| 1 |
;;; |
| 2 |
;;; tools/docprep - prepare document |
| 3 |
;;; |
| 4 |
;;; Copyright (c) 2001-2022 Shiro Kawai <[email protected]> |
| 5 |
;;; |
| 6 |
;;; Permission to use, copy, modify, distribute this software and |
| 7 |
;;; accompanying documentation for any purpose is hereby granted, |
| 8 |
;;; provided that existing copyright notices are retained in all |
| 9 |
;;; copies and that this notice is included verbatim in all |
| 10 |
;;; distributions. |
| 11 |
;;; This software is provided as is, without express or implied |
| 12 |
;;; warranty. In no circumstances the author(s) shall be liable |
| 13 |
;;; for any damages arising out of the use of this software. |
| 14 |
;;; |
| 15 |
|
| 16 |
(use file.util) |
| 17 |
(use gauche.charconv) |
| 18 |
(use gauche.parseopt) |
| 19 |
(use srfi.13) |
| 20 |
(use util.levenshtein) |
| 21 |
|
| 22 |
(define *version* |
| 23 |
(string-trim-both (file->string (or (find (cut sys-access <> R_OK) |
| 24 |
'("../VERSION" |
| 25 |
"./VERSION")) |
| 26 |
(error "No VERSION file?"))))) |
| 27 |
(define *node-table* '()) |
| 28 |
(define *header-table* '()) |
| 29 |
|
| 30 |
(define lang (make-parameter 'en)) |
| 31 |
(define srcdir (make-parameter "")) |
| 32 |
|
| 33 |
;; TRANSIENT: We use :converion-illegal-output 'replace to read utf8 files, |
| 34 |
;; so that docprep won't barf if Gauche is compiled with non-utf8 encoding. |
| 35 |
;; (The resulting info file may contain replace chars, but that's unavoidable). |
| 36 |
;; After 0.9.13 release and we drop other encoding support, remove |
| 37 |
;; :converion-illegal-output option, for it is no longer needed. |
| 38 |
|
| 39 |
(define (scan-nodes) |
| 40 |
(let ([current-node #f] |
| 41 |
[current-header #f]) |
| 42 |
(port-for-each |
| 43 |
(lambda (line) |
| 44 |
(rxmatch-case line |
| 45 |
[#/^@node\s+([^,]+)/ (#f node) |
| 46 |
(set! current-node (string-trim-right node))] |
| 47 |
[#/^@(chapter|(sub)*section|appendix\w*)\s+(.*)/ (#f #f #f header) |
| 48 |
(set! current-header (string-trim-right header))] |
| 49 |
[#/^@c NODE\s+([^,]*)(,(.*))?/ (#f jnode #f jheader) |
| 50 |
(let* ([jn (string-trim-right jnode)] |
| 51 |
[jh (if jheader (string-trim-both jheader) jn)]) |
| 52 |
(push! *node-table* (cons current-node jn)) |
| 53 |
(push! *header-table* (cons current-header jh)))] |
| 54 |
[#/^@include\s+(\S+)/ (#f file) |
| 55 |
(with-input-from-file (find-included-file file) |
| 56 |
(cut scan-nodes) |
| 57 |
:encoding 'utf8 |
| 58 |
:conversion-illegal-output 'replace)] |
| 59 |
)) |
| 60 |
read-line))) |
| 61 |
|
| 62 |
(define (filter pattern-in pattern-out) |
| 63 |
;; This regexp picks potential typo of extract directives. |
| 64 |
(define suspicious |
| 65 |
#/^@c ([A-Za-z]{2,7}:?)(?![,-])\s*(.*)/) |
| 66 |
;; Loop for 'active' lines. |
| 67 |
;; MODTAG tracks whether the previous line was "[SRFI-nn]", "@c MOD ...", or |
| 68 |
;; something else. It is to detect the error when those two are reversed. |
| 69 |
(define (in line modtag) |
| 70 |
(rxmatch-case line |
| 71 |
[test eof-object?] |
| 72 |
[pattern-in () (in (read-line) modtag)] |
| 73 |
[pattern-out () (out (read-line) modtag)] |
| 74 |
[#/^@include\s+(\S+)/ (#f file) |
| 75 |
(finish-srfi-line modtag) |
| 76 |
(with-input-from-file (find-included-file file) |
| 77 |
(cut filter pattern-in pattern-out) |
| 78 |
:encoding 'utf8 |
| 79 |
:conversion-illegal-output 'replace) |
| 80 |
(in (read-line) #f)] |
| 81 |
;; Extract-time include. @include will be tracked by |
| 82 |
;; texinfo-multiple-files-update and the includee must have @node |
