************************************************************************
159
160 2023-06-29 Thomas Guillem <thomas(at)gllm.fr>
161
162 miniserver: fix busy loop on socket error
163
164 In case of a socket error, the socket was not removed from the select
165 pool, causing future select() calls to return immediately, and
166 readFromSSDPSocket() to be called just after on the failing socket. This
167 was causing a high CPU load.
168
169 *******************************************************************************
170 Version 1.14.17
171 *******************************************************************************
172
173 2023-04-30 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
174
175 Include fuzzer and gtest in the tarball.
176
177 *******************************************************************************
178 Version 1.14.16
179 *******************************************************************************
180
181 2023-03-19 philippe44 <[email protected]>
182
183 Fix memory leak and ActionExAsync (#426)
184
185 This intends to fix #424. I've found at least 2 memory leaks
186
187 1- When sending Action(Ex)Async
188 The ActionRequest document was not released nor was the ActionResult.
189 I initially thought that this should be done in
190 UpnpActionComplete::UpnpActionComplete_delete() but I was wrong because
191 ActionRequest and ActionResult do not "belong" to the Evt object.
192 The sets methods only set a pointer and modifying generator.c is not the
193 right solution. The reason is that ActionRequest "belongs" to the ACTION
194 job and should be handled as such if we want as well that, when the job
195 is cancelled, the memory is released as well. So I've change the
196 free_func for ACTION to reflect that.
197
198 2- When using event subscription
199 The Event was not released upon Timer termination. So I've done as
200 above, created a special free_func that handles that properly.
201
202 In general, when something prevents the execution/submission of a job,
203 I've also used the actual free function instead of duplicating the
204 individual free(). Note that when terminating a timer, I'm using the
205 job's internal structure free_func() as it was already used to access
206 arg. One can argue that it should be an opaque structure and get/set or
207 the option to call the free function could be use instead.
208
209 Also, I realized that ActionExAsync has never been working properly
210 because when the job is executed, only an Action is made, never an
211 ActionEx although the Header document is in the job's argument. I've
212 fixed that as well and stopped there to not go too deep in that rabbit
213 hole.
214
215 I've also removed the upnp_timeout.c/h files as they were not used
216 anymore (and had a strange names anyway)
217
218 2023-01-04 Thoren Paulson <[email protected]>
219
220 Handle redundant initialization
221
222 2022-12-22 Thoren Paulson <[email protected]>
223
224 Clean up after failed initialization.
225
226 Some initialization failures (such as failure in UpnpGetIfInfo) would
227 leave resources like thread pools initialized. This change ensures if
228 UpnpInit2 fails, we call UpnpFinish before returning to clean up those
229 resources.
230
231 To ensure UpnpFinish is effective in this way, we set UpnpSdkInit to 1
232 earlier in UpnpInit2.
233
234 *******************************************************************************
235 Version 1.14.15
236 *******************************************************************************
237
238 2022-11-18 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
239
240 Fix for missing CMake files in the autotools distro.
241
242 *******************************************************************************
243 Version 1.14.14
244 *******************************************************************************
245
246 2022-09-27 philippe44 <[email protected]>
247
248 don't use strxxx_s on Windows
249
250 The use of these fails as 2nd argument is the same as 4th argument.
251 Reading the cpp_reference definitions of these, it can't be right.
252
253 2022-09-20 Rex Cheung <93339127+rexcheung21(at)users.noreply.github.com>
254
255 Fix #412
256
257 Skip leading UTF-8 BOM in XML document
258
259 2022-08-04 Hugo Beauzée-Luyssen <hugo(at)beauzee.fr>
260
261 ThreadPool: Fix non UCRT builds
262
263 *******************************************************************************
264 Version 1.14.13
265 *******************************************************************************
266
267 2021-04-28 Hugo Beauzée-Luyssen <hugo(at)beauzee.fr>
268
269 http: RecvMessage: Fix potential allocation failure
270
271 If for whatever reason the server replies without maxing out the buffer,
272 the code would end up doubling the buffer size nonetheless, which can
273 quickly cause malloc to fail, once we try to allocate 2^30 or more bytes
274 This was especially noticeable for large results.
275 Reported as https://code.videolan.org/videolan/vlc/-/issues/27173
276
277 *******************************************************************************
278 Version 1.14.12
279 *******************************************************************************
280
281 2021-04-28 Stephen Anthony Jones <frobones(at)gmail.com>
282
283 Fixes double free caused by closing a file pointer twice.
284
285 2021-09-25 Vollstrecker <[email protected]>
286
287 More soname fixes.
288
289 *******************************************************************************
290 Version 1.14.11
291 *******************************************************************************
292
293 2021-08-20 Vollstrecker <[email protected]>
294
295 More soname fixes.
296
297 *******************************************************************************
298 Version 1.14.10
299 *******************************************************************************
300
301 2021-08-17 Vollstrecker <[email protected]>
302
303 Fixes cmake build soname.
304
305 *******************************************************************************
306 Version 1.14.9
307 *******************************************************************************
308
309 2021-08-08 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
310
311 Fixes library soname.
312
313 *******************************************************************************
314 Version 1.14.8
315 *******************************************************************************
316
317 2021-08-02 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
318
319 Removes references to deprecated pthread_mutexattr_{g,s}etkind_np
320
321 *******************************************************************************
322 Version 1.14.7
323 *******************************************************************************
324
325 2021-05-08 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
326
327 Fixes broken windows build.
328
329 *******************************************************************************
330 Version 1.14.6
331 *******************************************************************************
332
333 2021-04-19 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
334
335 Fix for a DNS Rebind exploit. A special thanks for the collaboration
336 of the following people:
337 - Alaric Senat
338 - Fabrice Fontaine
339 - Gabriel Corona
340 - Ian Whyman
341 - Jean-Francois Dockes
342 - Marvin Scholz
343 - Werner Mahr
344
345 2021-04-06 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
346
347 Fix for Github #250:
348 When upnp uses ixml to parse SOAP messages which contains too many node,
349 services are unavailable.
350
351 *******************************************************************************
352 Version 1.14.5
353 *******************************************************************************
354
355 2021-04-06 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
356
357 Avoid potential memory leak in http_SendMessage()
358
359 2021-04-06 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
360
361 Get rid of alloca()'s.
362
363 2021-04-05 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
364
365 Non-recursive version of ixmlNode_free() avoids stack overflow attack.
366 Fixes CVE-2021-28302.
367
368 *******************************************************************************
369 Version 1.14.4
370 *******************************************************************************
371
372 2021-03-29 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
373
374 Add more missing CMake infrastructure to the tarball.
375
376 *******************************************************************************
377 Version 1.14.3
378 *******************************************************************************
379
380 2021-03-29 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
381
382 Add CMake infrastructure to the tarball.
383
384 2021-03-13 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
385
386 Fix for gena leak.
387
388 *******************************************************************************
389 Version 1.14.2
390 *******************************************************************************
391
392 2021-02-17 Jim Broadus <jbroadus(at)gmail.com>
393
394 upnpapi.c: Fix crash in UpnpGetIfInfo
395
396 Per getifaddrs documentation, the ifa_addr field of an ifaddrs structure
397 can be null. In a real world example, an entry may be provided for the
398 non-existent hardware address of a tunnel device. This behavior was
399 observed with the netlink based getifaddrs implementation in glibc.
400
401 2021-02-17 Karlchen <k_straussberger(at)netzland.net>
402
403 Handle interfaces without address
404
405 I got a crash on gerbera startup because at least one interface did not have `ifa_addr` set.
406
407 The reason is quite simple: they are the physical interfaces which are part of my channel
408 ```
409 2: eth1: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc fq_codel master bond0 state UP group default qlen 1000
410 link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff
411 3: eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc fq_codel master bond0 state UP group default qlen 1000
412 link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff
413 4: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
414 link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff
415 ```
416
417
418 2021-02-14 Vollstrecker <github(at)vollstreckernet.de>
419
420 CMake:
421 - Fix tests for all occasions. (Hopefully)
422 - Set right so-version and don't yell commands. Fixes #231
423 - Reverted libname-change till decided
424
425 2021-02-09 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
426
427 Fix IPv6 GENA subscription
428
429 Use LOCAL_PORT_V6_ULA_GUA or LOCAL_PORT_V6 depending on the IPv6 address
430
431 2021-02-08 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
432
433 samples: allow the user to specify the control point interface
434
435 This is useful to test UPnP on specific interface (e.g. IPv6-only).
436 This was already possible on the device sample.
437
438 2021-02-08 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
439
440 upnp/src/api/upnpapi.c: don't set gIF_IPV4 if there is no IPv4
441
442 Don't set gIF_IPV4 if no IPv4 is found to keep the default value of '\0'
443 otherwise SSDP will try to register IPv4 multicast with address 0.0.0.0
444 which will result in a runtime failure
445
446 *******************************************************************************
447 Version 1.14.1
448 *******************************************************************************
449
450 2021-02-05 Ian Whyman <ian(at)gerbera.io>
451
452 CMake Github Actions, including windows build.
453
454 2021-02-05 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
455
456 upnpapi.c: assume that getifaddrs is available
457
458 Assume that getifaddrs is available even if it is not POSIX-compliant,
459 this will simplify the code and, as a side effect, this allow pupnp to
460 work on an interface that does not have an IPv4 address.
461
462 2020-12-09 Vollstrecker <github(at)vollstreckernet.de>
463
464 CMake support.
465
466 2020-11-30 Marcelo Jimenez <marcelo.jimenez(at)gmail.com>
467
468 GitHub #203 by Vollstrecker, part 2.
469
470 Fixes the inclusion of alloca.h in WIN32.
471
472 2020-11-30 Marcelo Jimenez <marcelo.jimenez(at)gmail.com>
473
474 GitHub #203 by Vollstrecker, part 1.
475
476 Remove the now unused file ClientSubscription.c.
477
478 2020-11-09 Thomas Mitterfellner <thomas.mitterfellner(at)gmail.com>
479
480 Remove port >= 49152 restriction.
481
482 *******************************************************************************
483 Version 1.14.0
484 *******************************************************************************
485
486 2020-07-19 Marcelo Jimenez <marcelo.jimenez(at)gmail.com>
487
488 Reworked the miniserver code to deal with SO_REUSEADDR
489
490 Factored the common socket code and test for EADDRINUSE returning from
491 listen() when SO_REUSEADDR is turned on.
492
493 2020-06-17 Marcelo Jimenez <marcelo.jimenez(at)gmail.com>
494
495 Removed template classes.
496
497 The files generated through template classes had some drawbacks:
498
499 1 - You could not read the code.
500 2 - You could not step through the code with a debugger.
501 3 - Doxygen was unable to document it.
502
503 This patch removes the templates and creates an auto generator for these
504 boilerplate generated code in the same spirit of the templates, but
505 solves the above problems.
506
507 Still, dealing with documentation still needs some work, but should be
508 much easier now.
509
510 2020-06-11 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
511
512 upnpapi.c: retrieve gIF_IPV4_NETMASK with BSD
513 GitHub #183
514
515 BSD is using getifaddrs, update the code to retrieve the IPv4 netmask
516 (used in the CallStranger fix)
517
518 Signed-off-by: Fabrice Fontaine [email protected]
519
520 2020-06-09 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
521
522 Drop UpnpInit
523 GitHub #182
524
525 This function is deprecated since a long time moreover it is vulnerable
526 to CallStranger a.k.a. CVE-2020-12695 and can't be fixed without
527 breaking the API as HostIP only allows the user to pass an IP address
528 and not a netmask.
529
530 If we want to discover the netmask from the HostIP provided in UPnPInit,
531 we'll have to loop through all the available interfaces to find the
532 interface with the given IP address to finally retrieve the netmask.
533 This is a lot of work/modification for a deprecated function. Moreover,
534 in the end UPnPInit will be like a "deprecated" (i.e. IPv4-only) version
535 of UPnPInit2.
536
537 So it is time to remove this deprecated function.
538
539 As a result, also remove getlocalhostname and DEFAULT_INTERFACE which
540 are not needed anymore and replace UpnpInit by UpnpInit2 in comments.
541
542 Signed-off-by: Fabrice Fontaine [email protected]
543
544 2020-06-09 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
545
546 Partial fix for CallStranger on IPv4
547 GitHub #181
548
549 This is a partial fix for CallStranger a.k.a. CVE-2020-12695
550
551 Check that DeliveryURLs are in the expected network segment as requested
552 by the new UPnP UDA:
553 https://openconnectivity.org/upnp-specs/UPnP-arch-DeviceArchitecture-v2.0-20200417.pdf.
554
555 Here is an extract of the new requirement:
556
557 The subscription request containing a delivery URL not on the same
558 network segment as the fully qualified event subscription URL shall not
559 be accepted. For private networks this means that the delivery URL
560 provided will adhere to the following IP ranges:
561
562 . 10.0.0.0 - 10.255.255.255 (10/8 prefix)
563 . 172.16.0.0 - 172.31.255.255 (172.16/12 prefix)
564 . 192.168.0.0 - 192.168.255.255 (192.168/16 prefix)
565
566 In the context of pupnp, this means that the IPv4 netmask is now
567 retrieved when using UPnPInit2. Then, each DeliveryURL is checked
568 against the device's IPv4 address and netmask. If one of them are not
569 compliant, the whole subscription is rejected.
570
571 This first commit should be enhanced / updated to:
572
573 . remove UPnPInit (it is deprecated for a long time) or update it so
574 . the user can also pass the netmask
575 . fix IPv6
576 . fix Windows code
577 . retrieve the netmask in the BSD code of UPnPInit2
578
579 Signed-off-by: Fabrice Fontaine [email protected]
580
581 2020-06-04 Patrik Lantz pjlantz(at)github
582
583 Fixes #177
584
585 NULL pointer dereference in FindServiceControlURLPath
586
587 *******************************************************************************
588 Version 1.12.1
589 *******************************************************************************
590
591 2020-04-03 Rene Nielsen <rene.nielsen(at)microchip.com>
592
593 setsockopt(IP_MULTICAST_IF) fails on 64-bit CPUs
594
595 2020-04-03 Rosen Penev <rosenp(at)gmail.com>
596
597 GitHub PR #164
598
599 Do not use usleep when using newer POSIX C source.
600
601 usleep is deprecated and is optionally unavailable with uClibc-ng.
602
603 2020-03-29 pamapa <pamapa(at)github>
604
605 GitHub PR #162
606
607 Add Os/UserAgent information to UpnpActionRequest and UpnpFileInfo
608
609 2020-03-28 pamapa <pamapa(at)github>
610
611 GitHub PR #161
612
613 Add CtrlPtIPAddr to UpnpFileInfo class
614
615 2020-01-27 Ian Whyman <v00d00(at)github>
616
617 GitHub PR #160
618
619 List: Add extern C for C++ users
620
621 *******************************************************************************
622 Version 1.12.0
623 *******************************************************************************
624
625 2020-01-22 Stephen Bash <scbash(at)github>
626
627 Reduce spurious HTTP 416 errors due to ill-defined bytes header
628
629 I (re)discovered this behavior trying to use the Gerbera media server
630 with Chromecast (built in to my Vizio P55-F1 TV). Chromecast specifies
631 "bytes:0-" with no end range, which caused pupnp to return
632 RANGE_NOT_SATISFIABLE. Jin, the author of MediaTomb, of which Gerbera
633 is a continuation, fixed this in 2007 in MediaTomb's fork of pupnp, see
634 gerbera@ccd7994d45 "made sure that range requests specified as
635 "bytes=0-" do not trigger...", but never passed the fix upstream. When
636 restarted as Gerbera, pupnp was removed from the local tree in favor of
637 using the upstream version, and the patch was lost which lead to me
638 rediscovering it recently.
639
640 This is Jin's original patch applied to pupnp 1.8. This teaches pupnp to
641 ignore the bytes header if it matches "0-" and return HTTP_OK.
642
643 2020-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
644
645 GitHub issue #157
646
647 When building v1.8.6 (on GNU/Linux) with --disable-device, pupnp doesn't build:
648 /bin/bash ../libtool --tag=CC --mode=compile gcc -DHAVE_CONFIG_H -I. -I.. -I../upnp/inc -I./inc -I../ixml/inc -DNDEBUG -I./src/inc -I./src/threadutil -I/foo/contrib/x86_64-linux-gnu/include -fPIC -pthread -I/foo/contrib/x86_64-linux-gnu/include -DNDEBUG=1 -O3 -DUPNP_STATIC_LIB -Os -Wall -c -o src/genlib/util/libupnp_la-util.lo `test -f 'src/genlib/util/util.c' || echo './'`src/genlib/util/util.c
649 In file included from src/genlib/miniserver/miniserver.c:58:0:
650 ./src/inc/upnpapi.h:208:2: error: unknown type name 'service_info'
651 service_info **serv_info
652 ^~~~~~~~~~~~
653
654 This patch fixes compilation for the --disable-device case.
655
656
657 2019-11-20 ePirat Marvin Scholz <epirat07(at)gmail.com>
658
659 GitHub PRs:
660 #150: Remove unused files
661 #149: UpnpEnableWebserver: Error correctly when web server compiled-out
662 #148: CI: Remove .travis.yml
663 #147: Remove TRUE/FALSE defines and BOOL typedef
664 #146: CI: Add address sanitizer and leak sanitizer enabled test run
665 #145: Add list test
666 #144: Add -Wextra and -Wpedantic to get more warnings
667 #143: More warning fixes
668 #142: Doxygen changes
669 #141: Remove unused headers
670 #140: Remove IN, OUT and INOUT defines
671 #139: Remove unused defines in uri.h
672 #138: Use stdbool.h instead of BOOL typedef and defines
673 #137: Various compiler warning fixes
674 #136: configure.ac: Fix Windows detection
675 #135: configure.ac: Enable silent rules
676 #134: LinkedList: Simplify _WIN32 check
677 #133: Remove checks of __OSX__ define
678 #132: Add Github Actions based CI
679
680 2019-12-03 Jean-Francois Dockes <jf(at)dockes.org>
681
682 Github PR #130: Add forgotten Windows dll export qualifiers
683
684 2019-12-01 Jean-Francois Dockes <jf(at)dockes.org>
685
686 Github PR #129: Remove and replace the list.h file
687
688 2019-11-20 ePirat Marvin Scholz <epirat07(at)gmail.com>
689
690 Github PRs:
691
692 #128: Overhaul list.h to fix various issues
693 #127: Use rand_s in get_random_info on Windows
694 #126: Do not redeclare timezone struct if already defined in Mingw
695 #125: Fix interface filtering for Windows
696 - The mentioned fix is not only needed for MSVC, but when
697 cross-compiling for Windows with mingw-w64 too.
698 #124: config.h: Remove DEBUG_TARGET
699 - The DEBUG_TARGET is no longer used, it was previously used in
700 upnpdebug.c to disable logging to a file, which now is done
701 by just not calling UpnpSetLogFileNames.
702 #123: Check for -lpthread too
703 - Updates pthread m4 check from upstream
704
705 *******************************************************************************
706 Version 1.10.1
707 *******************************************************************************
708
709 2019-11-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
710
711 Forward port of Karlchen's github #122
712 Fix format string for ExtraHeaders
713
714 *******************************************************************************
715 Version 1.10.0
716 *******************************************************************************
717
718 2019-07-19 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
719
720 Fix sed command for upnp patch in configure.ac
721
722 2019-04-03 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
723
724 Proper allocation for an array of structs and a check for an unlikely
725 overflow when calling the SIOCGIFCONF ioctl().
726
727 It still bugs me the stack requirements of getlocalhostname() and
728 UpnpGetIfInfo().
729
730 *******************************************************************************
731 Version 1.8.6
732 *******************************************************************************
733
734 2019-11-20 Karl Straussberger (Karlchen@github)
735
736 Fix format string for ExtraHeaders #122
737
738 *******************************************************************************
739 Version 1.8.5
740 *******************************************************************************
741
742 2019-07-18 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
743
744 Fixed a wrong ifdef in ssdp_device.c that was causing problems with
745 ipv6. Thanks to BraneDev @ Github
746
747 *******************************************************************************
748 Version 1.8.4
749 *******************************************************************************
750
751 2017-11-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
752
753 GitHub #57 - 1.8.3 broke ABI without changing SONAME
754 Opened by jcowgill
755
756 This change in 1.8.3 broke the ABI and therefore the SONAME should have
757 been changed (ie: age reset to 0):
758
759 EXPORT_SPEC int UpnpAddVirtualDir(
760 /*! [in] The name of the new directory mapping to add. */
761 - const char *dirName);
762 + const char *dirName,
763 + /*! [in] The cookie to associated with this virtual directory */
764 + const void *cookie,
765 + /*! [out] The cookie previously associated, if mapping is already present */
766 + const void **oldcookie);
767
768 If only the cookie argument was added, you could probably get away with
769 this because all that would happen is that a garbage value is passed
770 around without being used. With the addition of oldcookie, any old
771 programs will not initialize this value and will probably segfault when
772 libupnp tries to write to it.
773
774 *******************************************************************************
775 Version 1.8.3
776 *******************************************************************************
777
778 2017-09-07 Dave Overton <david(at)insomniavisions.com>
779
780 Add user data/cookie to virtualDir callbacks
781
782 As with the main Device APIs (UpnpRegisterRootDevice etc), it is useful
783 to have a user data/cookie pointer returned with each callback.
784
785 This patch allows one cookie per registered path which enables a variety
786 of functionality in client apps.
787
788 2017-09-03 Uwe Kleine-König <[email protected]>
789
790 Fix large file system support
791
792 libupnp uses large file support (if available). If a program linking to
793 libupnp does not however it creates mismatches in call frames. See
794 Issue #51 for the results.
795
796 This simplifies LFS support by using AC_SYS_LARGEFILE_SENSITIVE instead of
797 manually defining _LARGE_FILE_SOURCE and _FILE_OFFSET_BITS (which is
798 useless on architectures where the size of off_t is fixed).
799
800 Furthermore additional logic is introduced to catch a library user without
801 64 bit wide off_t on such a platform.
802
803 upnp.h also makes use of off_t, but as this file includes FileInfo.h, the
804 latter is the single right place for this check.
805
806 This fixes #52 which is a generalized variant of #51.
807
808 2017-08-19 Uwe Kleine-König <[email protected]>
809
810 configure.ac: Drop copying of include files
811
812 The comment suggests this is for windows compilation. It should be easily
813 possible to add the source directory as an include path to the windows
814 compiler, too, so drop this. (Otherwise this should better be done using
815 AC_CONFIG_COMMANDS.)
816
817 2017-09-03 Uwe Kleine-König <[email protected]>
818
819 Let source code use autoconfig.h not the public upnpconfig.h
820
821 The former is the one supposed to be used for internal code. upnpconfig.h is only
822 for public stuff.
823
824 2017-08-19 Uwe Kleine-König <[email protected]>
825
826 configure.ac: Fix typo s/optionnal/optional/
827
828 2017-08-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
829
830 Fix broken samples when configured with --disable-ipv6.
831
832 *******************************************************************************
833 Version 1.8.2
834 *******************************************************************************
835
836 2017-07-24 Michael Osipov
837
838 Initialize in_addr and in6_addr to avoid garbage output if never written
839
840 If any of the address families isn't available in UpnpGetIfInfo(),
841 especially IPv6, always init both structs with zero to avoid garbage
842 output with inet_ntop() to gIF_IPV4 and gIF_IPV6.
843
844 See v00d00/gerbera#112 (https://github.com/v00d00/gerbera/issues/112)
845 for consequences: bind for IPv6 will fail.
846
847 2013-10-28 Vladimir Fedoseev <va-dos(at)users.sourceforge.net>
848
849 Attached patch allows to register multiple clients from single app.
850
851 2014-11-14 Philippe <philippe44ca(at)users.sourceforge.net>
852
853 Hi - I recently compiled libupnp on C++ Builder XE7 and had to do a few
854 changes to make it work. In case this helps, I've generated a small
855 patch file.
856
857 2015-04-30 Hugo Beauzée-Luyssen <chouquette(at)users.sourceforge.net>
858
859 When building using a strict mode (-std=c++11 instead of -std=gnu++11,
860 for instance), the WIN32 macro isn't defined. The attached patch fixes
861 it by using _WIN32 instead.
862
863 2015-02-06 Jean-Francois Dockes <[email protected]>
864
865 Queue events on their subscription object instead of adding them to the
866 thread pool immediately.
867
868 Events destined for a non-responding control point would flood the
869 thread pool and prevent correct dispatching to other clients, sometimes
870 to the point of disabling the device. Events are now queued without
871 allocating thread resources and properly discarded when a client is not
872 accepting them.
873
874 2015-02-03 Jean-Francois Dockes <[email protected]>
875
876 genaInitNotify()/genaInitNotifyExt() and
877 genaNotifyAll()/genaNotifyAllExt() are relatively complicated methods
878 which only differ by the format of an input parameter. This update
879 extracts the common code for easier maintenance, esp. relating to the
880 queueing modifications to follow.
881
882
883 *******************************************************************************
884 Version 1.8.1
885 *******************************************************************************
886
887 2017-04-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
888
889 Fix some compiler warning messages on md5.c
890
891 2017-03-07 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
892
893 Enable IPv6 by default
894
895 2017-03-07 Fabrice Fontaine <fontaine.fabrice(at)gmail.com>
896
897 Move threadutil source code to libupnp
898
899 With this patch, threadutil library is removed as the only public
900 header that has been kept in 1.8.x is ithread.h which is mainly a
901 wrapper to pthread with inline functions.
