| 1 |
|
| 2 |
source [file join [file dirname [info script]] parseunicode.tcl] |
| 3 |
|
| 4 |
proc print_rd {map} { |
| 5 |
global tl_lookup_table |
| 6 |
set aChar [list] |
| 7 |
set lRange [list] |
| 8 |
|
| 9 |
set nRange 1 |
| 10 |
set iFirst [lindex $map 0 0] |
| 11 |
set cPrev [lindex $map 0 1] |
| 12 |
set fPrev [lindex $map 0 2] |
| 13 |
|
| 14 |
foreach m [lrange $map 1 end] { |
| 15 |
foreach {i c f} $m {} |
| 16 |
|
| 17 |
if {$cPrev == $c && $fPrev==$f} { |
| 18 |
for {set j [expr $iFirst+$nRange]} {$j<$i} {incr j} { |
| 19 |
if {[info exists tl_lookup_table($j)]==0} break |
| 20 |
} |
| 21 |
|
| 22 |
if {$j==$i} { |
| 23 |
set nNew [expr {(1 + $i - $iFirst)}] |
| 24 |
if {$nNew<=8} { |
| 25 |
set nRange $nNew |
| 26 |
continue |
| 27 |
} |
| 28 |
} |
| 29 |
} |
| 30 |
|
| 31 |
lappend lRange [list $iFirst $nRange] |
| 32 |
lappend aChar $cPrev |
| 33 |
lappend aFlag $fPrev |
| 34 |
|
| 35 |
set iFirst $i |
| 36 |
set cPrev $c |
| 37 |
set fPrev $f |
| 38 |
set nRange 1 |
| 39 |
} |
| 40 |
lappend lRange [list $iFirst $nRange] |
| 41 |
lappend aChar $cPrev |
| 42 |
lappend aFlag $fPrev |
| 43 |
|
| 44 |
puts "/*" |
| 45 |
puts "** If the argument is a codepoint corresponding to a lowercase letter" |
| 46 |
puts "** in the ASCII range with a diacritic added, return the codepoint" |
| 47 |
puts "** of the ASCII letter only. For example, if passed 235 - \"LATIN" |
| 48 |
puts "** SMALL LETTER E WITH DIAERESIS\" - return 65 (\"LATIN SMALL LETTER" |
| 49 |
puts "** E\"). The resuls of passing a codepoint that corresponds to an" |
| 50 |
puts "** uppercase letter are undefined." |
| 51 |
puts "*/" |
| 52 |
puts "static int ${::remove_diacritic}(int c, int bComplex)\{" |
| 53 |
puts " unsigned short aDia\[\] = \{" |
| 54 |
puts -nonewline " 0, " |
| 55 |
set i 1 |
| 56 |
foreach r $lRange { |
| 57 |
foreach {iCode nRange} $r {} |
| 58 |
if {($i % 8)==0} {puts "" ; puts -nonewline " " } |
| 59 |
incr i |
| 60 |
|
| 61 |
puts -nonewline [format "%5d" [expr ($iCode<<3) + $nRange-1]] |
| 62 |
puts -nonewline ", " |
| 63 |
} |
| 64 |
puts "" |
| 65 |
puts " \};" |
| 66 |
puts "#define HIBIT ((unsigned char)0x80)" |
| 67 |
puts " unsigned char aChar\[\] = \{" |
| 68 |
puts -nonewline " '\\0', " |
| 69 |
set i 1 |
| 70 |
foreach c $aChar f $aFlag { |
| 71 |
if { $f } { |
| 72 |
set str "'$c'|HIBIT, " |
| 73 |
} else { |
| 74 |
set str "'$c', " |
| 75 |
} |
| 76 |
if {$c == ""} { set str "'\\0', " } |
| 77 |
|
| 78 |
if {($i % 6)==0} {puts "" ; puts -nonewline " " } |
| 79 |
incr i |
| 80 |
puts -nonewline "$str" |
| 81 |
} |
| 82 |
puts "" |
| 83 |
puts " \};" |
| 84 |
puts { |
| 85 |
unsigned int key = (((unsigned int)c)<<3) | 0x00000007; |
| 86 |
int iRes = 0; |
| 87 |
int iHi = sizeof(aDia)/sizeof(aDia[0]) - 1; |
| 88 |
int iLo = 0; |
| 89 |
while( iHi>=iLo ){ |
| 90 |
int iTest = (iHi + iLo) / 2; |
| 91 |
if( key >= aDia[iTest] ){ |
| 92 |
iRes = iTest; |
| 93 |
iLo = iTest+1; |
| 94 |
}else{ |
| 95 |
iHi = iTest-1; |
| 96 |
} |
| 97 |
} |
| 98 |
assert( key>=aDia[iRes] ); |
| 99 |
if( bComplex==0 && (aChar[iRes] & 0x80) ) return c; |
| 100 |
return (c > (aDia[iRes]>>3) + (aDia[iRes]&0x07)) ? c : ((int)aChar[iRes] & 0x7F);} |
| 101 |
puts "\}" |
| 102 |
} |
| 103 |
|