| 83 |
;; at the toplevel, so this can be used to include files without |
| 84 |
;; the restriction. |
| 85 |
[#/^@c xinclude\s+(\S+)/ (#f file) |
| 86 |
(finish-srfi-line modtag) |
| 87 |
(with-input-from-file (find-included-file file) |
| 88 |
(cut filter pattern-in pattern-out) |
| 89 |
:encoding 'utf8 |
| 90 |
:conversion-illegal-output 'replace) |
| 91 |
(in (read-line) #f)] |
| 92 |
[#/^@c COMMON$/ () (finish-srfi-line modtag) (in (read-line) #f)] |
| 93 |
[#/^@c MOD\s+(\S+)$/ (#f module) |
| 94 |
(display #"@{@t{~|module|}@}\n") |
| 95 |
(in (read-line) 'mod)] |
| 96 |
[#/^@c DEPRECATED$/ () |
| 97 |
(display "@b{Deprecated. }\n") |
| 98 |
(in (read-line) modtag)] |
| 99 |
[#/^\[SRFI/ (#f) |
| 100 |
(when (eq? modtag 'mod) |
| 101 |
(warn #"SRFI tag follows MOD directive: ~|line|\n")) |
| 102 |
(display line) |
| 103 |
(in (read-line) 'srfi)] |
| 104 |
[suspicious (#f word rest) |
| 105 |
(check-typo word rest line) |
| 106 |
(finish-srfi-line modtag) |
| 107 |
(in (read-line) #f)] |
| 108 |
[test (^_ (eq? (lang) 'en)) |
| 109 |
(finish-srfi-line modtag) |
| 110 |
(display (regexp-replace-all #/@VERSION@/ line *version*)) |
| 111 |
(newline) (in (read-line) #f)] |
| 112 |
[#/^@node\s+(.*)$/ (#f nodedesc) |
| 113 |
(process-node nodedesc) (in (read-line) #f)] |
| 114 |
[#/^@(chapter|(sub)*section|appendix\w*)\s+(.*)/ (#f cmd #f header) |
| 115 |
(process-header cmd header) (in (read-line) #f)] |
| 116 |
[#/^\* ([^:]+)::(.*)?/ (#f node desc) |
| 117 |
(process-menu node #f desc) (in (read-line) #f)] |
| 118 |
[#/^\* ([^:]+):\s+([^\)]+\))\.(.*)?/ (#f tag node desc) |
| 119 |
(process-menu node tag desc) (in (read-line) #f)] |
| 120 |
[else (finish-srfi-line modtag) |
| 121 |
(display |
| 122 |
($ regexp-replace-all* line |
| 123 |
#/@VERSION@/ *version* |
| 124 |
#/(@(?:px|x)?ref)\{([^\}]+)\}/ process-ref)) |
| 125 |
(newline) |
| 126 |
(in (read-line) #f)])) |
| 127 |
|
| 128 |
;; When we see '[SRFI-nn]' line, we don't emit newline imediately, |
| 129 |
;; for we don't want space if it is immediately followed by '{module}'. |
| 130 |
;; This one emits the delayed newline when the next line isn't the module. |
| 131 |
(define (finish-srfi-line modtag) |
| 132 |
(when (eq? modtag 'srfi) (newline))) |
| 133 |
|
| 134 |
|
| 135 |
;; Loop for 'inactive' lines. |
| 136 |
(define (out line modtag) |
| 137 |
(rxmatch-case line |
| 138 |
[test eof-object?] |
| 139 |
[pattern-in () (in (read-line) modtag)] |
| 140 |
[#/^@c COMMON$/ () (in (read-line) modtag)] |
| 141 |
[suspicious (#f word rest) |
| 142 |
(check-typo word rest line) |
| 143 |
(out (read-line) modtag)] |
| 144 |
[else (out (read-line) modtag)])) |
| 145 |
|
| 146 |
(in (read-line) #f)) |
| 147 |
|
| 148 |
;; Detect potential typo of extract directives. |
| 149 |
;; It is a bit complicated, since we need to allow general comment after @c. |
| 150 |
;; WORD is a 2-7 letter alphabetic word, and REST is what follows. |
| 151 |
(define (check-typo word rest line) |
| 152 |
(when (and (string-any char-upper-case? word) |
| 153 |
(or (and (equal? rest "") |
| 154 |
;; directives that doesn't take argument |
| 155 |
(let1 ds (re-distances word '("EN" "JP" "COMMON")) |
| 156 |
(<= 1 (apply min ds) 2))) |
| 157 |
;; directives that does take argument |
| 158 |
(and (>= (string-length word) 3) |
| 159 |
(let1 ds (re-distances word '("NODE" "MOD" "NOTE" |
| 160 |
"TODO:" "NB:")) |
| 161 |
(<= 1 (apply min ds) 2))))) |
| 162 |
(errorf "Unrecognized extract directive ~s at or near ~s:~d: ~a\n" |
| 163 |
word (port-name (current-input-port)) |