902 threadutil source code will now be a part of libupnp library.
903
904 *******************************************************************************
905 Version 1.8.0
906 *******************************************************************************
907
908 2014-01-15 Peng <howtofly(at)gmail.com>
909
910 Fix memory leaks.
911
912 2013-04-27 Thijs Schreijer <thijs(at)thijsschreijer.nl>
913
914 Renamed SCRIPTSUPPORT to IXML_HAVE_SCRIPTSUPPORT for consistency. Also
915 updated autoconfig and automake files, so it also works on non-windows.
916 Option is enabled by default, because it adds an element to the node
917 structure. Not using an available field is better than accidentally
918 using an unavailable field.
919
920 2012-07-11 Thijs Schreijer <thijs(at)thijsschreijer.nl>
921
922 Changed param to const UpnpAcceptSubscriptionExt() for consistency
923
924 2012-06-07 Thijs Schreijer <thijs(at)thijsschreijer.nl>
925
926 updated ixmlDocument_createAttributeEx() and ixmlDocument_createAttribute()
927 to use parameter DOMString instead of char * (same but now consistent)
928
929 2012-05-06 Thijs Schreijer <thijs(at)thijsschreijer.nl>
930
931 Added script support (directive SCRIPTSUPPORT) for better support of
932 garbage collected script languages. The node element gets a custom tag
933 through ixmlNode_setCTag() and ixmlNode_getCTag(). And a callback upon
934 releasing the node resources can be set using ixmlSetBeforeFree()
935 See updated readme for usage.
936
937 2012-03-24 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
938
939 SF Bug Tracker id 3510595 - UpnpDownloadXmlDoc : can't get the file
940
941 Submitted: Marco Virgulti ( mvirg83 ) - 2012-03-23 10:08:08 PDT
942
943 There is a problem, perhaps, during downloading a document by
944 UpnpDownloadXmlDoc. During debugging i've found that in an not exported
945 api (unfortunately i forgot the code line...) where it is set a
946 local variable "int timeout" to -1 then passed directly to another
947 function for sending data through tcp socket. I patched this setting it
948 to 0 (there is an IF section that exits if timeout < 0). It is normal
949 behavior or it is a bug?
950
951 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
952
953 Check for NULL pointer in TemplateSource.h
954
955 calloc can return NULL so check for NULL pointer in CLASS##_new and
956 CLASS##_dup.
957
958 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
959
960 Replace strcpy with strncpy in get_hoststr
961
962 Replace strcpy with strncpy to avoid buffer overflow.
963
964 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
965
966 Memory leak fix in handle_query_variable
967
968 variable was never freed.
969
970 2011-02-07 Chandra Penke <chandrapenke(at)mcntech.com>
971
972 Add HTTPS support using OpenSSL. HTTPS support is optional and can
973 be enabled by passing the --enable-open-ssl argument to the
974 configure script.
975
976 The following methods are introduced to the public API:
977 UpnpInitOpenSslContext
978
979 When enabled, HTTPS can be used by using "https://" instead of
980 "http://" when passing URLs to the HTTP Client API.
981
982 2011-02-07 Chandra Penke <chandrapenke(at)mcntech.com>
983
984 Refactor HTTP Client API to be more generic.
985
986 The following features are added:
987
988 - Support for persistent HTTP connections (reusing HTTP
989 connections). This is still a work in progress and relies on
990 applications to interpret the 'Connection' header
991 appropriately.
992
993 - Support for specifying request headers when making
994 requests. Useful for interacting with web services that require
995 custom headers.
996
997 - Support for retrieving response headers (this is a API only
998 change, some more work needs to be done to implement the actual
999 functionality. Specifically copy_msg_headers in httpreadwrite.c
1000 needs to be implemented)
1001
1002 - Common API for all HTTP methods.
1003
1004 - Support for PUT, and DELETE methods.
1005
1006 The following methods are introduced to the public HTTP Client API
1007
1008 UpnpOpenHttpConnection, UpnpCloseHttpConnection, UpnpMakeHttpRequest,
1009 UpnpWriteHttpRequest, UpnpEndHttpRequest, UpnpGetHttpResponse,
1010 UpnpReadHttpResponse.
1011
1012 Removed a lot of duplicate code in httpreadwrite.c
1013
1014 2011-01-17 Chandra Penke <chandrapenke(at)mcntech.com>
1015
1016 Include upnpconfig.h in FileInfo.h to automatically include large
1017 file macros
1018
1019 2011-01-17 Chandra Penke <chandrapenke(at)mcntech.com>
1020
1021 Fix for warnings Apple systems related to macros defined in list.h.
1022
1023 In list.h, in apple systems, undefine the macros prior to defining them.
1024
1025 2011-01-16 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1026
1027 Fix for UpnpFileInfo_get_LastModified() in http_MakeMessage().
1028
1029 UpnpFileInfo_get_LastModified() returns time_t, and http_MakeMessage()
1030 takes a "time_t *". Thanks to Chandra Penke for pointing the bug.
1031
1032 2010-11-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1033
1034 Template object for ssdp_ResultData.
1035
1036 2010-11-10 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
1037
1038 Support for "polling" select in sock_read_write.
1039
1040 Currently, in sock_read_write function, if the timeout is 0, pupnp
1041 realizes a "blocking" select (with an infinite timeout). With this
1042 patch, if timeout is set to 0, pupnp will realize a "polling" select
1043 and returns immediately if it can not read or write on the socket. This
1044 is very useful for GENA notifications when pupnp is trying to send
1045 events to a disconnected Control Point. "Blocking" select can now be
1046 done by putting a negative timeout value.
1047
1048 2010-09-18 Chandra Penke <chandrapenke(at)mcntech.com>
1049
1050 This is a minor build fix. The new Template*.h files added in the latest
1051 code need to be exported. Patch against the latest sources is attached.
1052
1053 2010-08-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1054 * upnp/src/api/Discovery.c: Fix a serious bug and memory leak in
1055 UpnpDiscovery_strcpy_DeviceType(). Thanks to David Blanchet for the
1056 patch.
1057
1058 2010-04-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1059 Separation of the ClientSubscription object.
1060
1061 2010-04-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1062 Protect the object destructors against null pointers on deletion, which
1063 should be something valid.
1064
1065 2010-03-27 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1066 SF Patch Tracker [ 2987390 ] upnp_debug vs. ixml_debug
1067 Thanks for the load of updates, I'm still assimilating them ! Could I make
1068 a suggestion though? The addition of printNodes(IXML_Node) to upnpdebug a
1069 dds a new dependency on ixml.h for anything using upnpdebug.h. I'm making
1070 quite a bit of use of upnpdebug in porting things to version 1.8.0, and I'd
1071 prefer it if printNodes could be added to ixmldebug.h instead. I'm attach
1072 ing a patch, what do you think ?
1073
1074 Nick
1075
1076 2010-03-27 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1077 * Forward port of svn revision 505:
1078 SF Patch Tracker [ 2836704 ] Patch for Solaris10 compilation and usage.
1079 Submitted By: zephyrus ( zephyrus00jp )
1080
1081 2010-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1082 * SF Patch Tracker [ 2969188 ] 1.8.0: patch for FreeBSD compilation
1083 Submitted By: Nick Leverton (leveret)
1084 Fix the order of header inclusion for FreeBSD.
1085
1086 2010-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1087 * Forward port of svn revision 502:
1088 SF Patch Tracker [ 2836704 ] Search for nested serviceList (not
1089 stopping at the first lis
1090 Submitted By: zephyrus ( zephyrus00jp )
1091
1092 Internet Gateway Device description contains nested serviceList (rootdevice
1093 -> servicelist, subdevice
1094 and subdevice has the lower-level serviceList, etc..)
1095
1096 Unfortunately, the sample code sample_util.c used by tv_device sample,
1097 etc.
1098 has a code that looks for only the first top-level serviceList.
1099 This results in the failure to read all the services of an IGD xml
1100 description.
1101
1102 Attached patch modifies this behavior and looks for the service by
1103 visiting all the serviceList in xml document in turn.
1104
1105 With the modified patch (ad additional modification), I could
1106 simulate an IGD device and created a modified control program for that.
1107
1108 Patch against 1.6.6
1109
1110 TIA.
1111
1112 2010-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1113 * SF Patch Tracker [ 2973319 ] Problem in commit 499
1114 Submitted By: Nick Leverton (leveret)
1115 Afraid that this doesn't compile, it seems retval should be retVal in two
1116 places.
1117
1118 2010-03-16 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1119 * Fix for the ithread_mutex_unlock() logic in UpnpInit().
1120 Thanks for Nicholas Kraft.
1121
1122 2010-03-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1123 * SF Patch Tracker [ 2962606 ] Autorenewal errors: invalid SID,
1124 too-short renewal interval
1125 Submitted By: Nick Leverton (leveret)
1126
1127 Auto-renewals send an invalid SID due to a missing UpnpString_get_String
1128 call. They also send a renewal interval of 0 instead of copying it from
1129 the original subscription.
1130
1131 2010-03-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1132 * SF Patch Tracker [ 2964685 ] patch for avoiding inet_ntoa (1.8.0)
1133 Submitted By: Nick Leverton (leveret)
1134
1135 Seems like SF's tracker won't let me add a patch to someone else's issue ?!
1136 This refers to https://sourceforge.net/support/tracker.php?aid=2724578
1137
1138 The calls to inet_ntoa are in getlocalhostname(), which is called from
1139 UpnpInit when it is returning the bound IP address.
1140 UpnpInit/getlocalhostname hasn't been updated to IPv6, I presume this is
1141 deliberate so that it doesn't start returning IPv6 addresses and
1142 overwriting the caller's IPv4-sized allocation.
1143
1144 The attached patch just updates getlocalhostname to use inet_ntop instead
1145 of inet_ntoa, and also documents the fact that UpnpInit is IPv4 only whilst
1146 UpnpInnit2 is both IPv4 and IPv6.
1147
1148 A fuller solution might be to change UpnpInit to use some variant on
1149 UpnpGetIfInfo. UpnpInit could still be left as IPv4 only if desired -
1150 perhaps UpnpGetIfInfo could take an option for the desired address family.
1151 getlocalhostname and its own copy of the interface scanning code would then
1152 be redundant. I don't have IPv6 capability here though so I'm reluctant to
1153 change the IPv6 code, as I have no way to test it.
1154
1155 2010-03-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1156 * SF Patch Tracker [ 2724578 ] patch for avoiding memory leaks when
1157 add devices
1158
1159 each time a device been added, UpnpInit() is called, on exit, UpnpFinish()
1160 is called, but the memories allocated by ThreadPoolInit() may lost because
1161 there's no code to call ThreadPoolShutdown() to release the memories. And
1162 inet_ntoa() is not thread safe, so in my patch, I substitute inet_ntoa()
1163 with inet_ntop().
1164
1165 2010-03-14 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1166 * SF Patch Tracker [ 2964687 ] Add new string based accessors to upnp
1167 object API
1168
1169 As per email to pupnp-devel, this is the patch to add the _strget_
1170 accessors for string-like objects in the interface.
1171
1172 Will add a further patch shortly to update the sample programs.
1173
1174 2008-06-27 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1175 * Nicholas Kraft's patch to fix some IPv6 copy/paste issues. He
1176 reported to be getting infinite loops with the svn code.
1177
1178 2008-06-13 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1179 * SF Bug Tracker [ 1984541 ]
1180 ixmlDocumentToString does not render the namespace tag.
1181 Submitted By: Beliveau - belivo
1182
1183 Undoing the patch that fixed this problem. In fact, there was no
1184 problem and the patch was wrong.
1185
1186 2008-06-11 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1187 * Ingo Hofmann's patch for "Content-Type in Subscription responses".
1188 Adds charset="utf-8" attribute to the CONTENT-TYPE header line.
1189
1190 Hi,
1191
1192 I have found an inconsistency regarding the text/xml content-type
1193 returned by libupnp. It looks like only subscription responses send
1194 "text/xml" where all other messages contain "text/xml; charset="utf-8"".
1195 Since I'm working on an DLNA device the latter behavior is mandatory.
1196 I changed the according lines in gena_device.c (see attached patch).
1197 I'm not sure if it would be ok for other device to have the charset
1198 field but it would help me a lot :)
1199
1200 Best regards,
1201 Ingo
1202
1203 2008-06-04 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1204 * SF Bug Tracker [ 1984541 ]
1205 ixmlDocumentToString does not render the namespace tag.
1206 Submitted By: Beliveau - belivo
1207
1208 The problem occurs when converting a xml document using
1209 ixmlDocumentToString containing a namespace tag created with
1210 ixmlDocument_createElementNS. The namespace tag doesn't get rendered.
1211
1212 example: The following code fragment prints:
1213
1214 <?xml version="1.0"?>
1215 <root></root>
1216
1217 instead of:
1218 <?xml version="1.0"?>
1219 <root xmlns="urn:schemas-upnp-org:device-1-0"></root>
1220
1221 Code:
1222
1223 #include <stdlib.h>
1224 #include <upnp/ixml.h>
1225
1226 int main()
1227 {
1228 IXML_Document* wDoc = ixmlDocument_createDocument();
1229 IXML_Element* wRoot = ixmlDocument_createElementNS(wDoc,
1230 "urn:schemas-upnp-org:device-1-0", "root");
1231 ixmlNode_appendChild((IXML_Node *)wDoc,(IXML_Node *)wRoot);
1232 DOMString wString = ixmlDocumentToString(wDoc);
1233 printf(wString);
1234 free(wString);
1235 ixmlDocument_free(wDoc);
1236
1237 return 0;
1238 }
1239
1240 The problem was in the printing routine, not in the library data
1241 structure.
1242
1243 2008-05-31 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1244 * Charles Nepveu's suggestion of not allocating a thread for
1245 MiniServer when it is not compiled.
1246
1247 2008-05-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1248 * Ported Peter Hartley's patch to compile with mingw.
1249
1250 2008-05-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1251 * Added some debug capability to ixml.
1252
1253 2008-05-02 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1254 * Merged Charles Nepveu's IPv6 work. libupnp now is IPv6 enabled.
1255
1256 2008-02-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1257 * Breaking API so that we now hide internal data structures.
1258
1259 2008-02-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
1260 * Rewrote Peter Hartley's patch to include a new extra header field in
1261 FileInfo.
1262
1263 *******************************************************************************
1264 Version 1.6.22
1265 *******************************************************************************
1266
1267 2017-07-07 James Cowgill <james410(at)cowgill.org.uk>
1268
1269 Replace MD5 implementation with public-domain version
1270
1271 Currently the RSA MD5 implementation is used. Unfortunately the license
1272 has some potential issues:
1273 * The license does not explicitly allow distributing derivative works.
1274 This was the original argument used in
1275 [Debian #459516](https://bugs.debian.org/459516).
1276 * The license contains an advertising clause similar to the BSD 4-clause
1277 license. This is incompatible with the GPL and if it were enforced,
1278 would require RSA to be mentioned by pretty much everyone who uses pupnp.
1279
1280 The simple solution is to replace it with a public domain
1281 implementation. I've taken OpenBSDs implementation and tweaked it
1282 slightly for use by pupnp by:
1283 - Adjusting the includes.
1284 - Removing the __bounded__ attributes which are specific to OpenBSD.
1285 - Using the standard integer types from stdint.h.
1286 - Using memset instead of explicit_bzero.
1287
1288 2016-12-16 Peter Pramberger <peterpramb(at)users.sf.net>
1289
1290 ixml/test/test_document.c is missing the string.h include, therefore
1291 the compiler complains about an implicit declaration.
1292
1293 *******************************************************************************
1294 Version 1.6.21
1295 *******************************************************************************
1296
1297 2016-12-16 Gabriel Burca <gburca(at)github>
1298
1299 If the error or info log files can not be created, use stderr and
1300 stdout instead.
1301
1302 2016-12-08 Uwe Kleine-König <uwe(at)kleine-koenig.org>
1303
1304 Fix out-of-bound access in create_url_list() (CVE-2016-8863)
1305
1306 If there is an invalid URL in URLS->buf after a valid one, uri_parse is
1307 called with out pointing after the allocated memory. As uri_parse writes
1308 to *out before returning an error the loop in create_url_list must be
1309 stopped early to prevent an out-of-bound access
1310
1311 Bug: https://sourceforge.net/p/pupnp/bugs/133/
1312 Bug-CVE: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2016-8863
1313 Bug-Debian: https://bugs.debian.org/842093
1314 Bug-Redhat: https://bugzilla.redhat.com/show_bug.cgi?id=1388771
1315
1316 2016-11-30 Uwe Kleine-König <uwe(at)kleine-koenig.org>
1317
1318 miniserver: fix binding to ipv6 link-local addresses
1319
1320 Linux requires to have sin6_scope_id hold the interface id when binding to
1321 link-local addresses. This is already in use in other parts of upnp, so
1322 portability shouldn't be in the way here. Without this bind(2) fails with
1323 errno=EINVAL (although ipv6(7) from man pages 4.08 specifies ENODEV in this
1324 case).
1325
1326 Fixes: https://bugs.debian.org/813249
1327
1328 2016-09-15 Mathew Garret <(at)mjg59 (twitter)>
1329
1330 SF Bug Tracker #132 CVE-2016-6255: write files via POST
1331
1332 Submitted by: Balint Reczey in 2016-08-02
1333
1334 From Debian's BTS https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=831857 :
1335 From: Salvatore Bonaccorso [email protected]
1336 To: Debian Bug Tracking System [email protected]
1337 Subject: libupnp: write files via POST
1338 Date: Wed, 20 Jul 2016 11:03:34 +0200
1339 Source: libupnp
1340 Version: 1:1.6.17-1
1341 Severity: grave
1342 Tags: security upstream
1343 Justification: user security hole
1344 Hi
1345 See http://www.openwall.com/lists/oss-security/2016/07/18/13 and
1346 https://twitter.com/mjg59/status/755062278513319936 .
1347 Proposed fix:
1348 https://github.com/mjg59/pupnp-code/commit/be0a01bdb83395d9f3a5ea09c1308a4f1a972cbd
1349 Regards,
1350 Salvatore
1351
1352 From Mathew Garret's commit: Don't allow unhandled POSTs to write to the filesystem by default
1353
1354 *******************************************************************************
1355 Version 1.6.20
1356 *******************************************************************************
1357
1358 2016-02-22 Jean-Francois Dockes <medoc(at)users.sf.net>
1359
1360 SF Bugs #131, Creator: Jean-Francois Dockes
1361
1362 I know it sounds crazy that nobody ever saw this, but the CONTENT-LENGTH
1363 value in GENA NOTIFY messages is too small by one.
1364
1365 It appears that most current control points don't notice the extra
1366 character (an LF, which is validly there but not included in
1367 Content-Length), probably because their protocol handler is reasonably
1368 lenient, and because the missing body LF does not prevent parsing the
1369 XML. But there is a least one anal CP (Linn Kazoo) which barfs, because
1370 it reads all data until connection close and the size mismatch triggers
1371 a bug.
1372
1373 "Proof":
1374 In gena_device.c:217 (notify_send_and_recv())
1375
1376 ret_code = http_SendMessage(&info, &timeout,
1377 "bbb",
1378 start_msg.buf, start_msg.length,
1379 propertySet, strlen(propertySet),
1380 CRLF, strlen(CRLF));
1381
1382 start_msg has all the headers, including the empty line.
1383 Content-length should be strlen(propertySet) + strlen(CRLF) (2)
1384 In gena_device.c:433 (AllocGenaHeaders())
1385
1386 rc = snprintf(headers, headers_size, "%s%s%"PRIzu"%s%s%s",
1387 HEADER_LINE_1,
1388 HEADER_LINE_2A,
1389 strlen(propertySet) + 1,
1390 HEADER_LINE_2B,
1391 HEADER_LINE_3,
1392 HEADER_LINE_4);
1393
1394 HEADER_LINE_2A is "CONTENT-LENGTH: ".
1395 The following value should be strlen(propertySet) + 2
1396
1397 2016-01-07 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
1398
1399 Fix for a reported integer overflow
1400
1401 2016-01-07 Jean-Francois Dockes <medoc(at)users.sf.net>
1402 2016-01-07 Nick Leverton <nick(at)leverton.org>
1403 SF Patches #60, Creator: Jean-Francois Dockes
1404
1405 When libupnp is configured with --enable-ipv6 but ipv6 is not available
1406 on the system (for example because the ipv6 code is not loaded in a Linux
1407 kernel as is the case by default on Raspbian), the ipv6 socket creation
1408 call will fail in miniserver.c and the library init will fail, even if
1409 the ipv4 initialization would have succeeded.
1410
1411 Let a library configured with --enable-ipv6 initialize in ipv4-only
1412 mode if ipv6 is not available instead of failing. This can happen
1413 if no ipv6 code is configured or loaded in the kernel.
1414
1415 Don't fail if IPv6 is unavailable.
1416 We might be an IPv6 enabled distro build running on an IPv4-only custom kernel.
1417
1418 2016-01-07 Nick Leverton <nick(at)leverton.org>
1419
1420 SF Bug Tracker #128, Creator: Nick Leverton
1421 redefining strndup causes "error: expected identifier or '(' before '__extension__'"
1422
1423 Fix redefinition of strnlen and strndup
1424 These are available when HAVE_STRNDUP and HAVE_STRNLEN are defined, but
1425 libupnp provides an extern prototype anyway. Recent versions of glibc
1426 define this prototype differently, causing the following compile error:
1427
1428 src/api/UpnpString.c:47:15: error: expected identifier or '(' before '__extension__'
1429 extern char *strndup(__const char *__string, size_t __n);
1430
1431 2016-01-07 Nick Leverton <nick(at)leverton.org>
1432
1433 SF Bug Tracker #129, Creator: Nick Leverton
1434 shutdown() on UDP sockets logs ENOTCONN message.
1435 https://sourceforge.net/p/pupnp/bugs/129/
1436
1437 Fix ENOTCONN "Error in shutdown: Transport endpoint is not connected"
1438
1439 When logging is enabled, ssdpserver logs bursts of
1440 "Error in shutdown: Transport endpoint is not connected"
1441 This is because shutdown() is not supported for UDP sockets and under
1442 recent UNIX specifications it returns ENOTCONN if used.
1443
1444 2016-01-07 Nick Leverton <nick(at)leverton.org>
1445
1446 SF Bug Tracker #127, Creator: Klaus Fischer
1447 Miniserver uses INADDR_ANY instead of HostIP
1448 https://sourceforge.net/p/pupnp/bugs/127/
1449
1450 The internal miniserver.c uses INADDR_ANY instead of the HostIP/IfName
1451 provided when initializing libupnp. But, this HostIP/IfName gets used
1452 for the UDP socket when multicasting SSDP messages. Because of this,
1453 miniserver may end up sending from different IP address than ssdpserver.
1454 This patch causes miniserver to use the already known interface address.
1455
1456 2016-01-07 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
1457
1458 SF Bug Tracker #130, Creator: Shaddy Baddah
1459 infinite loop in UpnpGetIfInfo() under WIN32
1460
1461 Original code makes no sense. This patch should fix it.
1462
1463 2015-02-04 Shaun Marko <[email protected]>
1464
1465 Bug tracker #124 Build fails with --enable-debug
1466
1467 Build environment
1468 Fedora 21
1469 X86-64
1470 * gcc 4.9.2
1471
1472 How to repeat
1473 $ ./configure --enable debug
1474 $ make
1475 libtool: compile: gcc -DHAVE_CONFIG_H -I. -I.. -I../upnp/inc -I./inc -I../threadutil/inc
1476 -I../ixml/inc -I./src/inc -pthread -g -O2 -Wall -MT src/api/libupnp_la-UpnpString.lo
1477 -MD -MP -MF src/api/.deps/libupnp_la-UpnpString.Tpo -c src/api/UpnpString.c
1478 -fPIC -DPIC -o src/api .libs/libupnp_la-UpnpString.o src/api/UpnpString.c:47:16:
1479 error: expected identifier or '(' before 'extension'
1480 extern char *strndup(const char *string, size_t __n);
1481 ^
1482 Makefile:1016: recipe for target 'src/api/libupnp_la-UpnpString.lo' failed
1483
1484 Reason for failure
1485 Build enables -O2 optimization flags which causes the inclusion of a
1486 macro implementation of strndup from include/bits/string2.h.
1487
1488 Workarounds
1489 Disable optimization when configuring or making:
1490 $ configure CFLAGS='-g -pthread -O0' --enable-debug
1491 $ make
1492 or
1493 $ configure --enable-debug
1494 $ make CFLAGS='-g -pthread -O0' Define NO_STRING_INLINES
1495 $ export CFLAGS="-DNO_STRING_INLINES -O2"
1496 $ ./configure --enable-debug
1497 $ make
1498
1499 Fix
1500 * Don't declare strndup in src/api/UpnpString.c if it exists
1501
1502 2015-02-01 Jean-Francois Dockes <[email protected]>
1503
1504 Out-of-tree builds seem to be currently broken, because ixml and
1505 threadutil files need an include path to include UpnpGlobal.h, and
1506 configure tries to copy files into a directory which it does not create.