| 104 |
proc print_isdiacritic {zFunc map} { |
| 105 |
|
| 106 |
set lCode [list] |
| 107 |
foreach m $map { |
| 108 |
foreach {code char flag} $m {} |
| 109 |
if {$flag} continue |
| 110 |
if {$code && $char == ""} { lappend lCode $code } |
| 111 |
} |
| 112 |
set lCode [lsort -integer $lCode] |
| 113 |
set iFirst [lindex $lCode 0] |
| 114 |
set iLast [lindex $lCode end] |
| 115 |
|
| 116 |
set i1 0 |
| 117 |
set i2 0 |
| 118 |
|
| 119 |
foreach c $lCode { |
| 120 |
set i [expr $c - $iFirst] |
| 121 |
if {$i < 32} { |
| 122 |
set i1 [expr {$i1 | (1<<$i)}] |
| 123 |
} else { |
| 124 |
set i2 [expr {$i2 | (1<<($i-32))}] |
| 125 |
} |
| 126 |
} |
| 127 |
|
| 128 |
puts "/*" |
| 129 |
puts "** Return true if the argument interpreted as a unicode codepoint" |
| 130 |
puts "** is a diacritical modifier character." |
| 131 |
puts "*/" |
| 132 |
puts "int ${zFunc}\(int c)\{" |
| 133 |
puts " unsigned int mask0 = [format "0x%08X" $i1];" |
| 134 |
puts " unsigned int mask1 = [format "0x%08X" $i2];" |
| 135 |
|
| 136 |
puts " if( c<$iFirst || c>$iLast ) return 0;" |
| 137 |
puts " return (c < $iFirst+32) ?" |
| 138 |
puts " (mask0 & ((unsigned int)1 << (c-$iFirst))) :" |
| 139 |
puts " (mask1 & ((unsigned int)1 << (c-$iFirst-32)));" |
| 140 |
puts "\}" |
| 141 |
} |
| 142 |
|
| 143 |
|
| 144 |
#------------------------------------------------------------------------- |
| 145 |
|
| 146 |
proc an_load_separator_ranges {} { |
| 147 |
global unicodedata.txt |
| 148 |
set lSep [an_load_unicodedata_text ${unicodedata.txt}] |
| 149 |
unset -nocomplain iFirst |
| 150 |
unset -nocomplain nRange |
| 151 |
set lRange [list] |
| 152 |
foreach sep $lSep { |
| 153 |
if {0==[info exists iFirst]} { |
| 154 |
set iFirst $sep |
| 155 |
set nRange 1 |
| 156 |
} elseif { $sep == ($iFirst+$nRange) } { |
| 157 |
incr nRange |
| 158 |
} else { |
| 159 |
lappend lRange [list $iFirst $nRange] |
| 160 |
set iFirst $sep |
| 161 |
set nRange 1 |
| 162 |
} |
| 163 |
} |
| 164 |
lappend lRange [list $iFirst $nRange] |
| 165 |
set lRange |
| 166 |
} |
| 167 |
|
| 168 |
proc an_print_range_array {lRange} { |
| 169 |
set iFirstMax 0 |
| 170 |
set nRangeMax 0 |
| 171 |
foreach range $lRange { |
| 172 |
foreach {iFirst nRange} $range {} |
| 173 |
if {$iFirst > $iFirstMax} {set iFirstMax $iFirst} |
| 174 |
if {$nRange > $nRangeMax} {set nRangeMax $nRange} |
| 175 |
} |
| 176 |
if {$iFirstMax >= (1<<22)} {error "first-max is too large for format"} |
| 177 |
if {$nRangeMax >= (1<<10)} {error "range-max is too large for format"} |
| 178 |
|
| 179 |
puts -nonewline " " |
| 180 |
puts [string trim { |
| 181 |
/* Each unsigned integer in the following array corresponds to a contiguous |
| 182 |
** range of unicode codepoints that are not either letters or numbers (i.e. |
| 183 |
** codepoints for which this function should return 0). |
| 184 |
** |
| 185 |
** The most significant 22 bits in each 32-bit value contain the first |
| 186 |
** codepoint in the range. The least significant 10 bits are used to store |
| 187 |
** the size of the range (always at least 1). In other words, the value |
| 188 |
** ((C<<22) + N) represents a range of N codepoints starting with codepoint |
| 189 |