| 164 |
(port-current-line (current-input-port)) |
| 165 |
line))) |
| 166 |
|
| 167 |
;; We search relative to the current directory first, then |
| 168 |
;; relative to $srcdir; for FILE may be the generated one. |
| 169 |
(define (find-included-file file) |
| 170 |
(if (file-exists? file) |
| 171 |
file |
| 172 |
(build-path (srcdir) file))) |
| 173 |
|
| 174 |
(define (lookup name table) |
| 175 |
(cond [(assoc (string-trim-both name) table) => cdr] |
| 176 |
[else name])) |
| 177 |
|
| 178 |
(define (process-node nodedesc) |
| 179 |
(display "@node ") |
| 180 |
(display (string-join |
| 181 |
(map (cut lookup <> *node-table*) (string-split nodedesc #\,)) |
| 182 |
", ")) |
| 183 |
(newline)) |
| 184 |
|
| 185 |
(define (process-header cmd header) |
| 186 |
(format #t "@~a ~a\n" cmd (lookup header *header-table*))) |
| 187 |
|
| 188 |
(define (process-menu node tag desc) |
| 189 |
(if tag |
| 190 |
(format #t "* ~a: ~a. ~a\n" tag (lookup node *node-table*) |
| 191 |
(string-trim-both (or desc ""))) |
| 192 |
(format #t "* ~a:: ~a\n" (lookup node *node-table*) |
| 193 |
(string-trim-both (or desc ""))))) |
| 194 |
|
| 195 |
(define (process-ref match) |
| 196 |
(let ([cmd (rxmatch-substring match 1)] |
| 197 |
[node (rxmatch-substring match 2)]) |
| 198 |
(format #f "~a{~a}" cmd (lookup node *node-table*)))) |
| 199 |
|
| 200 |
;; TRANSIENT: |
| 201 |
;; This script is used by 0.9.15 to prepare tarball of sources after 0.9.15, |
| 202 |
;; so it needs to work without the newly added parseopt's help string feature. |
| 203 |
;; We just comment out the help string thingy for now. Revive them after |
| 204 |
;; the new release. |
| 205 |
(define (usage) |
| 206 |
(print "Usage: gosh tools/docprep [options] infile") |
| 207 |
;;(print "Options:") |
| 208 |
;;(print (option-parser-help-string)) |
| 209 |
(exit 1)) |
| 210 |
|
| 211 |
;; TODO: Switch to use 'ja' (language code) instead of 'jp' (country code) |
| 212 |
(define (main args) |
| 213 |
(define language 'en) |
| 214 |
(let-args (cdr args) |
| 215 |
([outfile "o=s" |
| 216 |
;; TRANSIENT: See above comment |
| 217 |
;; "o=s{FILE}" |
| 218 |
;; ? "Writes output to {FILE} instead of stdout." |
| 219 |
] |
| 220 |
[#f "en" => (^[] (set! language 'en)) |
| 221 |
;; TRANSIENT: See above comment |
| 222 |
;; ? "Set output language to English" |
| 223 |
] |
| 224 |
[#f "jp" => (^[] (set! language 'jp)) |
| 225 |
;; TRANSIENT: See above comment |
| 226 |
;; ? "Set output language to Japanese" |
| 227 |
] |
| 228 |
[else _ (usage)] |
| 229 |
. a) |
| 230 |
|
| 231 |
(define (do-it) |
| 232 |
(case language |
| 233 |
[(en) (filter #/^@c EN$/ #/^@c JP$/)] |
| 234 |
[(jp) (filter #/^@c JP$/ #/^@c EN$/)])) |
| 235 |
|
| 236 |
(unless (= (length a) 1) (usage)) |
| 237 |
|
| 238 |
(parameterize ([srcdir (sys-normalize-pathname (sys-dirname (car a)) |
| 239 |
:absolute #t)] |
| 240 |
[lang language]) |
| 241 |
(when (eq? language 'jp) |
| 242 |
(with-input-from-file (car a) scan-nodes |
| 243 |
:encoding 'utf8 |
| 244 |
:conversion-illegal-output 'replace)) |
| 245 |
(with-input-from-file (car a) |
| 246 |
(^[] |
| 247 |
(if outfile |
| 248 |
(with-output-to-file outfile do-it :encoding 'utf8) |
| 249 |
(let1 out (wrap-with-output-conversion |
| 250 |
(current-output-port) 'utf8) |
| 251 |
(with-output-to-port out do-it) |
| 252 |
(close-output-port out)))) |
| 253 |
:encoding 'utf8 |
| 254 |
:conversion-illegal-output 'replace)) |
| 255 |
0)) |
| 256 |
|
| 257 |
;; Local variables: |
| 258 |
;; mode: Scheme |
| 259 |
;; end: |