1507 The patch fixes both issues.
1508
1509 2014-01-03 Peng <howtofly(at)gmail.com>
1510
1511 rewrite soap_device.c
1512
1513 1) separate HTTP handling from SOAP handling
1514 2) remove repeated validity check, each check is performed exactly once
1515 3) fix HTTP status code per UPnP spec, SOAP spec and RFC 2774
1516
1517 *******************************************************************************
1518 Version 1.6.19
1519 *******************************************************************************
1520
1521 2013-11-14 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
1522
1523 SF Bug Tracker #119, Creator: Klaus Fischer
1524 Access violation due to changed usage of pthreads-win32
1525
1526 Dear libupnp developers,
1527
1528 I have experienced a crash (access violation) when using libupnp on
1529 Windows. The crash is actually located in pthreads-win32 and happens
1530 when repeatedly de-/initializing libupnp on Win32 in the same process
1531 and both libupnp and pthreads-win32 are compiled as static libraries.
1532
1533 So I'm doing this:
1534 - UpnpInit()
1535 - UpnpFinish()
1536 - UpnpInit() <- Crash
1537
1538 I am already in touch with Ross Johnson on the pthreads-win32 mailing
1539 list regarding this issue:
1540
1541 http://sourceware.org/ml/pthreads-win32/2013/msg00020.html
1542
1543 He told me the problem is that the functions
1544 pthread_win32_process_attach/detach_np() should no longer be called
1545 directly, but are invoked automatically now since version 2.9.0 of
1546 pthreads-win32, which has been released approx. 1.5 years ago. Please
1547 refer to above link for in-depth information.
1548
1549 So for proper using of latest pthreads-win32 library, those function
1550 calls should vanish inside libupnp. Could you consider adapting libupnp
1551 in that way? I would really like to use both libraries out-of-the-box
1552 without local modifications, and this issue prevents that.
1553
1554 Best regards,
1555 Klaus
1556
1557 2013-11-08 Peng <howtofly(at)gmail.com>
1558
1559 Fix several minor bugs in soap_device.c
1560
1561 1) remove redundant free
1562 2) avoid user-provided ErrStr being overwritten by the default one
1563 3) eliminated memory leak possibility in handle_query_variable
1564
1565 2013-11-08 Peng <howtofly(at)gmail.com>
1566
1567 Fix return value check of parse_uri.
1568
1569 2012-06-19 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1570
1571 SF Bug Tracker #118, Creator: T.Iwamoto
1572 tv_ctrlpt crashes after detecting a later version of tvcontrol service
1573
1574 From: [email protected]
1575 Steps to reproduce:
1576 1. Extracts and build libupnp-1.6.18
1577 $ tar -xjf /path/to/archive/libupnp-1.6.18.tar.bz2
1578 $ cd libupnp-1.6.18
1579 $ ./configure
1580 $ make
1581 2. Applies the attached patch and remake.
1582 $ patch -p1 < /path/to/patch/libupnp-1.6.18.patch
1583 $ make
1584 3. Run tv_device.
1585 $ cd upnp/sample
1586 $ ./tv_device
1587 4. Run tv_ctrlpt; the tv_ctrlpt crashes soon.
1588 $ ./tv_ctrlpt
1589 Segmentation fault (core dumped)
1590 This is an issue report about the sample program of control point.
1591 The tv_ctrlpt crashes after detecting a tvdevice that contains tvcontrol:2 or higher version of tvcontrol service.
1592 tv_ctrlpt should detect correctly such devices due to forward compatibility of control points with device.
1593 For more information about the compatibility, please refer the following document:
1594 DLNA Architectures and Protocols Part 1 2011 December - 7.3.2.1.3 (GUN:GZJXU)
1595 The attached patch changes the sample programs as below:
1596 - device: changes version of tvcontrol service from 1 to 2. This change may occur in the future.
1597 - cp: nothing changed: cp knows version 1 of tvcontrol service only.
1598 I know many vendors implements their control points based on the tv_ctrlpt, so I hope to fix this issue ASAP.
1599
1600 ==
1601
1602 From: Yoichi NAKAYAMA
1603 SEGV is caused by strcpy with NULL argument.
1604 Attached patch will avoid SEGV in strcpy, but there may be other inconsistencies.
1605
1606 > I know many vendors implements their control points based on the tv_ctrlpt,
1607
1608 I don't think so. I think tv_ctrlpt is just a sample to be used with tv_device.
1609
1610 2013-10-28 Pino Toscano <pinotree(at)users.sourceforge.net>
1611
1612 Fix compilation on GNU/Hurd
1613
1614 2013-10-28 Peng <howtofly(at)gmail.com>
1615
1616 Fix return value of http_RecvPostMessage and update httpparser.c's comments
1617
1618 2013-10-17 Peng <howtofly(at)gmail.com>
1619
1620 Fix return value of process_request and related subroutines
1621 1) Only HTTP_XXX should be return
1622 2) Make default return value work for process_request
1623
1624 2013-10-15 Peng <howtofly(at)gmail.com>
1625
1626 Fix Content-Range generation bug
1627
1628 2013-09-10 zexian chen <chenzexian88(at)gmail.com>
1629
1630 Hi,
1631
1632 I had found some bugs about memory leak on libupnp-1.6.18.
1633
1634 It may lead to memory leak when calling ThreadPoolAdd() or
1635 ThreadPoolAddPersistent() which does not return 0.
1636
1637 See the attachment for patch.
1638
1639 2013-09-03 Peng <howtofly(at)gmail.com>
1640
1641 Fix return value of config_description_doc.
1642
1643 UPNP_E_XXX should not be used instead of IXML_XXX
1644
1645 2013-09-03 Peng <howtofly(at)gmail.com>
1646
1647 Remove faulty free in GetDescDocumentAndURL.
1648
1649 temp_str, which points to part of description, should not be freed.
1650
1651 2013-09-02 Peng <howtofly(at)gmail.com>
1652
1653 Suppose the UPnP device is listening on 192.168.1.102:49152. Use the following to send
1654 garbage bytes to the device:
1655 while true; do echo "\""; done | netcat 192.168.1.102 49152
1656
1657 The device just keeps receiving these bytes and its memory usage keeps growing.
1658 Malicious client may exploit it to exhaust the device's memory.
1659
1660 The attached patch eliminates this possibility.
1661
1662 2013-09-02 Peng <howtofly(at)gmail.com>
1663
1664 1) restore the scanner's original cursor position in case of
1665 insufficient input;
1666 2) free the memories allocated for a new header in case of a failure.
1667
1668 2013-08-13 Peng <howtofly(at)gmail.com>
1669
1670 Patch to fix behavior when char is signed
1671
1672 it seems to me that there is still something wrong:
1673
1674 1) the new is_qdtext_char() is incorrect.
1675 There is a trap if char is implemented as signed char.
1676 Suppose that c is '\xFF', it will be -1 when converted to an int.
1677 By definition, c should be qdtext:
1678 qdtext = <any TEXT except <">>
1679 TEXT = <any OCTET except CTLs, but including LWS>
1680 OCTET = <any 8-bit sequence of data>
1681
1682 2) the character after '\\' could be either part of a quoted-pair
1683 (together with '\\'), or a normal qdtext, since '\\' itself can
1684 be treated as a qdtext. This is equivalent to saying that the
1685 character after '\\' in a quoted string could be ANY octet.
1686
1687 A patch based on the above two observations is attached.
1688
1689 Peng
1690
1691 2013-08-13 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
1692
1693 Enforce RFC 2616 and accept "0" after a backslash for quoted-strings.
1694
1695 Reported by Peng <howtofly(at)gmail.com>
1696
1697 2013-08-13 Peng <howtofly(at)gmail.com>
1698
1699 Patch to make scanner_get_token more robust (avoid over-reading).
1700
1701 2013-07-30 Zheng Peng <darkelf2010(at)users.sf.net>
1702
1703 SF ticket #116 UpnpRemoveVirtualDir wrong linked list operation
1704
1705 What if pVirtualDirList has two nodes and what we want to delete is the
1706 first one. Patch attached.
1707
1708 2013-07-30 Sebastian Brandt <s.brandt(at)aixtrusion.de>
1709
1710 Dear libupnp-devels,
1711 when POST'ing to the simple web server in libupnp, the application crashes.
1712 This is caused by a missing "..." argument in webserver.c:1533.
1713 Seems it has been there for a long time ... 1.6.9 and 1.6.18 have it.
1714
1715 webserver.c:1533 calls http_MakeMessage
1716 /* Send response. */
1717 http_MakeMessage(&headers, 1, 1,
1718 "RTLSXcCc",
1719 ret, "text/html", X_USER_AGENT);
1720 The format parameter RTLSXcCc needs four arguments -
1721 R - response code - ret,
1722 T- content type - text/html,
1723 L - struct SendInstruction * - NOT PRESENT
1724 X - user agent - X_USER_AGENT
1725
1726 This results in a crash.
1727
1728 Changing to
1729 http_MakeMessage(&headers, 1, 1,
1730 "RTLSXcCc",
1731 ret, "text/html", &RespInstr, X_USER_AGENT);
1732 solves the situation.
1733
1734 Yours,
1735 Sebastian Brandt
1736
1737 *******************************************************************************
1738 Version 1.6.18
1739 *******************************************************************************
1740
1741 2012-12-06 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
1742
1743 Security fix for CERT issue VU#922681
1744
1745 This patch addresses three possible buffer overflows in function
1746 unique_service_name(). The three issues have the following CVE numbers:
1747
1748 CVE-2012-5958 Issue #2: Stack buffer overflow of Tempbuf
1749 CVE-2012-5959 Issue #4: Stack buffer overflow of Event->UDN
1750 CVE-2012-5960 Issue #8: Stack buffer overflow of Event->UDN
1751
1752 Notice that the following issues have already been dealt by previous
1753 work:
1754
1755 CVE-2012-5961 Issue #1: Stack buffer overflow of Evt->UDN
1756 CVE-2012-5962 Issue #3: Stack buffer overflow of Evt->DeviceType
1757 CVE-2012-5963 Issue #5: Stack buffer overflow of Event->UDN
1758 CVE-2012-5964 Issue #6: Stack buffer overflow of Event->DeviceType
1759 CVE-2012-5965 Issue #7: Stack buffer overflow of Event->DeviceType
1760
1761 2012-06-19 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1762
1763 Fix memory leak and access violation in UpnpSendAction(Ex)Async.
1764
1765 Free buffers after malloc or ixmlPrintNode failure.
1766 Free Param->Header before destructing Param.
1767
1768 2012-05-25 Anoop Mohan <anoop.anoop(at)gmail.com>
1769
1770 This patch fixes a bug in non blocking connect call where the sock
1771 option length for SO_ERROR was passed as 0 instead of sizeof(int).
1772
1773 2012-04-24 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1774
1775 Disable SetGenaCallback call if device is disabled.
1776
1777 If device is disabled, SetGenaCallback definition is disabled,
1778 but its call remains. A link error will occur in Win32.
1779
1780 2012-04-21 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1781
1782 Fix condition for allocation failure in get_content_type().
1783
1784 At the end of get_content_type() in webserver.c, it should check
1785 return value of ixmlCloneDOMString().
1786
1787 2012-04-21 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1788
1789 Fix problems detected as dead assignment warning by clang scan-build.
1790
1791 Wrong assignment by shutdown result hides the real error code
1792 of NewRequestHandler() in ssdp_device.c.
1793 Fix return code description of NewRequestHandler().
1794 Handle return code from ithread_create in sample applications.
1795 Remove unused assignments.
1796
1797 2012-04-20 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1798
1799 Avoid dereference of null pointer in ixmlNode_setNodeProperties.
1800
1801 The problem can occur if one of the arguments is NULL.
1802 Test argument and fix assertion.
1803
1804 2012-04-17 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1805
1806 Create intermediate directory per project on vc9.
1807
1808 Sample applications share sample_util.c and collisions of
1809 object file can occur in parallel build. Modify project files to
1810 split intermediate directories against it.
1811 Apply similar changes also to library projects, like vc10 projects.
1812
1813 2012-04-11 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1814
1815 Avoid access violation after parser_parse_chunky_headers call.
1816
1817 In parser_parse_chunky_headers, parser->msg.msg.buf can be changed
1818 by membuffer_delete call. Therefore if we save the pointer to
1819 parser->msg.entity.buf before calling membuffer_delete, it will
1820 induce access to released memory.
1821
1822 2012-04-06 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1823
1824 Remove possibility of access violation.
1825
1826 1. Test Instr before dereference it in http_RecvPostMessage.
1827 (Though it never becomes NULL because NULL is not passed to
1828 the static method)
1829 2. Avoid strdup(NULL) in ixmlElement_setAttributeNS.
1830 Those are detected by llvm scan-build.
1831
1832 2012-04-05 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1833
1834 SF Bug Tracker id 3507819 - Use of thread-unsafe gmtime() in httpreadwrite.c
1835
1836 Submitted: zephyrus ( zephyrus00jp ) - 2012-03-18 06:31:00 PDT
1837
1838 Define http_gmtime_r and web_server_asctime_r and use it.
1839 Those prefix are added since pthread for Win32 already
1840 has macro gmtime_r and asctime_r.
1841
1842 2012-04-05 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1843
1844 Fix type of local variable stopSock in RunMiniServer()
1845
1846 The variable is declared as SOCKET, but it is used to
1847 store return value of int receive_from_stopSock(...).
1848 The type was changed in the commit
1849 4b47e6a51d9c7049a862695b68de75699e023551 by mistake.
1850
1851 2012-04-03 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1852
1853 Compilation optimization
1854
1855 Do not compile the code related to the webserver in http_SendMessage
1856 when --disable-webserver is set
1857
1858 *******************************************************************************
1859 Version 1.6.17
1860 *******************************************************************************
1861
1862 2012-04-02 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1863
1864 SF Bug Tracker id 3514145 - Memory leak fix in threadutil
1865
1866 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-04-02 06:49:20 PDT
1867
1868 Put thread in a detached state when calling ithread_create otherwise in
1869 some circumstances, thread can end before the call to ithread_detach.
1870
1871 2012-03-30 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1872
1873 Add --enable-unspecified_server
1874
1875 Add --enable-unspecified_server configure option to set to "Unspecified"
1876 the OS name, OS version, product name and product version normally
1877 contained in the SERVER header as this could be used by an attacker.
1878
1879 2012-03-29 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1880
1881 Removing implicit casts in miniserver.c
1882
1883 Removing implicit integer or enum casts in miniserver.c.
1884
1885 2012-03-29 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1886
1887 SF Bug Tracker id 3512833 - Miniserver is wrongly disabled
1888
1889 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-29 07:36:34 PDT
1890
1891 Miniserver is disabled if EXCLUDE_GENA, EXCLUDE_SOAP and
1892 EXCLUDE_WEBSERVER are set.
1893 However, SSDP needs the Miniserver to answer to M-SEARCH requests.
1894 So, MiniServer should not be disabled if EXCLUDE_SSDP is not also set.
1895
1896 2012-03-26 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1897
1898 SF Bug Tracker id 3510693 - build fail with --disable-device
1899
1900 Use INCLUDE_DEVICE_APIS instead of UPNP_HAVE_DEVICE as in other sources.
1901 Don't use soap_device_callback if INCLUDE_DEVICE_APIS is not set,
1902 otherwise link error occur on Windows.
1903
1904 2012-03-26 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1905
1906 SF Bug Tracker id 3511149 - --disable-ssdp has no effect
1907
1908 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-03-25 18:14:34 PDT
1909
1910 There are typos in upnp/src/inc/config.h "EXCLUDE_SSSDP" (should be
1911 EXCLUDE_SSDP), therefore EXCLUDE_SSDP is always 0, and --disable-ssdp
1912 has no effect.
1913
1914 2012-03-24 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1915
1916 Remove implicit casts
1917
1918 Cast parameters of htonl in uint32_t in IN6_IS_ADDR_GLOBAL and
1919 IN6_IS_ADDR_ULA definitions.
1920 Remove comparison with 0 in while statement of vfmatch,
1921 http_SendMessage and http_MakeMessage.
1922
1923 2012-03-24 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
1924
1925 SF Bug Tracker id 3510693 - build fail with --disable-device
1926
1927 GetDeviceHandleInfo just fail without using undefined member DeviceAf
1928 if UPNP_HAVE_DEVICE is not defined.
1929 Move ContentTypeHeader definition to soap_common.c, since it is
1930 also used in soap_ctrlpt.c.
1931
1932 *******************************************************************************
1933 Version 1.6.16
1934 *******************************************************************************
1935
1936 2012-03-18 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1937
1938 Replace sprintf by snprintf in http_WriteHttpPost
1939
1940 Replace sprintf by snprintf in http_WriteHttpPost to avoid buffer
1941 overflow.
1942
1943 2012-03-18 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1944
1945 Add infoSize parameter to get_sdk_info
1946
1947 Add infoSize parameter to get_sdk_info function to replace sprintf call
1948 by a snprintf call.
1949
1950 2012-03-16 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1951
1952 Check return code in ixml
1953
1954 Check return code of ixmlDocument_CreateElementEx in
1955 ixmlDocument_CreateElement.
1956 Check return code of ixmlNode_setNodeName and ixmlNode_setNodeValue in
1957 ixmlNode_cloneCDATASect and ixmlNode_cloneTextNode.
1958
1959 2012-03-16 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1960
1961 Add more explicit casts and remove dead code
1962
1963 Comment unused SERVER from DeviceShutdown.
1964 Comment unused max from parse_hostport.
1965 Comment unused nodeptr from ixmlNode_cloneDoc.
1966 Comment unused newNode from Parser_hasDefaultNamespace.
1967 Comment unused Parser_parseReference function
1968 Check return code of shutdown and display an error if needed.
1969
1970 2012-03-15 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1971
1972 Fix previous commit
1973
1974 Replace HAVE_UPNP_OPTSSDP by UPNP_HAVE_OPTSSDP in upnpapi.c.
1975
1976 2012-03-15 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1977
1978 Add --disable-optssdp option
1979
1980 Modify configure.ac to add --disable-optssdp option. This option will
1981 remove OPT, 01-NLS and X_USER_AGENT headers from SSDP messages as those
1982 headers are optional. If --disable-gena and disable-optssdp are both
1983 used, uuid part will not be compiled anymore.
1984
1985 2012-03-15 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1986
1987 Bug fix in ixmlNode_allowChildren
1988
1989 Commit d48d73720bd325062c4d3b9ce85f3944be4f562d added a bug in
1990 ixmlNode_allowChildren, this function was returning FALSE instead of
1991 TRUE when newChild->nodeName was eELEMENT_NODE.
1992
1993 2012-03-15 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
1994
1995 Improve upnp/genlib/net
1996
1997 Change ret_code from int to parse_status_t in match.
1998 Set back return code of ReadResponseLineAndHeaders from parse_status_t
1999 to int as this function can return UPNP_E_BAD_HTTPMSG. As a result, do
2000 not cast the result of this function into parse_status_t in
2001 http_OpenHttpGetProxy and http_OpenHttpGetEx.
2002 Use switch with PARSE_OK in parsetools.c.
2003 Add missing explicit casts of integer constants in uri.c and
2004 httpreadwrite.c.
2005 Use switch, int and sa_family_t with AF_INET in uri.c.
2006 Print an error in http_Download if realloc failed.
2007
2008 2012-03-14 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2009
2010 Use switch instead of if with enums in upnpapi.c
2011
2012 Replace if statements with switch when using HND_DEVICE and HND_CLIENT
2013 enum constants.
2014 Correct also UpnpUnRegisterRootDeviceLowPower and UpnpUnRegisterClient
2015 as those functions were wrongly awaiting an UPNP_E_INVALID_HANDLE
2016 instead of HND_INVALID from GetHandleInfo.
2017
2018 2012-03-14 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2019
2020 Improve ssdp part
2021
2022 Do not compile CreateClientRequestPacketUlaGua if IPv6 is disable.
2023 Cast DestAddr->sa_family from sa_family_t into int when calling
2024 CreateServicePacket as this function has been set back to accept int in
2025 a692e591defe6ed9a617b9b4a083964a01f7bbab.
2026 Use switch instead of if with AF_INET and AF_INET6.
2027 Add missing casts from AF_INET and AF_INET6 into sa_family_t when using
2028 them to set sin_family and sin6_family.
2029 Add missing explicit casts into size_t or lu when using integer
2030 constants with strlen or unsigned long indexes.
2031 Set SSDP_PAUSE to be unsigned as it is used with usleep.
2032
2033 2012-03-14 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2034
2035 Use switch instead of if with enums in ixml
2036
2037 Replace if statements with switch when using enums in ixml.
2038 Remove unneeded initialization in ixmlAttr_init, Parser_init and
2039 ixmlNode_init which was added by wrongly added in commit
2040 06660b6383c438e4e2c9ca9854077cecc4da9e5d.
2041
2042 2012-03-14 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2043
2044 Use switch instead of if with enums in threadutil
2045
2046 Replace if statements with switch when using enums in threadutil.
2047
2048 2012-03-14 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2049
2050 Fix missing break in http_RecvMessage
2051
2052 There was a missing break in PARSE_INCOMPLETE_ENTITY due to commit
2053 2eb3e069badd5c8676738c3ead37f9551fd8448e.
2054
2055 2012-03-14 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2056
2057 Fix parse failure observed with tvdevice sample.
2058
2059 Commit c40d2bc0c9b60c43b641ac4669c7b8bbcd6134c5 has a problem
2060 at removing the parentheses in parser_parse_responseline.
2061 Difference of pointers was used with intention, don't cast
2062 them separately.
2063
2064 2012-03-14 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2065
2066 Fix compile error on Windows.
2067
2068 Include UpnpStdInt.h for ssize_t.
2069 Define sa_family_t in UpnpInet.h.
2070
2071 2012-03-14 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2072
2073 Avoid ambiguous change of SsdpEvent in unique_service_name.
2074
2075 Handle overflow before changing SsdpEvent.
2076 Because the behavior of "snprintf" is platform dependent in such case.
2077
2078 2012-03-14 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2079
2080 SF Bug Tracker id 3502958 - The commit 5944960e prevents a pupnp client
2081 (amule) from receiving replies from an IGD device.
2082
2083 Previous change broke the feature. The error of unique_service_name
2084 in ssdp_request_type should be ignored.
2085 This reverts commit 5944960e172a797a9fcc196291f4046cafa7f6ec.
2086
2087 2012-03-13 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2088
2089 Address family is an int
2090
2091 Reference: "man 2 socket".
2092
2093 2012-03-11 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2094
2095 Remove more implicit casts in upnp part
2096
2097 Remove more "implicit integer or enum conversions" errors as well as
2098 dead code.
2099
2100 2012-03-11 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2101
2102 Remove more implicit casts in upnp part
2103
2104 Remove more "implicit integer or enum conversions" as well as memset
2105 before snprintf.
2106
2107 2012-03-11 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2108
2109 Avoid out of range access in CheckOtherHTTPHeaders.
2110
2111 There was a problem in HDR_ACCEPT_LANGUAGE case.
2112 It may read from TmpBuf larger amount than allocated,
2113 since condition was always true.
2114 Terminate RespInstr->AcceptLanguageHeader correctly.
2115 Skip allocation if there is already sufficient buffer.
2116
2117 2012-03-10 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2118
2119 Remove some of the implicit cast in upnp part
2120
2121 Remove some of the "implicit integer or enum conversions" as well as
2122 some access to NULL reference in upnp part.
2123
2124 2012-03-10 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2125
2126 Remove lock in ThreadPoolInit
2127
2128 If ThreadPoolInit returned EAGAIN, tp->lock was not freed.
2129
2130 2012-03-10 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2131
2132 Improve ixml
2133
2134 Remove "implicit integer conversions" and
2135 "dereference NULL return value" errors in ixml part.
2136
2137 2012-03-10 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2138
2139 Exclude IPv6 stuff in SearchByTarget when UPNP_ENABLE_IPV6 is not defined.
2140
2141 2012-03-10 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2142
2143 Further measures against buffer overflows.
2144
2145 * Pass output buffer size to CreateClientRequestPacket(UlaGua)
2146 from SearchByTarget and detect overflow.
2147 * Handle SearchByTarget error in UpnpSearchAsync.
2148 * Pass output buffer size to addrToString and detect overflow.
2149 * Handle addrToString error in configure_urlbase.
2150 * Handle overflow in http_SendMessage.
2151 * Respect unique_service_name error in ssdp_request_type
2152 so as not to touch non-terminated buffer under Evt.
2153 * Treat large argument as error in UpnpAddVirtualDir.
2154 * Use strncpy with the standard way in readFromSSDPSocket.
2155 * Do not clear buffer before snprintf.
2156 * Clarify the last argument of GetDescDocumentAndURL has size LINE_SIZE.
2157 * For inet_ntop, use buffer with size INET6_ADDRSTRLEN or INET_ADDRSTRLEN.
2158
2159 2012-03-10 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2160
2161 Handle allocation error in strndup to avoid access violation.
2162
2163 Return NULL before calling strncpy.
2164 Platforms with HAVE_STRNDUP are not affected.
2165
2166 2012-03-10 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2167
2168 Synchronize autoconfig.h with upnpconfig.h.
2169
2170 It fixes WIN32 build where configure is not invoked.
2171
2172 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2173
2174 More compilation optimization
2175
2176 Do not compile most of service_table.c and client_table.c if
2177 --disable-gena is used.