** C. It is not possible to represent a range larger than 1023 codepoints |
| 190 |
** using this format. |
| 191 |
*/ |
| 192 |
}] |
| 193 |
puts -nonewline " static const unsigned int aEntry\[\] = \{" |
| 194 |
set i 0 |
| 195 |
foreach range $lRange { |
| 196 |
foreach {iFirst nRange} $range {} |
| 197 |
set u32 [format "0x%08X" [expr ($iFirst<<10) + $nRange]] |
| 198 |
|
| 199 |
if {($i % 5)==0} {puts "" ; puts -nonewline " "} |
| 200 |
puts -nonewline " $u32," |
| 201 |
incr i |
| 202 |
} |
| 203 |
puts "" |
| 204 |
puts " \};" |
| 205 |
} |
| 206 |
|
| 207 |
proc an_print_ascii_bitmap {lRange} { |
| 208 |
foreach range $lRange { |
| 209 |
foreach {iFirst nRange} $range {} |
| 210 |
for {set i $iFirst} {$i < ($iFirst+$nRange)} {incr i} { |
| 211 |
if {$i<=127} { set a($i) 1 } |
| 212 |
} |
| 213 |
} |
| 214 |
|
| 215 |
set aAscii [list 0 0 0 0] |
| 216 |
foreach key [array names a] { |
| 217 |
set idx [expr $key >> 5] |
| 218 |
lset aAscii $idx [expr [lindex $aAscii $idx] | (1 << ($key&0x001F))] |
| 219 |
} |
| 220 |
|
| 221 |
puts " static const unsigned int aAscii\[4\] = \{" |
| 222 |
puts -nonewline " " |
| 223 |
foreach v $aAscii { puts -nonewline [format " 0x%08X," $v] } |
| 224 |
puts "" |
| 225 |
puts " \};" |
| 226 |
} |
| 227 |
|
| 228 |
proc print_isalnum {zFunc lRange} { |
| 229 |
puts "/*" |
| 230 |
puts "** Return true if the argument corresponds to a unicode codepoint" |
| 231 |
puts "** classified as either a letter or a number. Otherwise false." |
| 232 |
puts "**" |
| 233 |
puts "** The results are undefined if the value passed to this function" |
| 234 |
puts "** is less than zero." |
| 235 |
puts "*/" |
| 236 |
puts "int ${zFunc}\(int c)\{" |
| 237 |
an_print_range_array $lRange |
| 238 |
an_print_ascii_bitmap $lRange |
| 239 |
puts { |
| 240 |
if( (unsigned int)c<128 ){ |
| 241 |
return ( (aAscii[c >> 5] & ((unsigned int)1 << (c & 0x001F)))==0 ); |
| 242 |
}else if( (unsigned int)c<(1<<22) ){ |
| 243 |
unsigned int key = (((unsigned int)c)<<10) | 0x000003FF; |
| 244 |
int iRes = 0; |
| 245 |
int iHi = sizeof(aEntry)/sizeof(aEntry[0]) - 1; |
| 246 |
int iLo = 0; |
| 247 |
while( iHi>=iLo ){ |
| 248 |
int iTest = (iHi + iLo) / 2; |
| 249 |
if( key >= aEntry[iTest] ){ |
| 250 |
iRes = iTest; |
| 251 |
iLo = iTest+1; |
| 252 |
}else{ |
| 253 |
iHi = iTest-1; |
| 254 |
} |
| 255 |
} |
| 256 |
assert( aEntry[0]<key ); |
| 257 |
assert( key>=aEntry[iRes] ); |
| 258 |
return (((unsigned int)c) >= ((aEntry[iRes]>>10) + (aEntry[iRes]&0x3FF))); |
| 259 |
} |
| 260 |
return 1;} |
| 261 |
puts "\}" |
| 262 |
} |
| 263 |
|
| 264 |
proc print_test_isalnum {zFunc lRange} { |
| 265 |
foreach range $lRange { |
| 266 |
foreach {iFirst nRange} $range {} |
| 267 |
for {set i $iFirst} {$i < ($iFirst+$nRange)} {incr i} { set a($i) 1 } |
| 268 |
} |
| 269 |
|
| 270 |
puts "static int isalnum_test(int *piCode)\{" |
| 271 |
puts -nonewline " unsigned char aAlnum\[\] = \{" |
| 272 |
for {set i 0} {$i < 70000} {incr i} { |
| 273 |
if {($i % 32)==0} { puts "" ; puts -nonewline " " } |
| 274 |
set bFlag [expr ![info exists a($i)]] |
| 275 |
puts -nonewline "${bFlag}," |
| 276 |
} |
| 277 |
puts "" |
| 278 |
puts " \};" |
| 279 |
|
| 280 |
puts -nonewline " int aLargeSep\[\] = \{" |
| 281 |
set i 0 |
| 282 |
foreach iSep [lsort -integer [array names a]] { |
| 283 |
if {$iSep<70000} continue |
| 284 |
if {($i % 8)==0} { puts "" ; puts -nonewline " " } |
| 285 |
puts -nonewline " $iSep," |
| 286 |
incr i |
| 287 |
} |
| 288 |
puts "" |
| 289 |
puts " \};" |
| 290 |
puts -nonewline " int aLargeOther\[\] = \{" |
| 291 |
set i 0 |
| 292 |
foreach iSep [lsort -integer [array names a]] { |
| 293 |
if {$iSep<70000} continue |
| 294 |
if {[info exists a([expr $iSep-1])]==0} { |
| 295 |
if {($i % 8)==0} { puts "" ; puts -nonewline " " } |
| 296 |
puts -nonewline " [expr $iSep-1]," |
| 297 |
incr i |
| 298 |
} |
| 299 |
if {[info exists a([expr $iSep+1])]==0} { |
| 300 |
if {($i % 8)==0} { puts "" ; puts -nonewline " " } |
| 301 |
puts -nonewline " [expr $iSep+1]," |
| 302 |
incr i |
| 303 |
} |
| 304 |
} |
| 305 |
puts "" |
| 306 |
puts " \};" |
| 307 |
|
| 308 |
puts [subst -nocommands { |
| 309 |
int i; |
| 310 |
for(i=0; i<sizeof(aAlnum)/sizeof(aAlnum[0]); i++){ |
| 311 |
if( ${zFunc}(i)!=aAlnum[i] ){ |
| 312 |
*piCode = i; |
| 313 |
return 1; |
| 314 |
} |
| 315 |
} |
| 316 |
for(i=0; i<sizeof(aLargeSep)/sizeof(aLargeSep[0]); i++){ |
| 317 |
if( ${zFunc}(aLargeSep[i])!=0 ){ |
| 318 |
*piCode = aLargeSep[i]; |
| 319 |
return 1; |
| 320 |
} |
| 321 |
} |
| 322 |
for(i=0; i<sizeof(aLargeOther)/sizeof(aLargeOther[0]); i++){ |
| 323 |
if( ${zFunc}(aLargeOther[i])!=1 ){ |
| 324 |
*piCode = aLargeOther[i]; |
| 325 |
return 1; |
| 326 |
} |
| 327 |
} |
| 328 |
}] |
| 329 |
puts " return 0;" |
| 330 |
puts "\}" |
| 331 |
} |
| 332 |
|
| 333 |
#------------------------------------------------------------------------- |
| 334 |
|
| 335 |
proc tl_create_records {} { |
| 336 |
global tl_lookup_table |
| 337 |
|
| 338 |
set iFirst "" |
| 339 |
set nOff 0 |
| 340 |
set nRange 0 |
| 341 |
set nIncr 0 |
| 342 |
|
| 343 |
set lRecord [list] |
| 344 |
foreach code [lsort -integer [array names tl_lookup_table]] { |
| 345 |
set mapping $tl_lookup_table($code) |
| 346 |
if {$iFirst == ""} { |
| 347 |
set iFirst $code |
| 348 |
set nOff [expr $mapping - $code] |
| 349 |
set nRange 1 |
| 350 |
set nIncr 1 |
| 351 |
} else { |
| 352 |
set diff [expr $code - ($iFirst + ($nIncr * ($nRange - 1)))] |
| 353 |
if { $nRange==1 && ($diff==1 || $diff==2) } { |
| 354 |
set nIncr $diff |
| 355 |
} |
| 356 |
|
| 357 |
if {$diff != $nIncr || ($mapping - $code)!=$nOff} { |
| 358 |
if { $nRange==1 } {set nIncr 1} |
| 359 |
lappend lRecord [list $iFirst $nIncr $nRange $nOff] |
| 360 |
set iFirst $code |
| 361 |
set nOff [expr $mapping - $code] |
| 362 |
set nRange 1 |
| 363 |
set nIncr 1 |
| 364 |
} else { |
| 365 |
incr nRange |
| 366 |
} |
| 367 |
} |
| 368 |
} |
| 369 |
|
| 370 |
lappend lRecord [list $iFirst $nIncr $nRange $nOff] |
| 371 |
|
| 372 |
set lRecord |
| 373 |
} |
| 374 |
|
| 375 |
proc tl_print_table_header {} { |
| 376 |
puts -nonewline " " |
| 377 |
puts [string trim { |
| 378 |
/* Each entry in the following array defines a rule for folding a range |
| 379 |
** of codepoints to lower case. The rule applies to a range of nRange |
| 380 |
** codepoints starting at codepoint iCode. |
| 381 |
** |
| 382 |
** If the least significant bit in flags is clear, then the rule applies |
| 383 |