2178 Do not compile urlconfig.c if --disable-webserver is used.
2179 Adding new UPNP_HAVE_xxx variables in upnpconfig.h and upnpconfig.h.in.
2180
2181 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2182
2183 Optimization of --disable-webserver
2184
2185 Do not compile webserver.c if --disable-webserver is used.
2186
2187 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2188
2189 Improve threadutil
2190
2191 Remove "dereference NULL return" errors and implicit conversions to
2192 double or enum types.
2193
2194 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2195
2196 Optimization of --disable-webserver
2197
2198 Do not compile miniserver.c if --disable-webserver is used.
2199
2200 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2201
2202 Adding configure options
2203
2204 Adding --disable-ssdp, --disable-soap, --disable-gena options to
2205 configure script.
2206
2207 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2208
2209 Bug fix of last commit
2210
2211 _snprintf was wrongly defined in ssdp_server.c
2212
2213 2012-03-09 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2214
2215 SF Bug Tracker id 3499781 - msvc doesn't have snprintf
2216
2217 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-03-08 10:18:39 PST
2218
2219 97a17ff5add73c97844e2fa74456bab4df0800f1 commit breaks build on
2220 windows/msvc since there is no snprintf.
2221
2222 Note:
2223 * Some existing sources use _snprintf when WIN32 is defined, but its
2224 behavior is a bit different from C99 snprintf.
2225 * snprintf does terminate the buffer, so the commit (use buffer size
2226 minus 1 as argument) changes the behavior at the boundary.
2227 * Truncation might be better than crash in some cases. But it may
2228 result in not good.
2229
2230 2012-03-08 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2231
2232 SF Bug Tracker id 3499878 - UpnpUnSubscribeAsync(): ‘retVal’ may be used uninitialized
2233
2234 Submitted: Marcelo Roberto Jimenez ( mroberto ) - 2012-03-08 12:38:57 PST
2235
2236 src/api/upnpapi.c: In function ‘UpnpUnSubscribeAsync’:
2237 src/api/upnpapi.c:2060:6: warning: ‘retVal’ may be used uninitialized in this function
2238
2239 2012-03-08 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2240
2241 SF Bug Tracker id 3175217 - Crash bug in Parser_addNamespace()
2242
2243 Submitted: Terry Farnham ( tfarnham ) - 2011-02-07 09:25:25 PST
2244
2245 Details: The strcmp(pNode->prefix,pCur->prefix) crashes on pCur->prefix
2246 being NULL. This occurs on invalidly formatted xml where a node uses an
2247 undefined namespace. I would expect to receive IXML_FAILED in this
2248 situation.
2249
2250 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2251
2252 Removing access to NULL pointers in node.c and element.c
2253
2254 Check that newNode is not NULL ixmlNode_cloneNodeTree and pass newAttr
2255 as the return node in the ixmlElement_setAttributeNodeNS call of
2256 ixmlElement_setAttributeNS.
2257
2258 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2259
2260 Memory leaks correction in upnpapi.c
2261
2262 Fix memory leaks in UpnpUnSubscribe, SendActionExAsync and
2263 RenewSubscription.
2264
2265 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2266
2267 SF Bug Tracker id 3496993 - Write after free in ixmlNode_insertBefore
2268
2269 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 04:54:40 PST
2270
2271 If ixmlNode_isParent(nodeptr, newChild) returns TRUE,
2272 ixmlNode_removeChild(nodeptr, newChild, NULL) will free newChild before
2273 the modifications of newChild->nextSibling and newChild->prevSibling.
2274
2275 2012-03-08 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2276
2277 Remove most of strcpy, sprintf and strcat
2278
2279 Replace strcpy, sprintf and strcat by strncpy, snprintf and strncat to
2280 avoid buffer overflows.
2281
2282 2012-03-07 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2283
2284 SF Bug Tracker id 3497714 - Buffer overflows
2285
2286 Fix compile error on WIN32.
2287
2288 Local variables must be declared first.
2289 Remove outdated comment.
2290
2291 2012-03-07 Yoichi NAKAYAMA <yoichi.nakayama(at)gmail.com>
2292
2293 Avoid access violation in assertion.
2294
2295 xmlParser->pCurElement was dereferenced before null check.
2296 Affects debug build only.
2297
2298 2012-03-07 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2299
2300 Remove SIZEOF_MISTACH error in notify_send_and_recv
2301
2302 Replace sizeof(CRLF) by strlen(CRLF) as CRLF is a const char*.
2303
2304 2012-03-07 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2305
2306 SF Bug Tracker id 3498442 - Memory leak in get_file_info
2307
2308 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-07 02:44:30 PST
2309
2310 info->contentType is not freed before being set to NULL.
2311
2312 2012-03-07 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2313
2314 SF Bug Tracker id 3498439 - Memory leak in removeServiceTable
2315
2316 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-07 02:35:46 PST
2317
2318 UDN is not freed.
2319
2320 2012-03-07 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2321
2322 SF Bug Tracker id 3498436 - Memory leak in Parser_processAttributeName
2323
2324 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-07 02:30:57 PST
2325
2326 attr is not freed if ixmlNode_setNodeProperties or
2327 ixmlNode_setAttributeNode return an error in
2328 Parser_processAttributeName.
2329
2330 2012-03-07 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2331
2332 Fixing an error in d6db7c555d0f11856ce5e3e479b16a4cf4689107 commit
2333
2334 Evt.Sid should not be cast into char* when calling sizeof otherwise
2335 size will be 4.
2336
2337 2012-03-06 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2338
2339 Removing two unused variables in ssdp_server.c
2340
2341 Removing first TempPtr allocation in unique_service_name as well as one
2342 of the dbgStr allocation in AdvertizeAndReply as those values were not
2343 used.
2344
2345 2012-03-06 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2346
2347 SF Bug Tracker id 3497714 - Buffer overflows
2348
2349 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-06 07:36:08 PST
2350
2351 Call to strcpy should be replaced by call to memset and strncpy to
2352 avoid getting buffer overflows.
2353
2354 2012-03-05 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2355
2356 SF Bug Tracker id 2989399 - UpnpSetVirtualDirCallbacks API removal in 1.6.x
2357
2358 Submitted: Nick Leverton ( leveret ) - 2010-04-19 07:44:10 PDT
2359
2360 Details: The recent codebase merge has removed a significant API call
2361 which is used by several pupnp devices such as mediatomb and gmediaserver.
2362 UpnpSetVirtualDirCallbacks() has been replaced by individual routines to
2363 set each callback. Essentially this means that 1.6.7 will in fact be a major
2364 bump and 1.6.6 devices can no longer link against it. Could we have the call
2365 reinstated please, perhaps as a wrapper around the individual calls ? As
2366 it is, all distros will have to patch their 1.6.x apps, rebuild and re-link them.
2367
2368 The other removed API calls and external variables don't seem to be used
2369 by any of the apps I have copies of, but UpnpSetVirtualDirCallbacks is
2370 important for maintaining compatibility within 1.6.x.
2371
2372 2012-03-05 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2373
2374 SF Bug Tracker id 3325246 - Memory Leak in XML Parser
2375
2376 Submitted: Terry Farnham ( tfarnham ) - 2011-06-23 09:45:54 PDT
2377
2378 Details: The following bit of xml results in a memory leak from the xml
2379 parser:
2380
2381 const char *xmlbuffer="<?xml version=\"1.0\" encoding=\"utf-8\"?>
2382 <root xmlns=\"urn:schemas-upnp-org:device-1-0\" xmlns:dlna=\"urn:schemas-dlna-org:device-1-0\">
2383 <dlna:X_DLNADOC xmlns:dlna=\"urn:schemas-dlna-org:device-1-0\">DMS-1.50</dlna:X_DLNADOC></root>";
2384
2385 When I execute the following code:
2386
2387 IXML_Document *doc = ixmlParseBuffer(xmlbuffer);
2388 ixmlDocument_free(doc);
2389
2390 It results in a memory leak in ixmlparser.c line 2107 where it calls
2391 safe_strdup( newElement->namespaceURI ); It's difficult to figure out why.
2392
2393 2012-03-05 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2394
2395 SF Bug Tracker id 3417134 - Crash seen in UpnpFinish
2396
2397 Submitted: Sunil ( sunilangadi ) - 2011-10-02 08:28:47 PDT
2398
2399 Details: I observed crash in the below mentioned log statement in
2400 function upnpfinish(file: upnpapi.c).
2401 UpnpPrintf(UPNP_INFO, API, __FILE__, __LINE__, "Exiting UpnpFinish:
2402 UpnpSdkInit is :%d:\n", UpnpSdkInit);
2403
2404 In particular it was crashing in ithread_self in
2405 UpnpDisplayFileAndLine(file upnpdebug.c) on WIN32.
2406
2407 Moving the call ithread_cleanup_library() below the upnp printf call
2408 mentioned above in function upnpfinish fixed the crash but I couldn't get
2409 to the root of the problem.
2410
2411 The problem was observed on WIN32.
2412
2413
2414 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2415
2416 SF Bug Tracker id 3497159 - Bug fix in Parser_readFileOrBuffer
2417
2418 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 14:20:58 PST
2419
2420 fileSize = ftell( xmlFilePtr ); can return a negative value, in this
2421 case the function should exit (at the moment, the function exits only
2422 if ftell returns 0).
2423
2424 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2425
2426 SF Bug Tracker id 3497140 - Bug fix in http_get_code_text
2427
2428 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 13:07:03 PST
2429
2430 Replace if( statusCode < 100 && statusCode >= 600 ) which can't be true
2431 by if( statusCode < 100 || statusCode >= 600 ).
2432
2433 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2434
2435 SF Bug Tracker id 3497126 - Resource leak in http_RecvPostMessage
2436
2437 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 12:33:59 PST
2438
2439 Fp is not closed when an error is raised on membuffer_append or
2440 sock_read.
2441
2442 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2443
2444 SF Bug Tracker id 3497034 - Buffer not null terminated in UpnpGetIfInfo
2445
2446 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 06:43:52 PST
2447
2448 gIF_NAME might be not null terminated.
2449
2450 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2451
2452 SF Bug Tracker id 3497033 - Buffer not null terminated in UpnpInit
2453
2454 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 06:42:18 PST
2455
2456 gIF_IPV4, gIF_IPV6 and gIF_IPV6_ULA_GUA might be not null terminated.
2457 Moreover, gIF_IPV4 should be 16 characters (INET_ADDRSTRLEN) and not 22
2458 and gIF_IPV6 should be 46 characters (INET6_ADDRSTRLEN) and not 65.
2459
2460 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2461
2462 Bug fix of latest commit (parse_hostport)
2463
2464 Missing parenthesis in memset.
2465
2466 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2467
2468 SF Bug Tracker id 3497027 - Buffer not null terminated in parse_hostport
2469
2470 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 06:28:38 PST
2471
2472 workbuf might be not null terminated.
2473
2474 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2475
2476 SF Bug Tracker id 3497009 - Resource leak in http_SendMessage
2477
2478 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 05:51:44 PST
2479
2480 Fp is not closed if fseeko(Fp, Instr->RangeOffset, SEEK_CUR) does not return 0.
2481
2482 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2483
2484 HInfo->ServiceTable initialization in UpnpRegisterRootDevice2 and UpnpRegisterRootDevice4
2485
2486 Initialize also HInfo->ServiceTable in UpnpRegisterRootDevice2 and
2487 UpnpRegisterRootDevice4 functions
2488
2489 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2490
2491 SF Bug Tracker id 3496703 - Handle_Info::ServiceList may have undefined value
2492
2493 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-03-03 20:49:25 PST
2494
2495 In UpnpRegisterRootDevice(), HInfo->ServiceTable is not initialized and
2496 getServiceTable() may leave it intact. It will cause crash on
2497 freeServiceTable() called from UpnpUnRegisterRootDevice().
2498
2499 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2500
2501 SF Bug Tracker id 3496702 - TvDeviceStop is called even if Start failed
2502
2503 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-03-03 20:35:08 PST
2504
2505 In sample tvdevice, error of device_main() is not handled, and
2506 TvDeviceStop() cause crash.
2507
2508 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2509
2510 SF Bug Tracker id 3496942 - Memory leak in config_description_doc
2511
2512 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 01:55:54 PST
2513
2514 element was not freed if membuffer_append_str(&url_str, "http://") does
2515 not return 0. Moreover addNew was not used.
2516
2517 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2518
2519 SF Bug Tracker id 3496938 - Missing structures initialization in some functions
2520
2521 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 01:31:16 PST
2522
2523 Memsetting to 0 some of the structures: finfo in process_request,
2524 job in readFromSSDPSocket, request in http_OpenHttpGetEx, job in
2525 genaNotifyThread, job in genaNotifyAllExt, job in genaNotifyAll,
2526 job in genaInitNotifyExt, job in genaInitNotify, LocalAddr in
2527 getlocalhostname.
2528
2529 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2530
2531 SF Bug Tracker id 3496934 - Memory leaks in getlocalhostname and UpnpGetIfInfo
2532
2533 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 01:25:42 PST
2534
2535 LocalSock is not closed if ioctl(LocalSock, SIOCGIFCONF, &ifConf);
2536 returns an error.
2537
2538 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2539
2540 SF Bug Tracker id 3496933 - Out-of-bounds access in CheckOtherHTTPHeaders
2541
2542 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-05 01:15:34 PST
2543
2544 An out-of-bands access is raised because size of
2545 RespInst->AcceptLanguageHeader is 200 and TmpBuf size is 180.
2546
2547 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2548
2549 Remove unused currentDevice variable in removeServiceTable
2550
2551 currentDevice is not used in this function.
2552
2553 2012-03-05 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2554
2555 SF Bug Tracker id 3496581 - Memory leak in getServiceList
2556
2557 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-03-03 08:43:23 PST
2558
2559 serviceNodeList is not freed if
2560 current->next = malloc(sizeof(service_info)); returns NULL.
2561
2562 2012-02-29 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2563
2564 SF Bug Tracker id 3495616 - Memory leak in ixmlElement_setAttributeNS
2565
2566 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-02-29 02:09:43 PST
2567
2568 newAttrNode is not freed if newAttr->n.nodeValue = strdup(value); returns
2569 NULL or if ixmlElement_setAttributeNodeNS(element, newAttr, NULL) does
2570 not return IXML_SUCCESS.
2571
2572 2012-02-28 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2573
2574 SF Bug Tracker id 3495286 - Double free in get_action_node
2575
2576 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-02-28 04:27:54 PST
2577
2578 ixmlFreeDOMString(ActNodeName); is called twice if
2579 ixmlParseBufferEx(ActNodeName, RespNode); does not return IXML_SUCCESS.
2580
2581 2012-02-28 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2582
2583 SF Bug Tracker id 3495280 - Memory leak in ixmlDocument_createElementEx
2584
2585 Submitted: Fabrice Fontaine ( ffontaine ) - 2012-02-28 04:01:02 PST
2586
2587 There is a memory leak in ixmlDocument_createElementEx:
2588 newElement->tagName is not freed if
2589 newElement->n.nodeName = strdup(tagName); returns NULL.
2590
2591 2012-02-27 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2592
2593 SF Bug Tracker id 3494865 - Use of non-initialized variable in parser_parse_requestline
2594
2595 Submitted: Marcelo Roberto Jimenez ( mroberto ) - 2012-02-26 16:50:23 PST
2596
2597 src/genlib/net/http/httpparser.c: In function ‘parser_parse_requestline’:
2598 src/genlib/net/http/httpparser.c:1319:28: warning: ‘index’ may be used uninitialized in this function
2599
2600 2012-02-24 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2601
2602 SF Bug Tracker id 3489999 - UpnpString leaks in genaSubscribe()
2603
2604 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-02-21 07:06:35 PST
2605
2606 In genaSubscribe() (defined in upnp/src/gena/gena_ctrlpt.c),
2607 ActualSID and EventURL will not be freed if ScheduleGenaAutoRenew
2608 returns UPNP_E_SUCCESS.
2609
2610 This fixes the an issue introduced by the previous fix.
2611
2612 2012-02-23 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2613
2614 SF Bug Tracker id 3489990 - some files are missing in tarball
2615
2616 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-02-21 06:52:57 PST
2617
2618 Released tarball (e.g. libupnp-1.6.15.tar.bz2) does not contain
2619 some files under upnp/{inc,src} used by the project files for
2620 windows (build/vc8/libupnp. vcproj and build/vc9/libupnp.vcproj).
2621
2622 This breaks build on Windows from tarball released after following changes
2623 http://pupnp.git.sourceforge.net/git/gitweb.cgi?p=pupnp/pupnp;a=commitdiff;h=0eba550da039be01211b56fea0d02d03f0a12343
2624 http://pupnp.git.sourceforge.net/git/gitweb.cgi?p=pupnp/pupnp;a=commitdiff;h=7a796b264ec7d5de5876fd6a2001c2d329709e02
2625
2626 2012-02-23 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2627
2628 SF Bug Tracker id 3489999 - UpnpString leaks in genaSubscribe()
2629
2630 Submitted: Yoichi NAKAYAMA ( yoichi ) - 2012-02-21 07:06:35 PST
2631
2632 In genaSubscribe() (defined in upnp/src/gena/gena_ctrlpt.c),
2633 ActualSID and EventURL will not be freed if ScheduleGenaAutoRenew
2634 returns UPNP_E_SUCCESS.
2635
2636 This fixes the original issue.
2637
2638 2012-02-23 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2639
2640 SF Bug Tracker: UpnpString leaks in genaSubscribe()
2641
2642 In genaSubscribe() (defined in upnp/src/gena/gena_ctrlpt.c),
2643 ActualSID and EventURL will not be freed if ScheduleGenaAutoRenew
2644 returns UPNP_E_SUCCESS.
2645
2646 2012-02-23 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2647
2648 Revert cb89781a55466703763c1b0ee67094eb401ddfe9 as suggested by
2649 Fabrice Fontaine.
2650
2651 2012-02-07 Edwin Stearns <edwin(at)vtilt.com>
2652
2653 Attached is a patch that resolved an issue I found with a server that
2654 gave its device description URI without a trailing slash (e.g.
2655 `http://127.0.0.1:5555`).
2656
2657 2012-01-04 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2658
2659 Bug fix for IPv4-mapped IPv6 addresses.
2660
2661 Setting IPv6 sockets with IPV6_V6ONLY flag to avoid getting IP packets
2662 with IPv4-mapped IPv6 addresses on IPv6 sockets.
2663
2664
2665 *******************************************************************************
2666 Version 1.6.15
2667 *******************************************************************************
2668
2669 2012-01-11 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2670
2671 Bug fix on M-SEARCH for IPv6 CPs.
2672
2673 Small bug fix on IPv6 Control Point: now CP will also send M-SEARCH on
2674 site-scope address (FF05::C) instead of only sending M-SEARCH on
2675 link-scope (FF02::C).
2676
2677 2012-01-11 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2678
2679 Retrieve IPv6 addresses in Upnp_Discovery.
2680
2681 Changing sockaddr_in into sockaddr_storage in Upnp_Discovery to be able
2682 to retrieve IPv6 addresses of devices in Control Points using pupnp.
2683
2684 2012-01-09 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2685
2686 SF Bug tracker, ID: 3469344
2687 Submitted: dimmman ( dimmman ) - 2012-01-04 01:44:29 PST
2688 Details: Looking at the code (v1.6.14, upnptools.c) for UpnpResolveURL
2689 and UpnpResolveURL2 it shows that the ExitFunction: always returns
2690 UPNP_E_SUCCESS.
2691
2692 I'm fairly sure it's a simple mistake that should have been "return ret;"
2693 in both cases.
2694
2695 Br,
2696 Jonny
2697
2698 *******************************************************************************
2699 Version 1.6.14
2700 *******************************************************************************
2701
2702 2011-10-31 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2703
2704 UPnP Low Power Support.
2705
2706 Adding two new functions (UpnpSendAdvertisementLowPower and
2707 UpnpUnRegisterRootDeviceLowPower) which can be used to specify values
2708 for the three SSDP headers defined by UPnP Low Power. Those headers are
2709 PowerState, SleepPeriod and RegistrationState.
2710
2711 2011-10-24 Fabrice Fontaine <fabrice.fontaine(at)orange.com>
2712
2713 Bug fix in IN6_IS_ADDR_GLOBAL.
2714
2715 Changing IN6_IS_ADDR_GLOBAL to accept all IPv6 addresses which have a
2716 2000::/3 prefix.
2717
2718 2011-07-20 Marc Essayan <marc.essayan(at)orange-ftgroup.com>
2719
2720 Bug Fix on M-SEARCH.
2721
2722 Do not answer to M-SEARCH using HTTP version 1.0 as specified by the
2723 UPnP Device Architecture.
2724
2725 2011-03-18 Iain Denniston <iain.denniston(at)gmail.com>
2726
2727 Fixes for compilation under Windows (specifically MSVC). Also added
2728 MSVC supported "_inline", and fixed some WIN32 specific warnings.
2729
2730 2011-03-08 Iain Denniston <iain.denniston(at)gmail.com>
2731
2732 Several fixes to correctly use SOCKET (and related) types instead of
2733 non-portable variations.
2734
2735 *******************************************************************************
2736 Version 1.6.13
2737 *******************************************************************************
2738
2739 2011-03-15 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
2740
2741 Putting ssdpReqSocks under compilation flag.
2742
2743 Putting all access to ssdpReqSock4 and ssdpReqSock6 under
2744 INCLUDE_CLIENT_APIS compilation flag to be able to compile when
2745 client part of library is disable.
2746
2747 2011-03-15 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
2748
2749 New UpnpRegisterRootDevice4 for legacy CPs.
2750
2751 Add a new UpnpRegisterRootDevice4 which allow user to specify a
2752 description URL to be returned for legacy CPs (for example, CPs
2753 searching for a v1 when the device is v2). Most of those CPs does not
2754 work if they found a v2 in the XML description, so this new function is
2755 only used to solve interoperability issues.
2756
2757 2011-03-08 Iain Denniston <iain.denniston(at)gmail.com>
2758 Fix for memory leak.
2759
2760 2011-03-08 Iain Denniston <iain.denniston(at)gmail.com>
2761 Fix and Update of MSVC9 solution and project files - now compile in all
2762 modes. Added x64 and static library targets.
2763
2764 2011-03-08 Iain Denniston <iain.denniston(at)gmail.com>
2765 Partial fix for UpnpGetIfInfo with MSVC - convert wchar string to
2766 char string (full fix requires a lot of work - potentially impacting
2767 all supported platforms)
2768
2769 2011-03-11 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2770
2771 Created the macros PRIzd and PRIzx to deal with MSVC lack of C99.
2772 Thanks to Iain Denniston for pointing it out.
2773
2774 2011-03-08 Iain Denniston <iain.denniston(at)gmail.com>
2775
2776 Fixes for headers when compiled under C++
2777
2778 2011-03-08 Iain Denniston <iain.denniston(at)gmail.com>
2779
2780 Fix for uuid_unpack incorrect shift precedence.
2781
2782 *******************************************************************************
2783 Version 1.6.12
2784 *******************************************************************************
2785
2786 2011-02-08 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2787
2788 Undo the "incorrectly exported include files".
2789
2790 Legacy applications like linux-igd and igd2-for-linux are using those
2791 API to create a thread pool for managing their GENA events.
2792
2793 Leave it to be reworked in 1.8.x.
2794
2795 *******************************************************************************
2796 Version 1.6.11
2797 *******************************************************************************
2798
2799 2011-02-07 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2800
2801 Remove PrintThreadPoolStats() from the public API. This function uses
2802 a ThreadPool object as an argument, which is not supposed to be
2803 exported.
2804
2805 2011-02-07 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
2806
2807 Major bug fix in IPv6 code.
2808
2809 Major bug fix in miniserver.c for IPv6, bug was introduced when
2810 changing implementation of get_port in November 20th 2010 ("gena:fix
2811 several compiler warnings" commit).
2812
2813 2011-02-06 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2814
2815 Fix for incorrectly exported include files.
2816
2817 The files FreeList.h, LinkedList.h, ThreadPool.h and TimerThread.h
2818 from the threautil library were being installed in the include
2819 directory of the library, incorrectly exposing internal data structure
2820 of the library.
2821
2822 2011-01-30 Chandra Penke <chandrapenke(at)mcntech.com>
2823
2824 Fix for compilation warnings of unused variables in upnpdebug.c in
2825 release builds.
2826
2827 2011-01-20 Chandra Penke <chandrapenke(at)mcntech.com>
2828
2829 Fix for Race condition can hang miniserver thread.
2830
2831 Add 'requiredThreads' field to the ThreadPool structure, to avoid
2832 a race condition when waiting for a new thread to be created. The
2833 race condition occurs when a thread is destroyed while the master
2834 thread is waiting for a new thread to be created.
2835
2836 Thanks to Chuck Thomason for pointing the problem.
2837
2838 Summary: Race condition can hang miniserver thread - ID: 3158591
2839
2840 Details:
2841 Hello,
2842
2843 I have found a race condition in the thread pool handling of
2844 libupnp-1.6.6 that periodically results in the miniserver thread
2845 getting blocked infinitely.
2846
2847 In my setup, I have the miniserver thread pool configured with 1
2848 job per thread, 2 threads minimum, and 50 threads maximum.