** to all nRange codepoints (i.e. all nRange codepoints are upper case and |
| 384 |
** need to be folded). Or, if it is set, then the rule only applies to |
| 385 |
** every second codepoint in the range, starting with codepoint C. |
| 386 |
** |
| 387 |
** The 7 most significant bits in flags are an index into the aiOff[] |
| 388 |
** array. If a specific codepoint C does require folding, then its lower |
| 389 |
** case equivalent is ((C + aiOff[flags>>1]) & 0xFFFF). |
| 390 |
** |
| 391 |
** The contents of this array are generated by parsing the CaseFolding.txt |
| 392 |
** file distributed as part of the "Unicode Character Database". See |
| 393 |
** http://www.unicode.org for details. |
| 394 |
*/ |
| 395 |
}] |
| 396 |
puts " static const struct TableEntry \{" |
| 397 |
puts " unsigned short iCode;" |
| 398 |
puts " unsigned char flags;" |
| 399 |
puts " unsigned char nRange;" |
| 400 |
puts " \} aEntry\[\] = \{" |
| 401 |
} |
| 402 |
|
| 403 |
proc tl_print_table_entry {togglevar entry liOff} { |
| 404 |
upvar $togglevar t |
| 405 |
foreach {iFirst nIncr nRange nOff} $entry {} |
| 406 |
|
| 407 |
if {$iFirst > (1<<16)} { return 1 } |
| 408 |
|
| 409 |
if {[info exists t]==0} {set t 0} |
| 410 |
if {$t==0} { puts -nonewline " " } |
| 411 |
|
| 412 |
set flags 0 |
| 413 |
if {$nIncr==2} { set flags 1 ; set nRange [expr $nRange * 2]} |
| 414 |
if {$nOff<0} { incr nOff [expr (1<<16)] } |
| 415 |
|
| 416 |
set idx [lsearch $liOff $nOff] |
| 417 |
if {$idx<0} {error "malfunction generating aiOff"} |
| 418 |
set flags [expr $flags + $idx*2] |
| 419 |
|
| 420 |
set txt "{$iFirst, $flags, $nRange}," |
| 421 |
if {$t==2} { |
| 422 |
puts $txt |
| 423 |
} else { |
| 424 |
puts -nonewline [format "% -23s" $txt] |
| 425 |
} |
| 426 |
set t [expr ($t+1)%3] |
| 427 |
|
| 428 |
return 0 |
| 429 |
} |
| 430 |
|
| 431 |
proc tl_print_table_footer {togglevar} { |
| 432 |
upvar $togglevar t |
| 433 |
if {$t!=0} {puts ""} |
| 434 |
puts " \};" |
| 435 |
} |
| 436 |
|
| 437 |
proc tl_print_if_entry {entry} { |
| 438 |
foreach {iFirst nIncr nRange nOff} $entry {} |
| 439 |
if {$nIncr==2} {error "tl_print_if_entry needs improvement!"} |
| 440 |
|
| 441 |
puts " else if( c>=$iFirst && c<[expr $iFirst+$nRange] )\{" |
| 442 |
puts " ret = c + $nOff;" |
| 443 |
puts " \}" |
| 444 |
} |
| 445 |
|
| 446 |
proc tl_generate_ioff_table {lRecord} { |
| 447 |
foreach entry $lRecord { |
| 448 |
foreach {iFirst nIncr nRange iOff} $entry {} |
| 449 |
if {$iOff<0} { incr iOff [expr (1<<16)] } |
| 450 |
if {[info exists a($iOff)]} continue |
| 451 |
set a($iOff) 1 |
| 452 |
} |
| 453 |
|
| 454 |
set liOff [lsort -integer [array names a]] |
| 455 |
if {[llength $liOff]>128} { error "Too many distinct ioffs" } |
| 456 |
return $liOff |
| 457 |
} |
| 458 |
|
| 459 |
proc tl_print_ioff_table {liOff} { |
| 460 |
puts -nonewline " static const unsigned short aiOff\[\] = \{" |
| 461 |
set i 0 |
| 462 |
foreach off $liOff { |
| 463 |
if {($i % 8)==0} {puts "" ; puts -nonewline " "} |
| 464 |
puts -nonewline [format "% -7s" "$off,"] |
| 465 |
incr i |
| 466 |
} |
| 467 |
puts "" |
| 468 |
puts " \};" |
| 469 |
|
| 470 |
} |
| 471 |
|
| 472 |
proc print_fold {zFunc} { |
| 473 |
|
| 474 |
set lRecord [tl_create_records] |
| 475 |
|
| 476 |