2849
2850 Just before the lockup occurs, the miniserver thread pool contains
2851 2 threads: one worker thread hanging around from a previous HTTP
2852 request job (let's call that thread "old_worker") and the
2853 miniserver thread itself.
2854
2855 A new HTTP request comes in. Accordingly, the miniserver enters
2856 schedule_request_job() and then ThreadPoolAdd(). In
2857 ThreadPoolAdd(), the job gets added to the medium-priority queue,
2858 and AddWorker() is called. In AddWorker(), jobs = 1 and threads =
2859 1, so CreateWorker gets called.
2860
2861 When we enter CreateWorker(), tp->totalThreads is 2, so
2862 currentThreads is 3. The function creates a new thread and then
2863 blocks on tp->start_and_shutdown. The miniserver thread expects
2864 the newly created thread to increment tp->totalThreads and then
2865 signal the condition variable to wake up the miniserver thread and
2866 let it proceed.
2867
2868 The newly created thread starts in the WorkerThread() function. It
2869 increments tp->totalThreads to 3, does a broadcast on the
2870 start_and_shutdown condition, and starts running its job. However,
2871 before the miniserver thread wakes up, "old_worker" times out. It
2872 sees that there are no jobs in any queue and that the total number
2873 of threads (3) is more than the minimum (2). As a result, it
2874 reduces tp->totalThreads to 2 and dies.
2875
2876 Now the miniserver thread finally wakes up. It checks
2877 tp->totalThreads and sees that its value is 2, so it blocks on
2878 tp->start_and_shutdown again. It has now "missed" seeing
2879 tp->totalThreads get incremented to 3 and will never be unblocked
2880 again.
2881
2882 When this issue does occur for a server device, the miniserver
2883 port remains open, but becomes unresponsive since the miniserver
2884 thread is stuck. SSDP alive messages keep getting sent out, as
2885 they are handled by a separate thread. Reproducing the issue is
2886 difficult due to the timing coincidence involved, but in my
2887 environment I am presently seeing it at least once a day. I
2888 figured out the sequence described above through addition of my
2889 own debug logs.
2890
2891 The relevant code involved in this bug has not changed
2892 substantially in libupnp-1.6.10, though I am planning to test
2893 against 1.6.10 as well in the near future.
2894
2895 Do you have any input for an elegant fix for this issue?
2896
2897 Thanks,
2898
2899 Chuck Thomason
2900
2901 2011-01-16 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2902
2903 Define _FILE_OFFSET_BITS, _LARGEFILE_SOURCE and _LARGE_FILE_SOURCE in
2904 upnpconfig.h.
2905
2906 Make these definitions available to programs using the library.
2907 Thanks to Chandra Penke for pointing the problem.
2908
2909 Summary: Problem with large file support in pupnp build - ID: 3158969
2910 Submitted: Chandra ( inactiveneurons ) - 2011-01-15 16:17:02 BRST
2911 Details:
2912 First off, I apologize in advance for the length of this comment, it's
2913 the only way I could describe the problem accurately.
2914
2915 Secondly, a brief thanks (again). The company I'm working for has been
2916 using pupnp for a massively cross-platform project which involves iphone,
2917 osx, windows, linux x86, arm, and mips hosts. It's amazing how well it
2918 works, so kudos to the maintainers!
2919
2920 We came across a problem when compiling with the following tool-chain:
2921 http://www.codesourcery.com/sgpp/lite/mips/portal/release824. The
2922 problem is the following:
2923
2924 In configure.ac the following lines exist to enable large file support:
2925
2926 AC_DEFINE([_LARGE_FILE_SOURCE], [], [Large files support])
2927 AC_DEFINE([_FILE_OFFSET_BITS], [64], [File Offset size])
2928
2929 Which in turn result in the following #defines in autoconfig.h:
2930
2931 #define _LARGE_FILE_SOURCE /**/
2932 #define _FILE_OFFSET_BITS 64
2933
2934 However, this file is not exported as part of the upnp build. Therefore,
2935 while the entire library gets built with large file support, it's
2936 possible that dependent libraries which only rely on the include files
2937 may not use large file support.
2938
2939 In the particular case of the above tool-chain, the 'off_t' type is 8
2940 bytes when large file support is enabled, but only 4 bytes when it's
2941 not. As a result part our stack built on top of pupnp, which did not
2942 have large file support (because it did not use the above autoconf
2943 directives), was relying on 'a off_t' that was 4 bytes.
2944
2945 This caused, among many things, for the UpnpFileInfo struct to break.
2946 Since the struct is completely invisible outside of pupnp (because of
2947 some template macro magic), pupnp thought that 'FileLength' field was
2948 8 bytes, but the header setter/getter methods being used by dependent
2949 libraries thought that it was 4, which caused some erratic behavior
2950 when going through pupnp's webserver and HTTP client API.
2951
2952 We put in a temporary work around by adding the following preprocessor
2953 flags: -D_LARGE_FILE_SOURCE, -D_FILE_OFFSET_BITS=64 as part of our
2954 build process. However, it's a hack, and I was wondering if I'm missing
2955 something and there's a better way to approach this.
2956
2957 Thanks,
2958 Chandra
2959
2960 2011-01-16 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2961
2962 Use config.h to test for the availability of strndup() and strnlen().
2963
2964 2011-01-14 Chandra Penke <chandrapenke(at)mcntech.com>
2965
2966 - Null termination of strndup() implementation on systems missing it.
2967 - Implementation of strnlen() on systems missing it.
2968
2969 2011-01-14 Chandra Penke <chandrapenke(at)mcntech.com>
2970
2971 Fixes transfer encoding in the HTTP client API, which is currently
2972 broken. The break was due to a regression caused by another
2973 fix (tracker 3056713), which fixed an out of memory crash when
2974 downloading large files. The previous fix changed the
2975 http_ReadHttpGet() implementation so that data already read by the
2976 user was discarded. However, it only worked for transfers where
2977 the content length was specified. This fix extends the previous
2978 implementation to cover chunked transfer encoding.
2979
2980 2011-01-14 Chandra Penke <chandrapenke(at)mcntech.com>
2981
2982 Minor change in membuffer.c to include "membuffer.h" without looking
2983 in the standard header path. This allows pupnp to build in xcode.
2984
2985 2010-12-18 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
2986
2987 SF Tracker: Patches
2988 Fedora mingw32 compilation - ID: 3138849
2989 Details:
2990 Hello. I trying compile libupnp-1.6.10 on the Fedora 14 MinGW
2991 Environment and get many errors. I create patch to fix it. With this
2992 patch i can get static library. This patch is very raw.
2993
2994 Submitted: Ivan Romanov (ivanromanov) - 2010-12-16 23:29:19 UTC
2995
2996 *******************************************************************************
2997 Version 1.6.10
2998 *******************************************************************************
2999
3000 2010-11-23 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3001
3002 Major bug fix in http_SendMessage.
3003
3004 Currently, http_SendMessage was not able to write to write a buffer
3005 due to a bad use of file_buf instead of buf. This bug was introduced by
3006 the 0197-Doxygen-reformating-compiler-warnings patch.
3007
3008 2010-11-23 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3009
3010 Returning the Sid in Upnp_Event_Subscribe.
3011
3012 Currently, Upnp_Event_Subscribe always contains an empty chain in the
3013 Sid parameter. This patch now saves the client Subscription ID in this
3014 parameter so Control Points can see and use the same SID in the
3015 Upnp_Event_Subscribe and in the Upnp_Event structures.
3016
3017 2010-11-22 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3018
3019 Two fixes from Juergen Lock <nox(at)jelal.kn-bremen.de>:
3020
3021 1. varargs: pass size of CRLF as size_t not as int:
3022
3023 --- upnp/src/gena/gena_device.c.orig
3024 +++ upnp/src/gena/gena_device.c
3025 @@ -225,7 +225,7 @@ static UPNP_INLINE int notify_send_and_r
3026 "bbb",
3027 start_msg.buf, start_msg.length,
3028 propertySet, strlen(propertySet),
3029 - "\r\n", 2);
3030 + "\r\n", sizeof "\r\n" - 1);
3031 if (ret_code) {
3032 membuffer_destroy(&start_msg);
3033 sock_destroy(&info, SD_BOTH);
3034
3035 2. Remove "b" arg here, there is no buffer passed: (this caused a pointer
3036 to be interpreted as a buffer size to be alloc'd/copied, hence the 32 GB.)
3037
3038 --- upnp/src/genlib/net/http/webserver.c.orig
3039 +++ upnp/src/genlib/net/http/webserver.c
3040 @@ -1262,7 +1262,7 @@ static int process_request(
3041 // Content-Range: bytes 222-3333/4000 HTTP_PARTIAL_CONTENT
3042 // Transfer-Encoding: chunked
3043 if (http_MakeMessage(headers, resp_major, resp_minor,
3044 - "R" "TLD" "s" "tcS" "b" "Xc" "sCc",
3045 + "R" "TLD" "s" "tcS" "Xc" "sCc",
3046 HTTP_OK, // status code
3047 finfo.content_type, // content type
3048 RespInstr, // language info
3049
3050 2010-11-15 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3051
3052 Added the convenience function UpnpResolveURL2() to upnptools.c.
3053
3054 This function avoids some unnecessary memory allocation.
3055 The memory alloc'd by this function must be freed later by the caller.
3056
3057 2010-11-10 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3058
3059 Add GENA_NOTIFICATION_xxx_TIMEOUT variable.
3060
3061 Currently, in notify_send_and_recv function, pupnp waits for
3062 HTTP_DEFAULT_TIMEOUT seconds when trying to send a GENA notification.
3063 When there is a lot of notifications with CPs which was disconnected
3064 without unsubscribing, all the pupnp threads are blocked on this
3065 timeout. To correct, this issue, this patch adds a new variable,
3066 GENA_NOTIFICATION_SENDING_TIMEOUT, which can be used to lower the
3067 timeout so GENA threads return quickly when writing is impossible. By
3068 the same mean, pupnp waits the CP's answer to the NOTIFY for
3069 HTTP_DEFAULT_TIMEOUT seconds, so this patch adds a new variable,
3070 GENA_NOTIFICATION_ANSWERING_TIMEOUT, to customize this value.
3071
3072 2010-11-10 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3073
3074 Add --disable-blocking-tcp-connections flag.
3075
3076 Currently, pupnp is using a blocking connect to sends GENA
3077 notifications. As a result, when there is a lot of notifications with
3078 CPs which were disconnected without unsubscribing, all the pupnp
3079 threads are blocked for 20s (timeout). To correct this issue, this
3080 patch replace the call to connect with a call to private_connect and add
3081 a compilation flag to disable blocking TCP connections, so if we are not
3082 able to connect to the CP, the notification is lost.
3083
3084 2010-11-07 Stefan Sommerfeld <zerocom(at)cs.tu-berlin.de>
3085
3086 Several patches for windows compatibility and fixing warnings.
3087
3088 2010-11-07 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3089
3090 PTHREAD_MUTEX_RECURSIVE on DragonFly is an enum.
3091
3092 SF Bug Tracker - ID: 3104527
3093 Submitted: OBATA Akio ( obache ) - 2010-11-07 07:10:28 BRST
3094
3095 In threadutil/inc/ithread.h, it is expected that
3096 PTHREAD_MUTEX_RECURSIVE is defined as macro. But on DragonFly BSD,
3097 it is defined as enum, so not works as expected.
3098
3099 Attachment patch treat that DragonFly BSD always
3100 have PTHREAD_MUTEX_RECURSIVE.
3101
3102 2010-11-07 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3103
3104 ftime(3) in -lcompat should not be checked.
3105
3106 SF Bug Tracker - ID: 3104521
3107 Submitted: OBATA Akio ( obache ) - 2010-11-07 07:03:44 BRST
3108
3109 In configure.ac
3110 AC_CHECK_FUNCS(ftime,, [AC_CHECK_LIB(compat, ftime)])
3111
3112 But since version 1.6.3, ftime(3) is not used, so it should be
3113 removed, or introduce unwanted linkage with -lcompat.
3114
3115 *******************************************************************************
3116 Version 1.6.9
3117 *******************************************************************************
3118
3119 2010-11-06 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3120
3121 Fix for bug introduced in samples code in svn revision 502, commit
3122 git:25c908c558c8e60eb386c155a6b93add447ffec0
3123
3124 Sample device and combo were aborting with the message:
3125 "***** SampleUtil_Initialize was called multiple times!"
3126
3127 2010-11-06 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3128
3129 Make multiple SSDP advertisements faster.
3130
3131 Put the loop to send multiple copies of each SSDP advertisements in
3132 ssdp_server.c instead of ssdp_device.c so we have only one call to
3133 imillisleep ( SSDP_PAUSE ) to speed up advertisements.
3134
3135 2010-11-05 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3136
3137 Removing unused NUM_COPY variable.
3138
3139 Previously, NUM_COPY was used in ssdp_device.c to send multiple copies
3140 of each advertisements but also multiple replies to each M-SEARCH
3141 request. As sending multiple replies is not compliant with HTTPU/MU
3142 spec, NUM_COPY has been set to 1 in an older patch. However, as this
3143 variable is not needed and has been replaced with SSDP_COPY, it has
3144 been removed.
3145
3146 2010-11-05 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3147
3148 Use SSDP_COPY to send multiple SSDP advertisements.
3149
3150 Currently, SSDP_COPY is used only to send multiple M-SEARCH requests (in
3151 ssdp_ctrlpt.c). With this patch, SSDP_COPY is also used to send multiple
3152 copies of each advertisements packets (in ssdp_device.c).
3153
3154 2010-11-01 Carl Benson <carl.benson(at)windriver.com>
3155
3156 Fix for Android build.
3157
3158 I had to do some modifications myself though, because the Android
3159 build system insists on having a file named "util.h" taking precedence
3160 in its include path, libupnp gets confused because of the same filename
3161 in upnp/src/inc/util.h
3162
3163 *******************************************************************************
3164 Version 1.6.8
3165 *******************************************************************************
3166
3167 2010-10-20 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3168
3169 Fix a long date memory leak in webserver.c:StrStr().
3170
3171 2010-10-19 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3172
3173 Bug fix in select of miniserver.c
3174
3175 Fix a bug in miniserver.c, in which maxMiniSock was wrongly declared as
3176 unsigned int and as a result it was being set to ((unsigned int)(-1)).
3177 As a result, after being incremented, it became zero, and this value
3178 was being used in the select() call.
3179
3180 Thanks to Fabrice Fontaine for helping and testing with this issue.
3181
3182 2010-10-15 Marcelo Roberto Jimenez <mroberto(at)users.sourceforge.net>
3183
3184 Fix for 100% CPU issue in select() in miniserv.c. I have also removed
3185 the sleep() call, it was just a workaround.
3186
3187 SF Bug Tracker [ 3086852 ] 99% CPU loop in miniserver.c on a non ipv6
3188 system.
3189
3190 Submitted by: Jin ( jin_eld ) - 2010-10-13 19:29:13 UTC
3191
3192 I cross compiled libupnp 1.6.7 for ARM9 using the --disable-ipv6
3193 option, my system is an ipv4 only setup.
3194
3195 I do not know why this problem only appears when running the app in the
3196 background (for instance using nohup &), but then it starts using 99%
3197 CPU.
3198
3199 I traced the problem down to the select() call in miniserver.c in the
3200 RunMiniServer() function. Select returns code 1, but errno is set to
3201 "Socket operation on non-socket", I also see this when running my app
3202 under strace.
3203
3204 I set all ...Sock6 variables to INVALID_SOCKET to make sure that they
3205 do not get added to the FD_SET and the problem is gone.
3206
3207 *******************************************************************************
3208 Version 1.6.7
3209 *******************************************************************************
3210
3211 2010-10-01 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3212
3213 Adding --disable-notification-reordering option
3214
3215 Adding a configure flag to disable GENA notification reordering as even
3216 with an imillisleep(1), this mechanism consumes too much CPU on embedded
3217 devices when there is a burst of notifications.
3218
3219 2010-09-30 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3220
3221 Bug fix when there is no service in embedded devices
3222
3223 When a device with embedded devices (like IGD) when created and one of
3224 the embedded devices did not have any service, there was a Segmentation
3225 Fault (see SF Tracker [ 2688125 ]).
3226
3227 Original SF Tracker issue follows:
3228
3229 SF Tracker [ 2688125 ] v1.6.6 crashes on sub devices without services
3230 Submitted by: Arno Willig ( akw ) - 2009-03-15 22:45:23 BRT
3231
3232 I discovered a bug, which will make libupnp (1.6.6) segfault, when you
3233 create a upnp description document with multiple devices which have
3234 sub devices, but no own services.
3235
3236 The crash occurs in genlib/service_table.c in line 977:
3237 end->next =
3238 getServiceList( currentDevice, &next_end, URLBase );
3239
3240 In this case "end" seems not to be defined, so end->next crashes.
3241
3242 Can anyone confirm this, please?
3243
3244 2010-09-28 Marc Essayan <marc.essayan(at)orange-ftgroup.com>
3245
3246 Bug fix on burst of GENA notification
3247
3248 When a lot of notifications were generated by a device in a short
3249 period of time then 100% of the CPU was used to reorder those
3250 notifications by pushing back the thread in the job queue. This
3251 mechanism has been modified so now thread sleep 1 ms before being
3252 pushed back into the job queue.
3253
3254 Removing DEFAULT_SCHED_PARAM parameter and use
3255 sched_get_priority_min(DEFAULT_POLICY) instead.
3256
3257 2010-09-22 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3258
3259 Bug fix on M-SEARCH response
3260
3261 Devices must respond to M-SEARCH requests for any supported version and the
3262 response should specify the same version as was contained in the search target.
3263 Previously, the device did not answer if the M-SEARCH request did not
3264 contain the same version number than the version number of the device.
3265
3266 2010-09-21 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3267
3268 Add Content-Language iff Accept-Language
3269
3270 Add Content-Language header in the response if and only if there is an
3271 Accept-Language header in the request.
3272
3273 2010-09-21 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3274
3275 Addition of WEB_SERVER_CONTENT_LANGUAGE parameter
3276
3277 This patch adds the WEB_SERVER_CONTENT_LANGUAGE parameter so the user can specify
3278 the language used by the device during Description and Presentation steps of UPnP
3279 through the HTTP CONTENT-LANGUAGE header.
3280 By default, the WEB_SERVER_CONTENT_LANGUAGE is an empty string so no
3281 CONTENT-LANGUAGE is added.
3282
3283 2010-09-18 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3284
3285 Customize the stack size of the threads used by pupnp through the new
3286 THREAD_STACK_SIZE variable.
3287
3288 This patch allows a user to customize the stack size of the threads used by
3289 pupnp through the new THREAD_STACK_SIZE variable. This is especially useful
3290 on embedded systems with limited memory where the user can set THREAD_STACK_SIZE
3291 to ITHREAD_STACK_MIN.
3292
3293 However, as this modification can have side effects, I set 0 as the default
3294 value, so threads will continue to use the default stack size of the system
3295 (which varies greatly as stated in
3296 https://computing.llnl.gov/tutorials/pthreads/).
3297
3298 2010-09-16 Fabrice Fontaine <fabrice.fontaine(at)orange-ftgroup.com>
3299
3300 Broken IPv6.
3301 IPv6 is currently broken in latest release of branch-1.6.x, so find a
3302 patch attached that correct the issue (small fixes on define, undef and
3303 retVal).
3304
3305 2010-09-10 Warwick Harvey <warwick.harvey(at)tieto.com>
3306 Patch to take notice of UPNP_USE_RWLOCK flag
3307
3308 The configure.ac file included with UPnP checks for the presence of the
3309 pthread_rwlock_t type, and then sets the value of the UPNP_USE_RWLOCK
3310 flag appropriately. However, this flag is not referenced at all in the
3311 source code, and thus the code does not compile on systems that don't
3312 have the pthread_rwlock_t type (such as Android).
3313
3314 Please find attached a patch (against the current 1.6.x head) that checks
3315 the value of this flag and falls back on using mutexes if read-write
3316 locks are not available.
3317
3318 2010-09-10 Jean Sigwald <jean.sigwald(at)orange-ftgroup.com>
3319 I discovered a reliable denial-of-service issue on the last stable
3320 release of libupnp (1.6.6) remotely triggerable by any
3321 unauthenticated user. The issue is related with a bad parsing of
3322 malformed XML.
3323
3324 2010-09-10 Chandra Penke <chandrapenke(at)mcntech.com>
3325 * SF Patch Tracker [ 2854711 ] Patch for Solaris10 compilation and usage
3326 Submitted By: zephyrus ( zephyrus00jp )
3327
3328 Patch for Solaris10 compilation and usage.
3329
3330 2010-09-10 Chandra Penke <chandrapenke(at)mcntech.com>
3331 Add support for conditionally enabling ipv6.
3332
3333 2010-09-10 Chandra Penke <chandrapenke(at)mcntech.com>
3334 Fix for compilation in debug builds.
3335
3336 Ensure internal methods are declared as static since debug builds don't inline.
3337
3338 2010-09-09 Chandra Penke <chandrapenke(at)mcntech.com>
3339 Fix for regression in SSDP code to send/receive messages over UDP
3340
3341 Sending messages over UDP is broken in some Apple OSes
3342 such as OS X and iOS. This might be broken in other OSes to but didn't
3343 verify.
3344
3345 The fix is to modify the socket length argument of sendto to use the correct
3346 sockaddr length depending on whether the socket is IPV4 or IPV6.
3347
3348 Also added some error checks and debugging related to the issue
3349
3350 2010-09-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3351 Using UpnpReadHttpGet to download large files causes the application to
3352 crash. This happens when the file being downloaded exceeds the device
3353 memory - entirely possible when transferring video files.
3354 The programmatic cause is that the logic implemented in the function
3355 http_ReadHttpGet (which UpnpReadHttpGet calls) reads the entire file
3356 into memory. The fix modifies the existing logic to discard data after
3357 it's been read; there's no reason to keep it around since the caller
3358 of UpnpReadHttpGet already has a copy of it.
3359
3360 This issue exists in 1.6.6 as well as the latest sources.
3361
3362 Patch submitted by Chandra (inactiveneurons).
3363
3364 2010-09-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3365 In the latest sources, http_RequestAndResponse and other methods that
3366 use connect() are broken. More specifically, connect() in these methods
3367 is returning with an EINVAL. The programmatic cause is that the address_len
3368 argument passed to connect() is different in IPV4 vs IPV6 (as described in:
3369 http://www.opengroup.org/onlinepubs/009695399/functions/connect.html).
3370 The current code always uses the IPV6 size. The fix modifies each use of
3371 connect() to use the correct size based on the address family being used.
3372
3373 Patch submitted by Chandra (inactiveneurons).
3374
3375 2010-09-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3376 Fix compilation error in upnp/src/gena/gena_ctrlpt.c (this is most
3377 likely an error on all platforms).
3378
3379 Patch submitted by Chandra (inactiveneurons).
3380
3381 2010-09-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3382 Fix compilation error in upnp/src/inc/ssdplib.h when compiling in OS X
3383 (the netinet/* headers are not available).
3384
3385 Patch submitted by Chandra (inactiveneurons).
3386
3387 2010-09-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3388 Fix compilation error in ixml/inc/ixml.h when compiling with an
3389 Objective-C compiler (when cross-compiling for iPhone devices).
3390
3391 Patch submitted by Chandra (inactiveneurons).
3392
3393 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3394 * Issue regarding the GENA notifications. A string termination indicator
3395 was added at the end of the notification ("\r\n") in
3396 notify_send_and_recv() in upnp/src/gena/gena_device.c.
3397 Patch by Fabrice Fontaine.
3398
3399 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3400 * The last part of Ronan Menard's patch.
3401
3402 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3403 * upnp/src/ssdp/ssdp_device.c: Fix for IPV6 ULA/GUA issues.
3404 * upnp/src/ssdp/ssdp_ctrlpt.c: Fix for IPV6 ULA/GUA issues.
3405 * upnp/src/ssdp/ssdp_server.c: Fix for IPV6 ULA/GUA issues.
3406
3407 Patch submitted by Ronan Menard.
3408
3409 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3410 * upnp/src/genlib/miniserver/miniserver.c: Fix for IPV6 ULA/GUA issues.
3411
3412 Patch submitted by Ronan Menard.
3413
3414 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3415 * gena_subscribe(): Fix for IPV6 ULA/GUA issues.
3416
3417 Patch submitted by Ronan Menard.
3418
3419 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3420 * SOCKET ssdpSock6UlaGua: created variable for later use.
3421
3422 Patch submitted by Ronan Menard.
3423
3424 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3425 * SSDP_IPV6_SITELOCAL: new macro.
3426
3427 Patch submitted by Ronan Menard.
3428
3429 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3430 * The scope of the macro NUM_HANDLE is now restricted to upnpapi.c.
3431
3432 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3433 * InitHandleList() has never been implemented, I guess no one has ever
3434 called it, so remove it.
3435 * GetFreeHandle() and FreeHandle() are now static as they should.
3436
3437 2010-08-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3438 * New internal buffer added to store global/ula IPV6 address.
3439 * Macros to test whether an IPV6 address is global or ula.
3440 * UpnpGetServerUlaGuaIp6Address(): added interface.
3441 * IN6_IS_ADDR_GLOBAL, IN6_IS_ADDR_ULA: new macros.
3442 * gIF_IPV6_ULA_GUA: new buffer.
3443 * UpnpRegisterRootDevice3(): Change to the test of already registered
3444 devices for IPV6.
3445 * UpnpGetIfInfo(): gua/ula issues.
3446
3447 Patch submitted by Ronan Menard.
3448
3449 2010-08-19 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3450 * libUPnP does support IPV6 now.
3451
3452 Patch submitted by Ronan Menard.
3453
3454 2010-08-19 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3455 * HTTP version equal to 1.0 should fail as required by the UPnP
3456 certification tool. Patch submitted by Ronan Menard.
3457
3458 2010-06-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3459 SF Bug Tracker [ 3022490 ] String declaration fix for patch applied in 3007407
3460 Hello,
3461
3462 When my patch for tracker ID 3007407 was accepted, the definition of the
3463 serviceList string was changed from
3464
3465 #define SERVICELIST_STR "serviceList"
3466
3467 to
3468
3469 static const char *SERVICELIST_STR = "serviceList";
3470
3471 During internal code review of the final patch, it was pointed out that
3472 sizeof(SERVICELIST_STR) == 4 since SERVICELIST_STR is now declared as
3473 a pointer instead of an array.
3474
3475 If you wish to use a variable instead of a define, I suggest the
3476 following instead:
3477
3478 static const char SERVICELIST_STR[] = "serviceList";
3479
3480 Thanks,
3481 Chuck Thomason
3482
3483 2010-06-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3484 SF Bug Tracker [ 3007407 ] Service traversal issue in AdvertiseAndReply()
3485 Submitted: Chuck Thomason ( cyt4 ) - 2010-05-26 15:07:39 UTC
3486
3487 When the UPnP server is started, one alive message is broadcast for each
3488 service in each device. It appears that libupnp's implementation of the
3489 alive message generation does not correctly navigate the XML description
3490 document when locating the services. This can result in the wrong UDN
3491 being used in the alive message sent for a service.
3492
3493 In my specific case (see attached XML), the root EchoSTB device contains
3494 no services, but its embedded MediaServer device contains 2 services.
3495 When the existing libupnp code traverses the EchoSTB device in the XML,
3496 it searches the global list of serviceLists within the document instead
3497 of searching for a serviceList that is its direct child node. The
3498 ContentDirectory and ConnectionManager services are then announced with
3499 the UDN of EchoSTB1 (the root device) instead of with the UDN of
3500 MediaServer, which is actually their parent device.
3501
3502 I discovered this behavior using libupnp-1.6.6. I have generated a patch
3503 against branch-1.6.x that corrects the XML navigation such that all
3504 services are traversed from their parent device, which results in the
3505 correct UDN being sent in the alive message for each service. I built
3506 from branch-1.6.x without this patch, tested, and confirmed that the
3507 issue still exists as I observed it in libupnp-1.6.6. I then built
3508 from branch-1.6.x with this patch, tested, and confirmed that the
3509 issue was resolved.
3510
3511 Thanks,
3512 Chuck Thomason
3513
3514 2010-05-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3515 SF Bug Tracker [ 2995758 ] libupnp 1.6.6, wrong bind when reuseaddr is
3516 1.
3517 Submitted: viallard anthony ( homer242 )
3518 When trying to use reuseaddr option in miniserver/miniserver.c, there
3519 isn't a affectation of the port chosen (serverAddr.sin_port isn't
3520 receive listen_port variable value).
3521
3522 2010-04-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3523 Define PROTOTYPES to be one by default in global.h. This affects the
3524 RSA MD5 code.
3525
3526 2010-03-27 Nick Leverton <[email protected]>
3527 Subscription auto-renewals copy the renewal time from old subscription.
3528
3529 2010-03-27 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3530 * Added API to ithread, created the following functions:
3531 - int ithread_initialize_library(void);
3532 - int ithread_cleanup_library(void);
3533 - int ithread_initialize_thread(void);
3534 - int ithread_cleanup_thread(void);
3535 * SF Bug Tracker [ 2876374 ] Access Violation when compiling with Visual Studio 2008
3536 Submitted: Stulle ( stulleamgym ) - 2009-10-10 19:05
3537
3538 Hi,
3539
3540 I am one of the devs of the MorphXT project and I use this lib in some
3541 other of my projects, too. When I tried to upgrade the lib earlier for one
3542 of my projects I had to realise that something did not work at first and
3543 while most of the things were reasonably ease to be fixed. Now, the last
3544 thing I encountered was not so easy to fix and I am uncertain if my fix is
3545 any good so I'll just post it here and wait for some comments.
3546
3547 The problem was that I got an Access Violation when calling "UpnpInit". It
3548 would call "ithread_rwlock_init(&GlobalHndRWLock, NULL)" which eventually
3549 led to calling "pthread_cond_init" and I got the error notice at
3550 "EnterCriticalSection (&ptw32_cond_list_lock);". It appeared that
3551 "ptw32_cond_list_lock" was NULL. Now, I found two ways to fix this. Firstly
3552 moving the whole block after at least one of the "ThreadPoolInit" calls
3553 will fix the issue. Secondly, you could add:
3554 #ifdef WIN32
3555 #ifdef PTW32_STATIC_LIB
3556 // to get the following working we need this... is it a good patch or
3557 not... I do not know!
3558 pthread_win32_process_attach_np();
3559 #endif
3560 #endif
3561 right before "ithread_rwlock_init(&GlobalHndRWLock, NULL)".
3562
3563 Just so you know, I am using libupnp 1.6.6 and libpthreads 2.8.0 and both
3564 are linked static into the binaries. I am currently using Visual Studio
3565 2008 for development with Windows being the target OS. Any comment at your
3566 end?
3567
3568 Regards, Stulle
3569
3570 2010-03-27 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3571 * SF Patch Tracker [ 2836704 ] Patch for Solaris10 compilation and usage.
3572 Submitted By: zephyrus ( zephyrus00jp )
3573 This second part covers the issue on linking with -lsocket -lnsl -lrt.
3574
3575 2010-03-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3576 * SF Bug Tracker [ 2392166 ] ithread_detach not called for finished worker thread
3577 Submitted: Ulrik ( ulsv_enea ) - 2008-12-05 08:24
3578
3579 Valgrind reports a memory leak due to that the function ithread_detach is
3580 not called for finished worker threads in ThreadPool.c.
3581
3582 ==21137== 2,176 bytes in 8 blocks are possibly lost in loss record 5 of 5
3583 ==21137== at 0x4C20F3F: calloc (vg_replace_malloc.c:279)
3584 ==21137== by 0x4010F58: _dl_allocate_tls (in /lib/ld-2.6.1.so)
3585 ==21137== by 0x544BA92: pthread_create@@GLIBC_2.2.5 (in
3586 /lib/libpthread-2.6.1.so)
3587 ==21137== by 0x5F94592: CreateWorker (ThreadPool.c:639)
3588 ==21137== by 0x5F95079: ThreadPoolInit (ThreadPool.c:784)
3589
3590 I'm using libupnp 1.6.6
3591
3592 For more info on pthread_detach, see:
3593 http://gelorakan.wordpress.com/2007/11/26/pthead_create-valgrind-memory-lea
3594 k-solved/
3595
3596 2010-03-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3597 * SF Bug Tracker [ 2392304 ] Memory leak in SSDP AdvertiseAndReply
3598 Submitted: Ulrik ( ulsv_enea ) - 2008-12-05 08:24
3599
3600 Valgrind reports a memory leak function in AdvertiseAndReply
3601 (ssdp/ssdp_server.c) in libupnp 1.6.6
3602
3603 There are continue statements in many places in AdvertiseAndReply. In some
3604 of those error handling cases the variable nodelist is not freed before
3605 continuing to the next iteration. The next iteration will take care of
3606 freeing the nodelist from the previous iteration in most cases, but not
3607 when breaking out of the for loop after the last element.
3608
3609 I believe this memory leak can be solved by making sure that the rows
3610
3611 ixmlNodeList_free( nodeList );
3612 nodeList = NULL;
3613
3614 are always executed, also in the beginning of the last iteration when we
3615 found out that there are not more elements.
3616
3617 ==29110== at 0x4C21C16: malloc (vg_replace_malloc.c:149)
3618 ==29110== by 0x5D8DE0E: ixmlNodeList_addToNodeList (nodeList.c:106)
3619 ==29110== by 0x5D8B7E2: ixmlNode_getElementsByTagNameRecursive
3620 (node.c:1438)
3621 ==29110== by 0x5D8E587: ixmlElement_getElementsByTagName
3622 (element.c:491)
3623 ==29110== by 0x5B6C0F1: AdvertiseAndReply (ssdp_server.c:201)
3624 ==29110== by 0x5B7AB74: UpnpSendAdvertisement (upnpapi.c:1495)
3625
3626 2010-03-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3627 * libupnp and multi-flows scenario patch
3628 Submitted by Carlo Parata from STMicroelectronics.
3629
3630 Hi Roberto and Nektarios,
3631 after an analysis of the problem of libupnp with a multi-flows scenario, I
3632 noticed that the only cause of the freezed system is the ThreadPool
3633 management. There are not mutex problems. In practice, if all threads in the
3634 thread pool are busy executing jobs, a new worker thread should be created if
3635 a job is scheduled (I inspired to tombupnp library). So I solved the problem
3636 with a little patch in threadutil library that you can find attached in this
3637 e-mail. I hope to have helped you.
3638
3639 2010-03-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3640 * SF Patch Tracker [ 2964973 ] install: will not overwrite just-created
3641 ...blah... with...
3642 Submitted: Nick Leverton ( leveret ) - 2010-03-07 05:18
3643
3644 Full error:
3645 /usr/bin/install: will not overwrite just-created
3646 `/tmp/build/libupnp-1.6.6/debian/tmp/usr/share/doc/libupnp3-dev/examples/s
3647 ample_util.c' with `common/sample_util.c'
3648
3649 This seems to be from Automake 1.11 which doesn't like having duplicate
3650 files in a Makefile.am. Patch attached, kindly provided by Stefan Potyra
3651 for Debian (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=543068)
3652
3653 This fix will be needed for both 1.6.x and 1.8.x branches.
3654
3655 2010-03-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3656 * Backport of svn revision 504:
3657 SF Patch Tracker [ 2969188 ] 1.8.0: patch for FreeBSD compilation
3658 Submitted By: Nick Leverton (leveret)
3659 Fix the order of header inclusion for FreeBSD.
3660
3661 2010-03-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3662 * SF Patch Tracker [ 2836704 ] Patch for Solaris10 compilation and usage.
3663 Submitted By: zephyrus ( zephyrus00jp )
3664 Obs by Marcelo: The issue with linking with -lsocket -lnsl -lrt is not
3665 covered in this changeset because I don't have solaris to test. I will
3666 need some help from zephyrus in this regard. The issue will be addressed
3667 in a future changeset.
3668
3669 Compilation for solaris
3670
3671 I have used gcc3.x and gcc4.x under solaris 10 for x86 / 64 bits.
3672
3673 A couple of Source file fixes were necessary for successful compilation
3674 and runtime behavior.
3675
3676 threadutil/src/ThreadPool.c
3677
3678 POSIX
3679 sched_setschduler() returns non-negative value for success.
3680
3681 Without the fix, UpnpInit() fails immediately.
3682
3683 upnp/src/api/upnpapi.c
3684
3685 There is a typo of a macro name "__sun" in one of the
3686 CPP conditional.
3687 Without the fix, the compilation aborts due to unknown constant
3688 in socket ioctl call.
3689
3690 A few structs and an array is not properly initialized.
3691 Well, I think it may be safe as is, but when I checked it
3692 using purify evaluation version, it was reported that
3693 uninitialized iszBuffer may cause read of uninitialized memory.
3694 So play it safe.
3695
3696 Configure issue.
3697 This has to be more of a configure magic.
3698 To link a program successfully using network, we need
3699 -lsocket and -lnsl library specifications on the link line.
3700 We also need -lrt for programs that use thread scheduling features.
3701
3702 The sample program under upnp/sample requires
3703 -lsocket -lnsl -lrt
3704 for successful linking.
3705 I added -lsocket -lnsl -lrt to Makefile.in.
3706 configure probably needs to take care of these.
3707
3708 I don't know much about configure, automake, etc., so
3709 I am just raising a flag here.
3710
3711 TIA
3712
3713 2010-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3714 * SF Patch Tracker [ 2836704 ] Search for nested serviceList (not
3715 stopping at the first lis
3716 Submitted By: zephyrus ( zephyrus00jp )
3717
3718 Internet Gateway Device description contains nested serviceList (rootdevice
3719 -> servicelist, subdevice
3720 and subdevice has the lower-level serviceList, etc..)
3721
3722 Unfortunately, the sample code sample_util.c used by tv_device sample,
3723 etc.
3724 has a code that looks for only the first top-level serviceList.
3725 This results in the failure to read all the services of an IGD xml
3726 description.
3727
3728 Attached patch modifies this behavior and looks for the service by
3729 visiting all the serviceList in xml document in turn.
3730
3731 With the modified patch (ad additional modification), I could
3732 simulate an IGD device and created a modified control program for that.
3733
3734 Patch against 1.6.6
3735
3736 TIA.
3737
3738 2010-03-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3739 * SF Patch Tracker [ 2203721 ] timeb.h check obsolete
3740
3741 2010-03-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3742 * SF Patch Tracker [ 2970872 ] Update ErrorMessages for latest return
3743 code list
3744 Submitted By: Nick Leverton ( leveret )
3745
3746 ErrorMessage[] in upnptools.c has got a bit out of sync, the attached
3747 patch (generated from grep 'define UPNP_E_') should bring it up to date.
3748
3749 2010-03-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3750 * SF Patch Tracker [ 2857611 ] Declare a few functions to have proper
3751 (void) argument list.
3752 Submitted By: zephyrus ( zephyrus00jp )
3753
3754 In a publicly installed headers, a few functions are declared without any
3755 arguments at all, a la "()".
3756 When I used gcc's -Wimplict and -Wstrict-prototypes to check for the
3757 mismatch of
3758 function prototype declarations and their usage in my own program,
3759 some headers from libupnp-1.6.6 produced warnings.
3760
3761 They are not strictly bugs, but pretty much annoying. This is 2009, and
3762 almost all the important compilers
3763 understand ISO-C.
3764
3765 So the offending functions are declared as "(void") to show that they have
3766 no arguments at all.
3767
3768 2010-03-14 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3769 * SF Patch Tracker [ 2546532 ] Missing carriage return between
3770 SOAPACTION and User-Agent headers.
3771
3772 There is something going wrong in soap_ctrlpt.c at line 931 (based on
3773 version 1.6.6 release).
3774
3775 The http_Makemessage call looks as follows:
3776
3777 if (http_MakeMessage(
3778 &request, 1, 1,
3779 "Q" "sbc" "N" "s" "s" "Ucc" "sss",
3780 SOAPMETHOD_POST, path.buf, path.length,
3781 "HOST: ", host.buf, host.length,
3782 content_length,
3783 ContentTypeHeader,
3784 "SOAPACTION:
3785 \"urn:schemas-upnp-org:control-1-0#QueryStateVariable\"",
3786 xml_start, var_name, xml_end ) != 0 ) {
3787 return UPNP_E_OUT_OF_MEMORY;
3788 }
3789
3790 This will result in the SOAPACTION header to be immediately followed by the
3791 User-Agent header, while a cr-lf should separate the two. I propose to fix
3792 this by changing the second "s" to "sc" to force the addition of a cr-lf
3793 after the SOAPACTION. This looks consistent to the other Makemessage calls.
3794
3795 2009-03-06 Oxygenic <oxygenic(at)users.sourceforge.net>
3796 * parameter problem fixed in soap_request_and_response(), 2nd call to
3797 http_RequestAndResponse() was wrong (thanks to Kim Kyungsan)
3798
3799 2008-07-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3800 * Added upnp/m4/libupnp.m4 to the distribution tarball.
3801
3802 2008-07-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3803 * Fixed a missing HandleUnlock() in upnp/src/gena/gena_device.c.
3804
3805 2008-07-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3806 * SF Bug Tracker [ 2026431 ] pupnp does not build on GNU/KfreeBSD.
3807 Submitted By: Nick Leverton - leveret
3808 Gnu/KFreeBSD is one of the Debian architectures, it includes a FreeBSD
3809 kernel with GNU user space (glibc etc). The Gnu/KfreeBSD developers
3810 provided the attached patch to test the appropriate #define and allow pupnp
3811 to build in their environment, and asked me to forward it to you.
3812
3813 Since the test is a simple check for defined(__GLIBC__), this would
3814 presumably also help with other ports of GNU libc to non-Linux kernels.
3815
3816 2008-07-16 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3817 * Andre Sodermans (wienerschnitzel) patch for building libupnp under
3818 windows systems with VC9. This one fixes a missing include.
3819
3820 2008-06-30 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3821 * Added an m4 macro to deal with finding libupnp in the users'
3822 configure script.
3823
3824 2008-06-11 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3825 * Fixed a buffer overflow due to a bug in the calculation of the
3826 CONTENT-TYPE header line size, the length was being calculated with
3827 the wrong string, there was a missing colon.
3828
3829 2008-06-02 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3830 * SF Bug Tracker [ 1942285 ]
3831 UpnpCreatePropertySet can leak memory.
3832 Submitted By: Bob Ciora - bobciora
3833
3834 In file upnp/src/api/upnptools.c, function UpnpCreatePropertySet can leak
3835 memory if no additional arguments are passed. This is because of the
3836 'return' statement at (or near) line 554.
3837
3838 The prior call to ixmlParseBufferEx may succeed. This causes a basic ixml
3839 tree to be created. The return statement at line 554 leaves this tree in
3840 memory without cleaning it up.
3841
3842 There are two options: either add code prior to the return at 554 to clean
3843 up the tree, or simply allow a NumArg parameter of 0 to be passed.
3844
3845 I prefer the second method -- there doesn't seem to be any need to pass
3846 *any* arguments to this function.
3847
3848 In my local copy of upnptools.c, I have simply replaced the "return NULL"
3849 in line 554 to "return PropSet".
3850
3851 I've attached the source file.
3852
3853 2008-05-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3854 * SF Bug Tracker [ 1903069 ]
3855 Subs (not services) not marked 'active'
3856 Submitted By: Bob Ciora - bobciora
3857
3858 If the UpnpAcceptSubscription is not called, the subscription is not marked
3859 as "active", so no state variables will ever be sent.
3860
3861 I have a "lazy" architecture where a service may not be ready to publish
3862 any state data at the time of a subscription. Subscriptions are still
3863 accepted, there's just nothing to send, so UpnpAcceptSubscription is never
3864 called. As a result, the subscription is never marked as "active" via the
3865 genaInitNotify functions.
3866
3867 A best course of action would be to modify UpnpAcceptSubscription<...>
3868 functions so that they can accept *no* initial state information, but can
3869 still result in the subscription being marked as active. Technically,
3870 then, the "active" flag should be set here, not in the genaInitNotify<...>
3871 functions.
3872
3873 But the UpnpAccept functions don't muck with the subscription table, and
3874 it's more work than it's worth to move that code from the gena functions to
3875 the upnpapi functions.
3876
3877 So--- what I've done to correct this problem is to modify both
3878 UpnpAcceptSubscription<...> functions (in upnpapi.c) to accept an empty
3879 state list and still call the gena layer functions. The gena layer
3880 genaInitNotify<...> functions (gena_device.c) then mark the subscription as
3881 "active" *before* checking for an empty state set.
3882
3883 In genaInitNotify, a check for "var_count <= 0" is added immediately after
3884 the "subs->active = 1;" line. If this occurs, then all cleanup is
3885 performed and the function returns GENA_SUCCESS (since now, an empty state
3886 list is not an error). The same check is made for "PropSet == 0" in
3887 genaInitNotifyExt (just after the "subs->active = 1;" line).
3888
3889 I've modified my proxy layer to call UpnpAcceptSubscriptionExt even when
3890 there is no state data to send. With the suggested changes to
3891 gena_device.c, later state changes are sent correctly.
3892
3893 This has solved my problem.
3894
3895 2008-04-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3896 * Fix in function SetSeed() in threadutil/src/ThreadPool.c for CYGWIN
3897 compilation. Thanks to Gary Chan.
3898
3899 *******************************************************************************
3900 Version 1.6.6
3901 *******************************************************************************
3902
3903 2008-04-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3904 * Added thread id's to the UpnpPrintf debug messages. Thanks to
3905 Charles Nepveu for the idea.
3906
3907 2008-04-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3908 * SF Bug Tracker [ 1948586 ]
3909 Uppercase U in in "xmlns:U" in Invoke Action causes seg. f.
3910 Submitted By: Thomas Norheim - kjakan_no
3911 Device no longer segfaults with the following malformed xml action:
3912 <u:SetColor xmlns:U="urn:schemas-upnp-org:service:tvpicture:1">
3913 <Color>2</Color>
3914 </u:SetColor>
3915
3916 2008-04-23 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3917 * Use -O0 in debug builds so that variables do not get optimized out.
3918
3919 2008-04-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3920 * Apostolos Syropoulos changes for OpenSolaris x86.
3921
3922 2008-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3923 * Andre Sodermans (wienerschnitzel) patch for building libupnp under
3924 windows systems with VC9.
3925
3926 2008-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3927 * Andre Sodermans (wienerschnitzel) patch for building libupnp under
3928 windows systems with VC8.
3929
3930 2008-03-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3931 * Fixed a printf format problem on the upnp_tv_device.c from both
3932 upnp/sample/tvdevice and upnp/sample/tvcombo directories. The variable
3933 port was a short int instead of an unsigned short and it was being
3934 print as a negative value.
3935
3936 2008-03-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3937 * SF Bug Tracker [ 1902668 ] Cannot compile on MSVC
3938 Submitted By Luke Kim - nereusuj
3939 Version 1.6.5 cannot be compiled because of some changes in 1.6.3.
3940 MSVC does not support stdint.h, gettimeofday(), sys/param.h, const int
3941 variables in array size and Windows does not define _WINDOWS_ but define
3942 _WINDOWS.
3943 * MSVC does not understand "const int"'s as declarators of array
3944 dimensions, we must use #define'd constants.
3945 * Use WIN32 instead of _WINDOWS_ or _WINDOWS.
3946
3947 2008-02-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3948 * No longer ignore "upnp:rootdevice" advertisement. Thanks to Bob Ciora.
3949
3950 2008-02-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3951 * Changed "sys_errlist[errno]", which is deprecated, by
3952 "strerror_r()", which is thread safe.
3953
3954 2008-02-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3955 * Slightly improved error report by showing the sys_errlist string
3956 corresponding to errno.
3957
3958 2008-02-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3959 * Got rid of two useless constants: UPNP_SOCKETERROR and
3960 UPNP_INVALID_SOCKET. They both mean the same, that a network API
3961 function has failed. -1 is the value to check, not an invented constant.
3962
3963 *******************************************************************************
3964 Version 1.6.5
3965 *******************************************************************************
3966
3967 2008-02-02 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3968 * Peter Hartley's fix for wrong sized variable being passed to
3969 http_MakeMessage() on 64 bit architectures.
3970
3971
3972 *******************************************************************************
3973 Version 1.6.4
3974 *******************************************************************************
3975
3976 2008-01-23 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3977 * Workaround for a problem with the new automake AM_CONDITIONAL macro
3978 from autotools-1.10. Thanks to Ingo Hofmann for helping with debugging
3979 this one.
3980
3981 2008-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3982 * Added quoting to macros AC_CONFIG_AUX_DIR, AC_CONFIG_MACRO_DIR and
3983 AC_CONFIG_SRCDIR in configure.ac. Also changed the name of the
3984 auxiliary directory in AC_CONFIG_AUX_DIR to build-aux.
3985
3986 2008-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3987 * Fix for setsockopt() in Threadpool.c to allow more than one process
3988 to join the multicast-group on OSX. Thanks to Ingo Hofmann.
3989
3990 2008-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3991 * Using defined(__OSX__) || defined(__APPLE__) instead of just
3992 defined(__OSX__) in the code. Thanks to Ingo Hofmann and Chris
3993 Pickel.
3994
3995 2008-01-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3996 * Fix for isFileInVirtualDir. Thanks to Peter Hartley for the patch.
3997
3998 2008-01-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
3999 * Putting back a "defined(__OSX__)" that has been removed in the
4000 previous *BSD patch. Thanks to Chris Pickel for pointing it out.
4001
4002 2008-01-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4003 * SF Patches Tracker [ 1865812 ] typo in docs comment
4004 Submitted By: Hartmut Holzgraefe - hholzgra
4005 typo in docs comment [-302]
4006 Also, the documentation file name was misspelled and was corrected in
4007 the Makefile.am.
4008
4009 *******************************************************************************
4010 Version 1.6.3
4011 *******************************************************************************
4012
4013 2007-12-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4014 * Using pthread flags for the whole project, not just at the places
4015 individually indicated by several Makefile.am files spread all over
4016 the directories. That was too much error prone.
4017
4018 2007-12-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4019 * Added a configure test to check if pthread_rwlock_t is available.
4020 Define _GNU_SOURCE if needed. The fallback behavior will only be
4021 implemented if _GNU_SOURCE proves to be insufficient on some
4022 platforms. Thanks to Jonathan Casiot (no_dice) and Robert Gingher
4023 (robsbox).