set lHigh [list] |
| 477 |
puts "/*" |
| 478 |
puts "** Interpret the argument as a unicode codepoint. If the codepoint" |
| 479 |
puts "** is an upper case character that has a lower case equivalent," |
| 480 |
puts "** return the codepoint corresponding to the lower case version." |
| 481 |
puts "** Otherwise, return a copy of the argument." |
| 482 |
puts "**" |
| 483 |
puts "** The results are undefined if the value passed to this function" |
| 484 |
puts "** is less than zero." |
| 485 |
puts "*/" |
| 486 |
puts "int ${zFunc}\(int c, int eRemoveDiacritic)\{" |
| 487 |
|
| 488 |
set liOff [tl_generate_ioff_table $lRecord] |
| 489 |
tl_print_table_header |
| 490 |
foreach entry $lRecord { |
| 491 |
if {[tl_print_table_entry toggle $entry $liOff]} { |
| 492 |
lappend lHigh $entry |
| 493 |
} |
| 494 |
} |
| 495 |
tl_print_table_footer toggle |
| 496 |
tl_print_ioff_table $liOff |
| 497 |
|
| 498 |
puts [subst -nocommands { |
| 499 |
int ret = c; |
| 500 |
|
| 501 |
assert( sizeof(unsigned short)==2 && sizeof(unsigned char)==1 ); |
| 502 |
|
| 503 |
if( c<128 ){ |
| 504 |
if( c>='A' && c<='Z' ) ret = c + ('a' - 'A'); |
| 505 |
}else if( c<65536 ){ |
| 506 |
const struct TableEntry *p; |
| 507 |
int iHi = sizeof(aEntry)/sizeof(aEntry[0]) - 1; |
| 508 |
int iLo = 0; |
| 509 |
int iRes = -1; |
| 510 |
|
| 511 |
assert( c>aEntry[0].iCode ); |
| 512 |
while( iHi>=iLo ){ |
| 513 |
int iTest = (iHi + iLo) / 2; |
| 514 |
int cmp = (c - aEntry[iTest].iCode); |
| 515 |
if( cmp>=0 ){ |
| 516 |
iRes = iTest; |
| 517 |
iLo = iTest+1; |
| 518 |
}else{ |
| 519 |
iHi = iTest-1; |
| 520 |
} |
| 521 |
} |
| 522 |
|
| 523 |
assert( iRes>=0 && c>=aEntry[iRes].iCode ); |
| 524 |
p = &aEntry[iRes]; |
| 525 |
if( c<(p->iCode + p->nRange) && 0==(0x01 & p->flags & (p->iCode ^ c)) ){ |
| 526 |
ret = (c + (aiOff[p->flags>>1])) & 0x0000FFFF; |
| 527 |
assert( ret>0 ); |
| 528 |
} |
| 529 |
|
| 530 |
if( eRemoveDiacritic ){ |
| 531 |
ret = ${::remove_diacritic}(ret, eRemoveDiacritic==2); |
| 532 |
} |
| 533 |
} |
| 534 |
}] |
| 535 |
|
| 536 |
foreach entry $lHigh { |
| 537 |
tl_print_if_entry $entry |
| 538 |
} |
| 539 |
|
| 540 |
puts "" |
| 541 |
puts " return ret;" |
| 542 |
puts "\}" |
| 543 |
} |
| 544 |
|
| 545 |
proc code {txt} { |
| 546 |
set txt [string trimright $txt] |
| 547 |
set txt [string trimleft $txt "\n"] |
| 548 |
set n [expr {[string length $txt] - [string length [string trim $txt]]}] |
| 549 |
set ret "" |
| 550 |
foreach L [split $txt "\n"] { |
| 551 |
append ret "[string range $L $n end]\n" |
| 552 |
} |
| 553 |
return [uplevel "subst -nocommands {$ret}"] |
| 554 |
} |
| 555 |
|
| 556 |
proc intarray {lInt} { |
| 557 |
set ret "" |
| 558 |
set n [llength $lInt] |
| 559 |
for {set i 0} {$i < $n} {incr i 10} { |
| 560 |
append ret "\n " |
| 561 |
foreach int [lrange $lInt $i [expr $i+9]] { |
| 562 |
append ret [format "%-7s" "$int, "] |
| 563 |
} |
| 564 |
} |
| 565 |
append ret "\n " |
| 566 |
set ret |
| 567 |
} |
| 568 |
|
| 569 |
proc categories_switch {Cvar first lSecond} { |
| 570 |
upvar $Cvar C |
| 571 |
set ret "" |
| 572 |
append ret "case '$first':\n" |
| 573 |
append ret " switch( zCat\[1\] ){\n" |
| 574 |
foreach s $lSecond { |
| 575 |
append ret " case '$s': aArray\[$C($first$s)\] = 1; break;\n" |
| 576 |
} |
| 577 |
append ret " case '*': \n" |
| 578 |
foreach s $lSecond { |
| 579 |
append ret " aArray\[$C($first$s)\] = 1;\n" |
| 580 |
} |
| 581 |
append ret " break;\n" |
| 582 |
append ret " default: return 1;" |
| 583 |
append ret " }\n" |
| 584 |
append ret " break;\n" |
| 585 |
} |
| 586 |
|
| 587 |
# Argument is a list. Each element of which is itself a list of two elements: |
| 588 |
# |
| 589 |
# * the codepoint |
| 590 |
# * the category |
| 591 |
# |
| 592 |
# List elements are sorted in order of codepoint. |
| 593 |
# |
| 594 |
proc print_categories {lMap} { |
| 595 |
set categories { |
| 596 |
Cc Cf Cn Cs |
| 597 |
Ll Lm Lo Lt Lu |
| 598 |
Mc Me Mn |
| 599 |
Nd Nl No |
| 600 |
Pc Pd Pe Pf Pi Po Ps |
| 601 |
Sc Sk Sm So |
| 602 |
Zl Zp Zs |
| 603 |
|
| 604 |
LC Co |
| 605 |
} |
| 606 |
|
| 607 |
for {set i 0} {$i < [llength $categories]} {incr i} { |
| 608 |
set C([lindex $categories $i]) [expr 1+$i] |
| 609 |
} |
| 610 |
|
| 611 |
set caseC [categories_switch C C {c f n s o}] |
| 612 |
set caseL [categories_switch C L {l m o t u C}] |
| 613 |
set caseM [categories_switch C M {c e n}] |
| 614 |
set caseN [categories_switch C N {d l o}] |
| 615 |
set caseP [categories_switch C P {c d e f i o s}] |
| 616 |
set caseS [categories_switch C S {c k m o}] |
| 617 |
set caseZ [categories_switch C Z {l p s}] |
| 618 |
|
| 619 |
set nCat [expr [llength [array names C]] + 1] |
| 620 |
puts [code { |
| 621 |
int sqlite3Fts5UnicodeCatParse(const char *zCat, u8 *aArray){ |
| 622 |
aArray[0] = 1; |
| 623 |
switch( zCat[0] ){ |
| 624 |
$caseC |
| 625 |
$caseL |
| 626 |
$caseM |
| 627 |
$caseN |
| 628 |
$caseP |
| 629 |
$caseS |
| 630 |
$caseZ |
| 631 |
|
| 632 |
default: |
| 633 |
return 1; |
| 634 |
} |
| 635 |
return 0; |
| 636 |
} |
| 637 |
}] |
| 638 |
|
| 639 |
set nRepeat 0 |
| 640 |
set first [lindex $lMap 0 0] |
| 641 |
set class [lindex $lMap 0 1] |
| 642 |
set prev -1 |
| 643 |
|
| 644 |
set CASE(0) "Lu" |
| 645 |
set CASE(1) "Ll" |
| 646 |
|
| 647 |
foreach m $lMap { |
| 648 |
foreach {codepoint cl} $m {} |
| 649 |
set codepoint [expr "0x$codepoint"] |
| 650 |
if {$codepoint>=(1<<20)} continue |
| 651 |
|
| 652 |
set bNew 0 |
| 653 |
if {$codepoint!=($prev+1)} { |
| 654 |
set bNew 1 |
| 655 |
} elseif { |
| 656 |
$cl==$class || ($class=="LC" && $cl==$CASE([expr $nRepeat & 0x01])) |
| 657 |
} { |
| 658 |
incr nRepeat |
| 659 |
} elseif {$class=="Lu" && $nRepeat==1 && $cl=="Ll"} { |
| 660 |
set class LC |
| 661 |
incr nRepeat |
| 662 |
} else { |
| 663 |
set bNew 1 |
| 664 |
} |
| 665 |
if {$bNew} { |
| 666 |
lappend lEntries [list $first $class $nRepeat] |
| 667 |
set nRepeat 1 |
| 668 |
set first $codepoint |
| 669 |
set class $cl |
| 670 |
} |
| 671 |
set prev $codepoint |
| 672 |
} |
| 673 |
if {$nRepeat>0} { |
| 674 |
lappend lEntries [list $first $class $nRepeat] |
| 675 |
} |
| 676 |
|
| 677 |
set aBlock [list 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] |
| 678 |
set aMap [list] |
| 679 |
foreach e $lEntries { |
| 680 |
foreach {cp class nRepeat} $e {} |
| 681 |
set block [expr ($cp>>16)] |
| 682 |
if {$block>0 && [lindex $aBlock $block]==0} { |
| 683 |
for {set i 1} {$i<=$block} {incr i} { |
| 684 |