4024
4025 2007-12-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4026 * Removed unused iasnprintf.{c,h} files.
4027
4028 2007-12-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4029 * Removed STATSONLY() macro from ThreadPool.{c,h}.
4030 * Removed time() usage from ThreadPool.c.
4031 * Fixed STATS = 0 compilation.
4032
4033 2007-12-16 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4034 * Library was not compiling on FreeBSD 7. Code now no longer uses
4035 ftime(), using gettimeofday() instead. Thanks to Josh Carroll.
4036
4037
4038 *******************************************************************************
4039 Version 1.6.2
4040 *******************************************************************************
4041
4042 2007-12-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4043 * Fixed a compilation error due to a missing #ifdef in
4044 upnp/src/genlib/miniserver/miniserver.c. Thanks to Eugene Christensen.
4045
4046 2007-11-12 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4047 * "make check" was failing because ixml/test/test_document.sh did not
4048 have the executable flag set. Thanks to Steve Bresson.
4049
4050 2007-11-12 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4051 * Fixed a memory leak in upnpapi.c to delete gMiniServerThreadPool in
4052 the call to UpnpFinish(). Thanks to Fabrice Fontaine.
4053
4054 2007-11-09 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4055 * Added a isleep() call to the error handler of select() in
4056 RunMiniServer(), so that it does not take 100% cpu in case select()
4057 fails repeatedly.
4058
4059 *******************************************************************************
4060 Version 1.6.1
4061 *******************************************************************************
4062
4063 2007-11-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4064 * SF Bug Tracker [ 1825278 ] AdvertiseAndReply sleeps with handle lock out
4065 Applied patch from Alex (afaucher) to change some write locks to read
4066 locks.
4067
4068 2007-11-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4069 * Adjusting libtool library numbers to reflect the last changes.
4070
4071 2007-11-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4072 * SF Bug Tracker [ 1825278 ] AdvertiseAndReply sleeps with handle lock out
4073 GlobalHndMutex, which was a mutex is now GlobalHndRWLock, which is a
4074 rwlock. HandleLock() is mapped to HandleWriteLock() while all other
4075 instances have not been checked. One instance in AdvertiseAndReply()
4076 has been changed to HandleReadLock(). Thanks to Alex (afaucher) for the
4077 bug report and suggestions.
4078
4079 2007-11-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4080 * Added support for rwlocks.
4081
4082 2007-11-05 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4083 * SF Bug Tracker [ 1825929 ] worker thread still alive after UpnpFinish()
4084 Submitted By: Luke Kim - nereusuj
4085 Worker thread still alive after calling UpnpFinish() because
4086 ThreadPoolShutdown() is in the #ifdef DEBUG block.
4087 421
4088 422 #ifdef DEBUG
4089 423 ThreadPoolShutdown( &gSendThreadPool );
4090 424 ThreadPoolShutdown( &gRecvThreadPool );
4091
4092 2007-08-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4093 * Changed the calls to virtualDirCallback.open(filename, UPNP_WRITE)
4094 to (virtualDirCallback.open)(filename, UPNP_WRITE) (notice the
4095 parenthesis) due to a change in glibc that produces compilation
4096 errors.
4097
4098 2007-08-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4099 * Initialization of the "randomness" struct so that valgrind does not
4100 complain.
4101
4102 2007-08-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4103 * Merge of patch submitted By Keith Brindley - brindlk
4104 SF Bug Tracker [ 1762758 ] Seek not working for large files
4105 Problem:
4106 Requests from the uPnP client to seek to a position beyond 2GB in a large
4107 file are handled as a request to see from the 2GB point.
4108
4109 Impact:
4110 Varies depending on client. The Xbox 360 kills the connection when it
4111 realizes.
4112
4113 Solution:
4114 GetNextRange function (webserver.c) is updated to handle large file sizes.
4115 Fix should also recognize when built on a 32bit platform rather than 64 and
4116 handle accordingly.
4117
4118 2007-08-05 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4119 * Merge of Mac OS X patch from Stéphane Corthésy (davelopper),
4120 SF Bug Tracker [ 1686420 ] Modifications for MacOSX.
4121 Some of the proposed changes were already done by Rene Hexel's patch.
4122
4123 *******************************************************************************
4124 Version 1.6.0
4125 *******************************************************************************
4126
4127 2007-06-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4128 * More fixes to Mac OS X and NetBSD from Rene Hexel:
4129 [pupnp-devel] NetBSD & Mac OS X packages and patches
4130
4131 Okay, I found a couple more things. I have attached a patch file
4132 against the trunk (version 206) that make the repository code compile
4133 and run on both Mac OS X and NetBSD.
4134
4135 This fixes the following issues:
4136
4137 upnp/src/api/upnpapi.c: SIOCGIFCONF didn't work properly, use
4138 getifaddrs() instead (on BSD systems).
4139
4140 threadutil/src/ThreadPool.c: priorities only work if
4141 _POSIX_PRIORITY_SCHEDULING is defined (and greater than 0).
4142
4143 threadutil/src/LinkedList.c and threadutil/src/iasnprintf.c: use
4144 stdlib.h instead of malloc.h on all BSD systems (not just FreeBSD).
4145 This is important, because malloc.h does not exist on Darwin/Mac OS X.
4146
4147 Cheers
4148 ,
4149 Rene
4150
4151 2007-06-09 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4152 * [pupnp-devel] NetBSD & Mac OS X packages and patches.
4153 Rene Hexel's <[email protected]> patch to compile in NetBSD and Mac OS X.
4154
4155 2007-05-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4156 * Updated the macro files acx_pthread.m4, ax_cflags_gcc_option.m4,
4157 ax_cflags_warn_all.m4, m4/ax_cflags_warn_all_ansi.m4,
4158 m4/type_socklen_t.m4.
4159
4160 2007-05-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4161 * Fixed an issue with the installation of the file upnpdebug.h. Since
4162 the last modifications that removed the macro DEBUG_ONLY, this file
4163 must be installed even on a non-debug build.
4164
4165 2007-05-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4166 * SF Bug Tracker [ 1711325 ] Bad DestAddr in Upnp_Discovery structure
4167 Submitted By: Bob Ciora
4168 The field DestAddr of the structure Upnp_Discovery is now a full
4169 SOCKADDRIN instead of a pointer to SOCKADDRIN. Commented code suggests
4170 that in a previous moment, the function ssdp_handle_ctrlpt_msg() did
4171 not use a postponed thread to call ctrlpt_callback(). Now the code
4172 uses a thread, and most probably the original data would get lost and
4173 the pointer would point to an invalid memory region. This fix caused
4174 an interface change in the library and the minor library version was
4175 bumped. Also, the libtool library numbers were changed accordingly.
4176
4177 *******************************************************************************
4178 Version 1.4.7
4179 *******************************************************************************
4180
4181 2007-05-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4182 * Added support for the Basic Device
4183 (http://www.upnp.org/standardizeddcps/basic.asp) as suggested by
4184 Titus Winters.
4185
4186 2007-05-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4187 * Fixed the file libupnp.pc.in to generate a correct path for the
4188 include files.
4189
4190 2007-05-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4191 * Removing the Dbg_Level, InitLog, SetLogFileNames and CloseLog
4192 defines. These were just aliases, no reason to keep them.
4193
4194 2007-05-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4195 * Changed the comments of the include files that expose the UPnP API
4196 to use only C89 comments and no C99 comments.
4197
4198 2007-05-24 Nektarios K. Papadopoulos <npapadop(at)inaccessnetworks.com>
4199 * Added tvcombo sample that demonstrates coexistence of a device and a
4200 control point in the same application.
4201
4202 2007-05-24 Nektarios K. Papadopoulos <npapadop(at)inaccessnetworks.com>
4203 * SF Tracker FR [ 1570020 ].
4204 * Enable both device and control point in the same application. Resolve
4205 deadlock in the SSDP processing threads.
4206 * Fix Threadpool expansion condition.
4207 Thanks to Siva Chandran P. for the original patch.
4208
4209 2007-05-24 Nektarios K. Papadopoulos <npapadop(at)inaccessnetworks.com>
4210 * Modified tvdevice (control and picture) service descriptions to make
4211 compatible with WinXP/IE control point. 'in' arguments must appear before
4212 'out' arguments in argument list.
4213 Thanks to Martin Tremblay for pointing out the solution originally provided
4214 by MORIOKA Yasuhiro.
4215
4216 2007-05-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4217 * More MSVC fixes, using XINLINE instead of inline, MSVC has troubles
4218 with inline. Thanks to David Maass for reporting.
4219 * Changed XINLINE to UPNP_INLINE.
4220
4221 2007-05-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4222 * Added the file build/inc/msvc/inttypes.h. This file is for use with
4223 MSVC only, because it does not provide C99 compatibility.
4224
4225 2007-05-18 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4226 * Removed all uses of the DEVICEONLY(x) macro.
4227
4228 2007-05-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4229 * Removed all uses of the DBGONLY(x) macro. A static inline empty
4230 function now is used and the compiler takes care of optimizing it out.
4231
4232 2007-05-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4233 * Fixed a bug in UpnpPrintf, function could call va_start() and return
4234 before calling va_end().
4235
4236 2007-05-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4237 * EXPORT_SPEC missing on some declarations in ixml/inc/ixml.h.
4238 Thanks to David Maass.
4239
4240 2007-05-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4241 * sizeof is unsigned, so %zu is more adequate than %zd.
4242
4243 2007-05-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4244 * Using an invented printf directive PRIzu that on MSVC
4245 expands to "lu", and on normal C99 compilers expands to "zu".
4246
4247 2007-05-15 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4248 * Rewrote raw_find_str. Now it no longer uses strcasestr(), but it
4249 transforms the first input buffer into lowercase.
4250
4251 2007-05-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4252 * Fix for debug printf format strings. size_t are not expected
4253 in a string format like "%.*s".
4254
4255 2007-05-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4256 * Added inttypes.h as a header requirement in configure.ac.
4257
4258 2007-05-11 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4259 * Moved upnp_tv_ctrlpt and upnp_tv_device executables from folder
4260 upnp to folder upnp/sample. Moved folder upnp/sample/tvdevice/web
4261 to folder upnp/sample/web. This way, if someone compiles the
4262 tarball and executes upnp_tv_device from its creation directory,
4263 there will be no error -108 for not finding directory web.
4264
4265 *******************************************************************************
4266 Version 1.4.6
4267 *******************************************************************************
4268
4269 2007-04-19 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4270 * Fix for freebsd host_os in configure.ac.
4271
4272 *******************************************************************************
4273 Version 1.4.5
4274 *******************************************************************************
4275
4276 2007-04-19 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4277 * Case insensitive comparison in raw_find_str() (httpparser.c) as
4278 suggested by Craig Nelson in SF Tracker [ 1689382 ] DLINK DIR-625
4279 patch.
4280
4281 2007-04-07 Nektarios K. Papadopoulos <npapadop(at)inaccessnetworks.com>
4282 * Fix for a bug in makeAction where va_arg was being called one
4283 extra time.
4284
4285 2007-04-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4286 * SF Tracker [ 1703533 ] Patch to make it compile under FreeBSD
4287 Submitted By: Timothy Redaelli - drittz
4288 I made some patches to make it compile under FreeBSD using
4289 gethostbyaddr_r when supported.
4290
4291 2007-04-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4292 * [pupnp-devel] Type mixup on x86_64 causes UPNP_E_OUT_OF_MEMORY
4293 Submitted By: Glen Masgai
4294 after an UpnpSendActionAsync() for example, i get UPNP_E_OUT_OF_MEMORY
4295 in the callback using 1.4.4 on a x86_64 system. This happens in
4296 http_MakeMessage(), which in some cases get called with wrong types
4297 (int instead of size_t) in combination with format "b" and "Q".
4298 The attached patch should fix this.
4299
4300 *******************************************************************************
4301 Version 1.4.4
4302 *******************************************************************************
4303
4304 2007-04-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4305 * SF Tracker [ 1695399 ] Typo in util.h
4306 Submitted By: Luke Kim - nereusuj
4307 Unix sleep is in seconds but WIN32 Sleep is in milliseconds.
4308
4309 2007-04-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4310 * SF Tracker [ 1652827 ] UpnpRegisterRootDevice returned -104
4311 Submitted By: Michael Andersen - miwer
4312
4313 Issue was found to be related to sizeof (size_t) != sizeof (int)
4314 on AMD64 systems. Emil Ljungdahl's AMD64 patch has been applied along
4315 with some other fixes. Original user report follows:
4316
4317 When I run upnpd I get the above mentioned error (UPNP_E_OUT_OF_MEMORY).
4318 I've tried with 1.4.1 and 1.4.2-RC3, it's the same. I don't understand why,
4319 because I have plenty of RAM, and I even tried closing some applications,
4320 but it didn't help.
4321
4322 $ upnpd eth1 br0
4323
4324 The following is logged in the /var/log/messages:
4325 Feb 6 01:33:47 server upnpd[6933]: Error registering the root device with
4326 descDocUrl: http://192.168.0.1:49152/gatedesc.xml
4327 Feb 6 01:33:47 server upnpd[6933]: UpnpRegisterRootDevice returned -104
4328
4329
4330 I tried enabling debugging and it looks like it cannot allocate memory
4331 through the membuffer_append function. It's weird because it's only a few
4332 bytes.
4333
4334 Please note, that I enabled some extra debugging lines that were commented,
4335 in order to get more information. See attached files.
4336
4337 *******************************************************************************
4338 Version 1.4.3
4339 *******************************************************************************
4340
4341 2007-03-13 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4342 * SF Tracker [ 1663004 ] Compile on Cygwin
4343 Submitted By: Jon Foster - jongfoster
4344 This patch gives basic support for building under Cygwin - it compiles,
4345 links, and a simple UPnP device application can initialize. I'm not sure
4346 if it actually works yet, but this is definitely a step in the right
4347 direction.
4348
4349 Patch is against the 1.4.1 release. Changes are:
4350
4351 * threadutil/inc/ithread.h: Fix the ithread mutex support to use
4352 documented, portable APIs (if present) rather than the Non-Portable (_NP)
4353 ones it uses now. This is required because Cygwin implements only the
4354 portable API.
4355
4356 * threadutil/src/ThreadPool.c: Fake SetPolicyType() to do nothing on Cygwin
4357 because otherwise it fails. Should probably investigate why it fails and
4358 add a proper implementation later.
4359
4360 * upnp/src/api/upnpapi.c: On Cygwin, zero out the GlobalHndMutex structure
4361 before initializing it. Without this, the initialization fails. This
4362 appears to be a bug in Cygwin.
4363
4364 * upnp/src/genlib/net/uri/uri.c: Use gethostbyname() on Cygwin.
4365
4366 2007-03-05 Oxy <oxygenic(at)users.sourceforge.net>
4367 * Code adapted and typedefs added to compile cleanly under Windows
4368 with Borland C++ Builder and MS Visual C++
4369
4370 2007-03-03 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4371
4372 * Fixed nasty segmentation fault bug on membuffer.c.
4373
4374 *******************************************************************************
4375 Version 1.4.2
4376 *******************************************************************************
4377
4378 2007-02-09 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4379
4380 * 32/64 bits portability issues on *printf.
4381 Use %zd for size_t, and cast to (long long) for off_t.
4382
4383 2007-02-02 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4384
4385 * Bumped the program version to 1.4.2 in config.ac.
4386
4387 * Now requires autoconf 2.60.
4388
4389 * Fixed docdir use.
4390
4391 * Does not install the documentation by default.
4392
4393 * Use dist-bzip2 to create a .bz2 distribution file.
4394
4395 2007-01-23 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4396
4397 * SF Tracker [ 1634922 ] Support for large files (>= 2 GiB), part 2
4398 Submitted By: Jonathan Casiot - no_dice
4399 Summary: This patch hopefully fixes the remaining types and related
4400 code to enable files >= 2 GiB to be streamed. Jonathan claims to have
4401 tested this with a patched version of ushare-0.9.8 and a D-Link DSM-520.
4402
4403 2007-01-09 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4404
4405 * SF Tracker [ 1628629 ] Multicast interface patch
4406 Submitted By: Fredrik Svensson - svefredrik
4407 This patch fixes two problems:
4408 1) Specify the IP address for the interface when we do
4409 setsockopt IP_ADD_MEMBERSHIP. This makes it possible to run
4410 when no default router has been configured.
4411 2) Explicitly set the multicast interface through setsockopt
4412 IP_MULTICAST_IF. Avoids socket error -207 in some cases.
4413
4414 * SF Tracker [ 1628590 ] XML parsing segfault patch
4415 Submitted By: Fredrik Svensson - svefredrik
4416 This patch fixes a segmentation fault problem that occurs
4417 when parsing XML code than some routers produce.
4418
4419 2007-01-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4420
4421 * SF Tracker [ 1628552 ] XML white space patch
4422 Submitted By: Fredrik Svensson - svefredrik
4423
4424 * SF Tracker [ 1628562 ] Maximum total jobs patch
4425 Submitted By: Fredrik Svensson - svefredrik
4426 Also, I incremented the library versions and included some
4427 comments in the file configure.ac so that we do not bump
4428 the library version excessively, only the necessary numbers
4429 on the next release.
4430
4431 * SF Tracker [ 1628575 ] Linksys WRT54G patch
4432 Submitted By: Fredrik Svensson - svefredrik
4433
4434 * SF Tracker [ 1628636 ] SSDP packet copy patch
4435 Submitted By: Fredrik Svensson - svefredrik
4436 Changed NUM_COPY to 1 since, according to section 9.2 of the
4437 HTTPU/MU spec, we should never send more than one copy of a
4438 reply to an SSDP request. Ref. section 9.2 of
4439 http://www.upnp.org/download/draft-goland-http-udp-04.txt
4440
4441 2006-12-23 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4442
4443 * Thorough revision of every call of http_MakeMessage() due to a
4444 bug introduced in rev.79 "largefile patch added".
4445 http_MakeMessage() has a worst than brain damaged "printf" like
4446 interface. In rev.79, the "N" format parameter must be an off_t.
4447 Every call of this function with an "N" format parameter and an
4448 int passed on the stack would fail terribly.
4449
4450 * SF Bug tracker [ 1590469 ]
4451 Typo in ixmlparser.c
4452 Submitted By: Erik Johansson - erijo
4453
4454 * SF Bug Tracker [ 1590466 ] Invalid xml output
4455 Submitted By: Erik Johansson - erijo
4456
4457 * SF Patch tracker [ 1581161 ] VStudio2005 patch
4458 Submitted By: David Maass - darkservant
4459
4460 * SF Patch tracker [ 1587272 ] constified ixml
4461 Submitted By: Erik Johansson
4462
4463 * Finished constifications as suggested by Erik Johansson in
4464 SF Patch tracker [ 1587272 ].
4465
4466 2006-07-05 Nektarios K. Papadopoulos <npapadop(at)inaccessnetworks.com>
4467 * [bug-id] 1580440
4468 [submitted-by] Erik Johansson - erijo
4469 [patched-by] Erik Johansson - erijo
4470 The SOAP HTTP message that's generated on upnp errors
4471 is missing a \r\n\ between header and body.
4472
4473 2006-07-07 Oxy <virtual_worlds(at)gmx.de>
4474
4475 * support for large files (>2 GBytes) added
4476
4477
4478 *******************************************************************************
4479 Version 1.4.1
4480 *******************************************************************************
4481
4482 2006-07-07 Oxy <virtual_worlds(at)gmx.de>
4483
4484 * full support for Windows added, static library and DLL are fully
4485 working, code compiles with Borland Builder C++ and MS Visual
4486 C/C++
4487
4488 2006-07-05 Nektarios K. Papadopoulos <npapadop(at)inaccessnetworks.com>
4489
4490 * Include prebuilt documentation (html,pdf), dropping doc++
4491 dependency.
4492
4493 2006-07-03 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4494
4495 * Patch for FreeBSD, thanks to Markus Strobl.
4496
4497 2006-06-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4498
4499 * Fix for missing "else" in httpreadwrite.c. Thanks to npapadop
4500 for the patch.
4501
4502 2006-06-26 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
4503
4504 * Fix for va_list initialization in x86_64 architectures.
4505
4506 2006-06-08 Oxy <virtual_worlds(at)gmx.de>
4507
4508 * Patch to fix memory leaks and reasons for crashes added (thanks
4509 to loigu)
4510
4511 *******************************************************************************
4512 Version 1.4.0
4513 *******************************************************************************
4514
4515 2006-05-26 Oxy <virtual_worlds(at)gmx.de>
4516
4517 * defines in iasnprintf.h changed to work with GCC-version < 3
4518
4519 2006-05-22 Oxy <virtual_worlds(at)gmx.de>
4520
4521 * BSD-patch added (not tested yet on an BSD system)
4522
4523 2006-05-19 Oxy <virtual_worlds(at)gmx.de>
4524
4525 * Patch added for bug: ixml parser collapsed on empty args (arg="")
4526
4527 2006-05-18 Oxy <virtual_worlds(at)gmx.de>
4528
4529 * DSM-320 patch added (fetched from project MediaTomb)
4530
4531 * httpGet additions added, Added proxy support by introducing
4532 UpnpOpenHttpGetProxy. UpnpOpenHttpGet now just calls
4533 UpnpOpenHttpGetProxy with the proxy url set to NULL.
4534
4535 * Bugfix for typo ("\0" / "0") in ixmlparser.c
4536
4537 * Bugfix for M-Search packet
4538
4539 *******************************************************************************
4540 FORK FROM DEAD libupnp
4541 *******************************************************************************
4542
4543 2006-04-29 Rémi Turboult <r3mi(at)users.sourceforge.net>
4544
4545 * THANKS: new file with list of contributors
4546
4547 * upnp/src/gena/gena_device.c (respond_ok): add 'Content-Length: 0'
4548 in subscription response. Patch by Chaos (Bug # 1455367).
4549
4550 2006-04-08 Rémi Turboult <r3mi(at)users.sourceforge.net>
4551
4552 * upnp/doc/UPnP_Programming_Guide.pdf: replace this document with
4553 the one in libupnp-doc-1.2.1 because current CVS version
4554 was corrupted.
4555
4556 2006-04-06
4557
4558 * changes applied to several files to work under Sparc Solaris, temporarily
4559 requiring a define SPARC_SOLARIS
4560
4561 2006-04-03 Rémi Turboult <r3mi(at)users.sourceforge.net>
4562
4563 * upnp/Makefile.am: install upnp samples in $(docdir)/examples
4564
4565 2006-03-28 Rémi Turboult <r3mi(at)users.sourceforge.net>
4566
4567 * configure.ac: add --with-docdir option to choose where documentation
4568 is installed (or -without-docdir to not install the documentation)
4569
4570 2006-03-27 Rémi Turboult <r3mi(at)users.sourceforge.net>
4571
4572 * ixml/test: add simple test suite for xml parser
4573
4574 2006-03-26 Rémi Turboult <r3mi(at)users.sourceforge.net>
4575
4576 * ixml/src/ixmlparser.c (Parser_processCDSect): fix bug which prevents
4577 CDATA sections which contain a 0 (zero) to be parsed (instead the
4578 parsing of the whole document is aborted). Patch by Arno Willig
4579 (Patch # 1432124).
4580
4581 * configure.ac, upnp/Makefile.am: add "--disable-samples" configure
4582 option, and move samples compilation from check_PROGRAMS to
4583 noinst_PROGRAMS
4584
4585 2006-03-25 Rémi Turboult <r3mi(at)users.sourceforge.net>
4586
4587 * upnp/src/genlib/miniserver/miniserver.c (get_miniserver_sockets):
4588 fix bug if new socket created has fd 0 (can only occur when stdin
4589 has been closed). Patch by Oskar Liljeblad 2004-07-02 :
4590 http://sourceforge.net/mailarchive/message.php?msg_id=8870528
4591
4592 2006-03-21 Rémi Turboult <r3mi(at)users.sourceforge.net>
4593
4594 * upnp/test/test_init.c: add some version checks and exit if failure
4595
4596 2006-03-05 Rémi Turboult <r3mi(at)users.sourceforge.net>
4597
4598 * libupnp version 1.3.1
4599
4600 * upnp/inc/upnpconfig.h.in: add new define UPNP_VERSION_PATCH
4601
4602 * upnp/test/test_init.c: add simple test to run during checks
4603
4604 * upnp/inc/upnp.h: include "upnpdebug.h" only if debug enabled
4605 in the library (else header file is not installed)
4606
4607 * upnp/Makefile.am (libupnp_la_LDFLAGS): add inter-library libtool
4608 dependencies between upnp and ixml / threadutil, so that programs
4609 linking against upnp only still work.
4610
4611 2006-03-04 Rémi Turboult <r3mi(at)users.sourceforge.net>
4612
4613 * libupnp version 1.3.0
4614
4615 2006-03-03 Rémi Turboult <r3mi(at)users.sourceforge.net>
4616
4617 * upnp/src/genlib/net/http/httpreadwrite.c (get_sdk_info): use
4618 package version string from configure to set sdk info
4619
4620 * upnp/Makefile.am: add sample/tvdevice/web/ files in EXTRA_DIST
4621 + do not distribute generated upnpconfig.h file.
4622
4623 2006-02-28 Rémi Turboult <r3mi(at)users.sourceforge.net>
4624
4625 * upnp/src/inc/config.h, configure.ac: use only new defines
4626 UPNP_HAVE_xx instead of INCLUDE_yyy_APIS and INTERNAL_WEB_SERVER
4627
4628 * upnp/Makefile.am, ixml/Makefile.am: add -export-symbols-regex to
4629 the library LDFLAGS in order to export only the symbols defined
4630 in the API
4631
4632 2006-02-27 Rémi Turboult <r3mi(at)users.sourceforge.net>
4633
4634 * configure.ac: add libtool versions for the 3 libraries
4635
4636 * ixml/src/ixml.c (copy_with_escape): add missing 'static' to function
4637
4638 * threadutil/src/ThreadPool.c (SetSeed): add missing 'static'
4639
4640 2006-02-26 Rémi Turboult <r3mi(at)users.sourceforge.net>
4641
4642 * threadutil/inc/iasnprintf.h: add gcc __printf__ format attribute
4643 to "iasnprintf"
4644
4645 * upnp/src/api/upnpapi.c: fix invalid UpnpPrintf formats
4646
4647 * upnp/src/gena/gena_device.c: fix invalid UpnpPrintf formats
4648
4649 * upnp/src/inc/config.h: move upnp/inc/config.h to internal
4650 sources (this file is no longer installed with the libraries)
4651
4652 * upnp/inc/upnpdebug.h: new file created from debug definitions
4653 previously in upnp/inc/config.h
4654
4655 * upnp/src/api/config.c: rename to upnp/src/api/upnpdebug.c
4656
4657 * upnp/inc/upnpconfig.h.in: new file to contain information on
4658 the configuration of the installed libraries (generates installed
4659 file <upnp/upnpconfig.h>)
4660
4661 2006-02-22 Rémi Turboult <r3mi(at)users.sourceforge.net>
4662
4663 * upnp/ : add missing include of config.h in some .c files
4664
4665 2006-02-21 Rémi Turboult <r3mi(at)users.sourceforge.net>
4666
4667 * upnp/inc/upnp.h: move some definitions which should not be
4668 exported into "upnp/src/inc/util.h"
4669
4670 * import all modifications below from libupnp in djmount 0.51
4671 into official libupnp
4672
4673 2006-01-17 Rémi Turboult <r3mi(at)users.sourceforge.net>
4674
4675 * threadutil/Makefile.am (libthreadutil_la_SOURCES): remove extraneous
4676 file
4677
4678 2006-01-15 Rémi Turboult <r3mi(at)users.sourceforge.net>
4679
4680 * configure.ac: add checks for large-file support
4681
4682 * upnp/inc/config.h: rename to "upnpconfig.h". The new "config.h" file
4683 is the one generated by autoconf.
4684
4685 * m4/type_socklen_t.m4: added new check for socklen_t (fallback to
4686 int if not defined)
4687
4688 * upnp/src/genlib/miniserver/miniserver.c,
4689 upnp/src/ssdp/ssdp_server.c: use socklen_t where appropriate
4690 (instead of int)
4691
4692 * upnp/src/genlib/net/http/httpreadwrite.c (get_sdk_info): remove
4693 XINLINE declaration (unused and too late)
4694
4695 * ixml/src/node.c (ixmlNode_getNodeType): fix compilation warning
4696 on const return type
4697
4698 2006-01-12 Rémi Turboult <r3mi(at)users.sourceforge.net>
4699
4700 * upnp/src/inc/readwrite.h : suppress unused C++ header file
4701
4702 2006-01-11 Rémi Turboult <r3mi(at)users.sourceforge.net>
4703
4704 * upnp/inc/config.h, upnp/src/inc/upnpapi.h,
4705 upnp/src/inc/httpreadwrite.h: remove internal configuration variable
4706 MINIMUM_DELAY (no clear purpose)
4707
4708 2005-12-05 Rémi Turboult <r3mi(at)users.sourceforge.net>
4709
4710 * upnp/inc/upnp.h: re-declare obsolete method UpnpSetContentLength,
4711 for binary compatibility with previous libupnp version.
4712
4713 * upnp/src/api/upnpapi.c: correct type of g_maxContentLength to size_t
4714
4715 2005-11-01 Rémi Turboult <r3mi(at)users.sourceforge.net>
4716
4717 * autoconfiscate library : replace all makefiles by Makefile.am
4718 for automake support, + preliminary autoconf support
4719 (generated config.h not yet used in source files)
4720
4721 2005-10-18 Rémi Turboult <r3mi(at)users.sourceforge.net>
4722
4723 * upnp/src/makefile: fix location of DEBUG STATIC libupnp library
4724
4725 * upnp/sample/tvctrlpt/linux/Makefile,
4726 upnp/sample/tvdevice/linux/Makefile: fix STATIC library support
4727
4728 2005-10-16 Rémi Turboult <r3mi(at)users.sourceforge.net>
4729
4730 * threadutil/src/Makefile (clean): remove built library
4731
4732 2005-08-28 Rémi Turboult <r3mi(at)users.sourceforge.net>
4733
4734 * ixml/src/ixml.h, ixml/src/ixml.c (ixmlRelaxParser) : new function
4735
4736 * ixml/src/ixmlparser.h, ixml/src/ixmlparser.c (Parser_setErrorChar) :
4737 new function
4738
4739 2005-08-02 Rémi Turboult <r3mi(at)users.sourceforge.net>
4740
4741 * ixml/src/Makefile: correct bug for static library being incorrectly
4742 stripped when building non-debug
4743
4744 2005-06-09 Rémi Turboult <r3mi(at)users.sourceforge.net>
4745
4746 * ixml/src/element.c (ixmlElement_removeAttributeNode):
4747 remove some compilation warning
4748
4749 * ixml/inc/ixml.h, ixml/src/document.c :
4750 add some missing const's in public API
4751
4752 * upnp/inc/upnptools.h, upnp/src/api/upnptools.c :
4753 add missing const's in public API
4754
4755 2005-05-28 Rémi Turboult <r3mi(at)users.sourceforge.net>
4756
4757 * upnp/inc/config.h: suppress HTTP_READ_BYTES (unused)
4758 and replace by DEFAULT_SOAP_CONTENT_LENGTH (previously in upnpapi.h)
4759
4760 * upnp/inc/upnp.h, upnp/src/api/upnpapi.c : replace
4761 UpnpSetContentLength (which was not using its Handle argument)
4762 by global function UpnpSetMaxContentLength.
4763 Remove "hard" limitation to 32K (not suitable for using in UPnP AV).
4764
4765 * upnp/src/inc/upnpapi.h : removed DEFAULT_SOAP_CONTENT_LENGTH
4766 (moved to config.h) and MAX_SOAP_CONTENT_LENGTH (now unused)
4767
4768 * upnp/src/api/upnptools.c : add more error message strings
4769
4770 * upnp/src/genlib/net/http/httpreadwrite.c : return OUT_OF_BOUNDS
4771 instead of BAD_HTTPMSG when allowed Content Length is exceeded.
4772
4773 * upnp/src/genlib/net/http/httpreadwrite.c : corrected an incorrect
4774 sprintf format
4775
4776 2005-05-27 Rémi Turboult <r3mi(at)users.sourceforge.net>
4777
4778 * upnp/makefile, upnp/src/makefile,
4779 ixml/Makefile, ixml/src/Makefile,
4780 threadutil/Makefile, threadutil/src/Makefile :
4781 implement STATIC library support (from patch at
4782 http://sourceforge.net/tracker/?group_id=7189&atid=307189 )
4783
4784 2005-05-26 Rémi Turboult <r3mi(at)users.sourceforge.net>
4785
4786 * upnp/src/api/upnpapi.c, upnp/src/soap/soap_device.c,
4787 upnp/src/soap/makefile :
4788 corrections for compilation with CLIENT=1 only
4789
4790 * importing "libupnp-1.2.1a" as baseline
4791
4792 --------------------------------------------------------------------------------
4793
4794 History of configure.ac:
4795
4796 dnl # Release 1.4.1:
4797 dnl #AC_SUBST([LT_VERSION_IXML], [2:2:0])
4798 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [2:2:0])
4799 dnl #AC_SUBST([LT_VERSION_UPNP], [2:2:0])
4800 dnl #
4801 dnl ############################################################################
4802 dnl # Release 1.4.6:
4803 dnl # "current:revision:age"
4804 dnl #
4805 dnl # - Code has changed in ixml
4806 dnl # revision: 2 -> 3
4807 dnl # - Code has changed in threadutil
4808 dnl # revision: 2 -> 3
4809 dnl # - Interface added in threadutil
4810 dnl # current: 2 -> 3
4811 dnl # revision: 3 -> 0
4812 dnl # age: 0 -> 1
4813 dnl # - Code has changed in upnp
4814 dnl # revision: 2 -> 3
4815 dnl #
4816 dnl #AC_SUBST([LT_VERSION_IXML], [2:3:0])
4817 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [3:0:1])
4818 dnl #AC_SUBST([LT_VERSION_UPNP], [2:3:0])
4819 dnl #
4820 dnl ############################################################################
4821 dnl # Release 1.6.0:
4822 dnl # "current:revision:age"
4823 dnl #
4824 dnl # - Code has changed in ixml
4825 dnl # revision: 3 -> 4
4826 dnl # - Code has changed in threadutil
4827 dnl # revision: 0 -> 1
4828 dnl # - Code has changed in upnp
4829 dnl # revision: 3 -> 4
4830 dnl # - Interface changed in upnp
4831 dnl # current: 2 -> 3
4832 dnl # revision: 4 -> 0
4833 dnl # - Interface removed in upnp
4834 dnl # age: 0 -> 0
4835 dnl #
4836 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4837 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [3:1:1])
4838 dnl #AC_SUBST([LT_VERSION_UPNP], [3:0:0])
4839 dnl #
4840 dnl ############################################################################
4841 dnl # Release 1.6.1:
4842 dnl # "current:revision:age"
4843 dnl #
4844 dnl # - Code has changed in threadutil
4845 dnl # revision: 1 -> 2
4846 dnl # - Interface added in threadutil
4847 dnl # current: 3 -> 4
4848 dnl # revision: 2 -> 0
4849 dnl # - Interface added in threadutil
4850 dnl # age: 1 -> 2
4851 dnl # - Code has changed in upnp
4852 dnl # revision: 0 -> 1
4853 dnl #
4854 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4855 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [4:0:2])
4856 dnl #AC_SUBST([LT_VERSION_UPNP], [3:1:0])
4857 dnl #
4858 dnl ############################################################################
4859 dnl # Release 1.6.2:
4860 dnl # "current:revision:age"
4861 dnl #
4862 dnl # - Code has changed in upnp
4863 dnl # revision: 1 -> 2
4864 dnl #
4865 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4866 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [4:0:2])
4867 dnl #AC_SUBST([LT_VERSION_UPNP], [3:2:0])
4868 dnl #
4869 dnl ############################################################################
4870 dnl # Release 1.6.3:
4871 dnl # "current:revision:age"
4872 dnl #
4873 dnl # - Code has changed in threadutil
4874 dnl # revision: 0 -> 1
4875 dnl #
4876 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4877 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [4:1:2])
4878 dnl #AC_SUBST([LT_VERSION_UPNP], [3:2:0])
4879 dnl #
4880 dnl ############################################################################
4881 dnl # Release 1.6.4:
4882 dnl # "current:revision:age"
4883 dnl #
4884 dnl # - Code has changed in threadutil
4885 dnl # revision: 1 -> 2
4886 dnl # - Code has changed in upnp
4887 dnl # revision: 2 -> 3
4888 dnl #
4889 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4890 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [4:2:2])
4891 dnl #AC_SUBST([LT_VERSION_UPNP], [3:3:0])
4892 dnl #
4893 dnl ############################################################################
4894 dnl # Release 1.6.5:
4895 dnl # "current:revision:age"
4896 dnl #
4897 dnl # - Code has changed in upnp
4898 dnl # revision: 3 -> 4
4899 dnl #
4900 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4901 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [4:2:2])
4902 dnl #AC_SUBST([LT_VERSION_UPNP], [3:4:0])
4903 dnl #
4904 dnl ############################################################################
4905 dnl # Release 1.6.6:
4906 dnl # "current:revision:age"
4907 dnl #
4908 dnl # - Code has changed in threadutil
4909 dnl # revision: 2 -> 3
4910 dnl # - Code has changed in upnp
4911 dnl # revision: 4 -> 5
4912 dnl #
4913 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4914 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [4:3:2])
4915 dnl #AC_SUBST([LT_VERSION_UPNP], [3:5:0])
4916 dnl #
4917 dnl ############################################################################
4918 dnl # Release 1.6.7:
4919 dnl # "current:revision:age"
4920 dnl #
4921 dnl # - Code has changed in threadutil
4922 dnl # revision: 3 -> 4
4923 dnl # - Code has changed in upnp
4924 dnl # revision: 5 -> 6
4925 dnl # - Interfaces have been changed, added and removed in upnp
4926 dnl # current: 3 -> 4
4927 dnl # revision: 6 -> 0
4928 dnl # - Interfaces removed in upnp:
4929 dnl # age: -> 0
4930 dnl #
4931 dnl #AC_SUBST([LT_VERSION_IXML], [2:4:0])
4932 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [5:0:2])
4933 dnl #AC_SUBST([LT_VERSION_UPNP], [3:5:0])
4934 dnl #
4935 dnl # Obs.: 1.6.7 was released with a version error, the correct numbers should
4936 dnl # have been:
4937 dnl #AC_SUBST([LT_VERSION_UPNP], [4:0:0])
4938 dnl #
4939 dnl ############################################################################
4940 dnl # Release 1.6.8:
4941 dnl # "current:revision:age"
4942 dnl #
4943 dnl # - Code has changed in ixml
4944 dnl # revision: 4 -> 5
4945 dnl # - Code has changed in threadutil
4946 dnl # revision: 0 -> 1
4947 dnl # - Code has changed in upnp
4948 dnl # revision: 0 -> 1
4949 dnl #
4950 dnl #AC_SUBST([LT_VERSION_IXML], [2:5:0])
4951 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [5:1:2])
4952 dnl #AC_SUBST([LT_VERSION_UPNP], [4:1:0])
4953 dnl #
4954 dnl ############################################################################
4955 dnl # Release 1.6.9:
4956 dnl # "current:revision:age"
4957 dnl #
4958 dnl # - Code has changed in threadutil
4959 dnl # revision: 1 -> 2
4960 dnl # - Code has changed in upnp
4961 dnl # revision: 1 -> 2
4962 dnl #
4963 dnl #AC_SUBST([LT_VERSION_IXML], [2:5:0])
4964 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [5:2:2])
4965 dnl #AC_SUBST([LT_VERSION_UPNP], [4:2:0])
4966 dnl #
4967 dnl ############################################################################
4968 dnl # Release 1.6.10:
4969 dnl # "current:revision:age"
4970 dnl #
4971 dnl # - Code has changed in ixml
4972 dnl # revision: 5 -> 6
4973 dnl # - Code has changed in threadutil
4974 dnl # revision: 2 -> 3
4975 dnl # - Code has changed in upnp
4976 dnl # revision: 2 -> 3
4977 dnl # - interface has changed in upnp
4978 dnl # current: 4 -> 5
4979 dnl # revision: 3 -> 0
4980 dnl #
4981 dnl #AC_SUBST([LT_VERSION_IXML], [2:6:0])
4982 dnl ############################################################################
4983 dnl # Release 1.6.11:
4984 dnl # "current:revision:age"
4985 dnl #
4986 dnl # - Code has changed in threadutil
4987 dnl # revision: 3 -> 4
4988 dnl # - Code has changed in upnp
4989 dnl # revision: 3 -> 4
4990 dnl # - interface has changed in upnp
4991 dnl # current: 5 -> 6
4992 dnl # revision: 4 -> 0
4993 dnl #
4994 dnl #AC_SUBST([LT_VERSION_IXML], [2:6:0])
4995 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [5:4:2])
4996 dnl #AC_SUBST([LT_VERSION_UPNP], [6:0:0])
4997 dnl #
4998 dnl ############################################################################
4999 dnl # Release 1.6.12:
5000 dnl # "current:revision:age"
5001 dnl #
5002 dnl # - interface added in threadutil
5003 dnl # current: 5 -> 6
5004 dnl # revision: 4 - > 0
5005 dnl # age: 2 -> 3
5006 dnl # - interfaces removed in threadutil
5007 dnl # age: 3 -> 0
5008 dnl #
5009 dnl #AC_SUBST([LT_VERSION_IXML], [2:6:0])
5010 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [6:0:0])
5011 dnl #AC_SUBST([LT_VERSION_UPNP], [6:0:0])
5012 dnl #
5013 dnl ############################################################################
5014 dnl # Release 1.6.13:
5015 dnl # "current:revision:age"
5016 dnl #
5017 dnl # - Code has changed in upnp
5018 dnl # revision: 0 -> 1
5019 dnl # - interface added in upnp
5020 dnl # current: 6 -> 7
5021 dnl # revision: 1 - > 0
5022 dnl # age: 0 -> 1
5023 dnl #
5024 dnl #AC_SUBST([LT_VERSION_IXML], [2:6:0])
5025 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [6:0:0])
5026 dnl #AC_SUBST([LT_VERSION_UPNP], [7:0:1])
5027 dnl #
5028 dnl ############################################################################
5029 dnl # Release 1.6.14:
5030 dnl # "current:revision:age"
5031 dnl #
5032 dnl # - Code has changed in upnp
5033 dnl # revision: 0 -> 1
5034 dnl # - interface added in upnp
5035 dnl # current: 7 -> 8
5036 dnl # revision: 1 - > 0
5037 dnl # age: 1 -> 2
5038 dnl #
5039 dnl #AC_SUBST([LT_VERSION_IXML], [2:6:0])
5040 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [6:0:0])
5041 dnl #AC_SUBST([LT_VERSION_UPNP], [8:0:2])
5042 dnl #
5043 dnl ############################################################################
5044 dnl # Release 1.6.15:
5045 dnl # "current:revision:age"
5046 dnl #
5047 dnl # - Code has changed in upnp
5048 dnl # revision: 0 -> 1
5049 dnl #
5050 dnl #AC_SUBST([LT_VERSION_IXML], [2:6:0])
5051 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [6:0:0])
5052 dnl #AC_SUBST([LT_VERSION_UPNP], [8:1:2])
5053 dnl #
5054 dnl ############################################################################
5055 dnl # Release 1.6.16:
5056 dnl # "current:revision:age"
5057 dnl #
5058 dnl # - Code has changed in ixml
5059 dnl # revision: 6 -> 7
5060 dnl # - Code has changed in threadutil
5061 dnl # revision: 0 -> 1
5062 dnl # - Code has changed in upnp
5063 dnl # revision: 1 -> 2
5064 dnl # - interface changed/added/removed in upnp
5065 dnl # current++(9); revision = 0
5066 dnl # - interface added in upnp
5067 dnl # age++(3)
5068 dnl #
5069 dnl #AC_SUBST([LT_VERSION_IXML], [2:7:0])
5070 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [6:1:0])
5071 dnl #AC_SUBST([LT_VERSION_UPNP], [9:0:3])
5072 dnl #
5073 dnl ############################################################################
5074 dnl # Release 1.6.17:
5075 dnl # "current:revision:age"
5076 dnl #
5077 dnl # - Code has changed in threadutil
5078 dnl # revision: 1 -> 2
5079 dnl # - Code has changed in upnp
5080 dnl # revision: 0 -> 1
5081 dnl #
5082 dnl #AC_SUBST([LT_VERSION_IXML], [2:7:0])
5083 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [6:2:0])
5084 dnl #AC_SUBST([LT_VERSION_UPNP], [9:1:3])
5085 dnl #
5086 dnl ############################################################################
5087 dnl # Release 1.6.18:
5088 dnl # "current:revision:age"
5089 dnl #
5090 dnl # -
5091 dnl #
5092 dnl #AC_SUBST([LT_VERSION_IXML], [::])
5093 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [::])
5094 dnl #AC_SUBST([LT_VERSION_UPNP], [::])
5095 dnl #
5096 dnl ############################################################################
5097 dnl # Release 1.8.0:
5098 dnl # "current:revision:age"
5099 dnl #
5100 dnl # - Start from a new ground.
5101 dnl #
5102 dnl #AC_SUBST([LT_VERSION_IXML], [10:0:0])
5103 dnl #AC_SUBST([LT_VERSION_THREADUTIL], [10:0:0])
5104 dnl #AC_SUBST([LT_VERSION_UPNP], [10:0:0])
5105 dnl #
5106 dnl ############################################################################
5107 dnl # Release 1.8.1:
5108 dnl # "current:revision:age"
5109 dnl #
5110 dnl # - Code has changed in ixml
5111 dnl # revision: 0 -> 1
5112 dnl # - Code has changed in upnp
5113 dnl # revision: 0 -> 1
5114 dnl # - interface changed/added/removed in upnp
5115 dnl # current++(11); revision = 0
5116 dnl # - interface added in upnp
5117 dnl # age++(1)
5118 dnl #
5119 dnl #AC_SUBST([LT_VERSION_IXML], [10:1:0])
5120 dnl #AC_SUBST([LT_VERSION_UPNP], [11:0:1])
5121 dnl #
5122 dnl ############################################################################
5123 dnl # Release 1.8.2:
5124 dnl # "current:revision:age"
5125 dnl #
5126 dnl # - Code has changed in upnp
5127 dnl # revision: 0 -> 1
5128 dnl #
5129 dnl #AC_SUBST([LT_VERSION_IXML], [10:1:0])
5130 dnl #AC_SUBST([LT_VERSION_UPNP], [11:1:1])
5131 dnl #
5132 dnl ############################################################################
5133 dnl # Release 1.8.3:
5134 dnl # "current:revision:age"
5135 dnl #
5136 dnl # - Code has changed in upnp
5137 dnl # revision: 1 -> 2
5138 dnl # - interfaces changed/added/removed:
5139 dnl # current: 11 -> 12, revision 2 -> 0
5140 dnl # - interfaces added:
5141 dnl # age: 1 -> 2
5142 dnl #
5143 dnl #AC_SUBST([LT_VERSION_IXML], [10:1:0])
5144 dnl #AC_SUBST([LT_VERSION_UPNP], [12:0:2])
5145 dnl #
5146 dnl ############################################################################
5147 dnl # Release 1.8.4:
5148 dnl # "current:revision:age"
5149 dnl #
5150 dnl # - Code has changed in upnp
5151 dnl # revision: 0 -> 1
5152 dnl # - interfaces changed/added/removed:
5153 dnl # current: 12 -> 13, revision 1 -> 0
5154 dnl # - interfaces removed:
5155 dnl # age: 2 -> 0
5156 dnl #
5157 dnl #AC_SUBST([LT_VERSION_IXML], [10:1:0])
5158 dnl #AC_SUBST([LT_VERSION_UPNP], [13:0:0])
5159 dnl #
5160 dnl ############################################################################
5161 dnl # Release 1.8.5:
5162 dnl # "current:revision:age"
5163 dnl #
5164 dnl # - Code has changed in ixml
5165 dnl # revision: 1 -> 2
5166 dnl # - Code has changed in upnp
5167 dnl # revision: 0 -> 1
5168 dnl # - interfaces changed/added/removed:
5169 dnl # current: 13 -> 14, revision 1 -> 0
5170 dnl # - interfaces added:
5171 dnl # age: 0 -> 1
5172 dnl #
5173 dnl #AC_SUBST([LT_VERSION_IXML], [10:2:0])
5174 dnl #AC_SUBST([LT_VERSION_UPNP], [14:0:1])
5175 dnl #
5176 dnl ############################################################################
5177 dnl # Release 1.8.6:
5178 dnl # "current:revision:age"
5179 dnl #
5180 dnl # -
5181 dnl #
5182 dnl #AC_SUBST([LT_VERSION_IXML], [::])
5183 dnl #AC_SUBST([LT_VERSION_UPNP], [::])
5184 dnl #
5185 dnl ############################################################################
5186 dnl # Release 1.10.0:
5187 dnl # "current:revision:age"
5188 dnl #
5189 dnl # - First release of 1.10
5190 dnl #
5191 dnl #AC_SUBST([LT_VERSION_IXML], [11:0:0])
5192 dnl #AC_SUBST([LT_VERSION_UPNP], [15:0:0])
5193 dnl #
5194 dnl ############################################################################
5195 dnl # Release 1.10.1
5196 dnl # "current:revision:age"
5197 dnl #
5198 dnl # - Code has changed in upnp
5199 dnl # revision: 0 -> 1
5200 dnl #
5201 dnl #AC_SUBST([LT_VERSION_IXML], [11:0:0])
5202 dnl #AC_SUBST([LT_VERSION_UPNP], [15:1:0])
5203 dnl #
5204 dnl ############################################################################
5205 dnl # Release 1.12.0
5206 dnl # "current:revision:age"
5207 dnl #
5208 dnl # - Code has changed in upnp
5209 dnl # revision: 1 -> 2
5210 dnl # - interfaces changed/added/removed:
5211 dnl # current: 15 -> 16
5212 dnl # revision: 2 -> 0
5213 dnl # - interfaces removed or changed:
5214 dnl # age: 0 -> 0
5215 dnl #
5216 dnl #AC_SUBST([LT_VERSION_IXML], [11:0:0])
5217 dnl #AC_SUBST([LT_VERSION_UPNP], [16:0:0])
5218 dnl #
5219 dnl ############################################################################