if {[lindex $aBlock $i]==0} { |
| 685 |
lset aBlock $i [llength $aMap] |
| 686 |
} |
| 687 |
} |
| 688 |
} |
| 689 |
lappend aMap [expr {$cp & 0xFFFF}] |
| 690 |
lappend aData [expr {($nRepeat << 5) + $C($class)}] |
| 691 |
} |
| 692 |
for {set i 1} {$i<[llength $aBlock]} {incr i} { |
| 693 |
if {[lindex $aBlock $i]==0} { |
| 694 |
lset aBlock $i [llength $aMap] |
| 695 |
} |
| 696 |
} |
| 697 |
|
| 698 |
set aBlockArray [intarray $aBlock] |
| 699 |
set aMapArray [intarray $aMap] |
| 700 |
set aDataArray [intarray $aData] |
| 701 |
puts [code { |
| 702 |
static u16 aFts5UnicodeBlock[] = {$aBlockArray}; |
| 703 |
static u16 aFts5UnicodeMap[] = {$aMapArray}; |
| 704 |
static u16 aFts5UnicodeData[] = {$aDataArray}; |
| 705 |
|
| 706 |
int sqlite3Fts5UnicodeCategory(u32 iCode) { |
| 707 |
int iRes = -1; |
| 708 |
int iHi; |
| 709 |
int iLo; |
| 710 |
int ret; |
| 711 |
u16 iKey; |
| 712 |
|
| 713 |
if( iCode>=(1<<20) ){ |
| 714 |
return 0; |
| 715 |
} |
| 716 |
iLo = aFts5UnicodeBlock[(iCode>>16)]; |
| 717 |
iHi = aFts5UnicodeBlock[1+(iCode>>16)]; |
| 718 |
iKey = (iCode & 0xFFFF); |
| 719 |
while( iHi>iLo ){ |
| 720 |
int iTest = (iHi + iLo) / 2; |
| 721 |
assert( iTest>=iLo && iTest<iHi ); |
| 722 |
if( iKey>=aFts5UnicodeMap[iTest] ){ |
| 723 |
iRes = iTest; |
| 724 |
iLo = iTest+1; |
| 725 |
}else{ |
| 726 |
iHi = iTest; |
| 727 |
} |
| 728 |
} |
| 729 |
|
| 730 |
if( iRes<0 ) return 0; |
| 731 |
if( iKey>=(aFts5UnicodeMap[iRes]+(aFts5UnicodeData[iRes]>>5)) ) return 0; |
| 732 |
ret = aFts5UnicodeData[iRes] & 0x1F; |
| 733 |
if( ret!=$C(LC) ) return ret; |
| 734 |
return ((iKey - aFts5UnicodeMap[iRes]) & 0x01) ? $C(Ll) : $C(Lu); |
| 735 |
} |
| 736 |
|
| 737 |
void sqlite3Fts5UnicodeAscii(u8 *aArray, u8 *aAscii){ |
| 738 |
int i = 0; |
| 739 |
int iTbl = 0; |
| 740 |
while( i<128 ){ |
| 741 |
int bToken = aArray[ aFts5UnicodeData[iTbl] & 0x1F ]; |
| 742 |
int n = (aFts5UnicodeData[iTbl] >> 5) + i; |
| 743 |
for(; i<128 && i<n; i++){ |
| 744 |
aAscii[i] = (u8)bToken; |
| 745 |
} |
| 746 |
iTbl++; |
| 747 |
} |
| 748 |
aAscii[0] = 0; /* 0x00 is never a token character */ |
| 749 |
} |
| 750 |
}] |
| 751 |
} |
| 752 |
|
| 753 |
proc print_test_categories {lMap} { |
| 754 |
|
| 755 |
set lCP [list] |
| 756 |
foreach e $lMap { |
| 757 |
foreach {cp cat} $e {} |
| 758 |
if {[expr 0x$cp] < (1<<20)} { |
| 759 |
lappend lCP "{0x$cp, \"$cat\"}, " |
| 760 |
} |
| 761 |
} |
| 762 |
|
| 763 |
set aCP "\n" |
| 764 |
for {set i 0} {$i < [llength $lCP]} {incr i 4} { |
| 765 |
append aCP " [join [lrange $lCP $i $i+3]]\n" |
| 766 |
} |
| 767 |
|
| 768 |
|
| 769 |
puts [code { |
| 770 |
static int categories_test (int *piCode){ |
| 771 |
struct Codepoint { |
| 772 |
int iCode; |
| 773 |
const char *zCat; |
| 774 |
} aCP[] = {$aCP}; |
| 775 |
int i; |
| 776 |
int iCP = 0; |
| 777 |
|
| 778 |
for(i=0; i<1000000; i++){ |
| 779 |
u8 aArray[40]; |
| 780 |
int cat = 0; |
| 781 |
int c = 0; |
| 782 |
memset(aArray, 0, sizeof(aArray)); |
| 783 |
if( aCP[iCP].iCode==i ){ |
| 784 |
sqlite3Fts5UnicodeCatParse(aCP[iCP].zCat, aArray); |
| 785 |
iCP++; |
| 786 |
}else{ |
| 787 |
aArray[0] = 1; |
| 788 |
} |
| 789 |
|
| 790 |
c = sqlite3Fts5UnicodeCategory((u32)i); |
| 791 |
if( aArray[c]==0 ){ |
| 792 |
*piCode = i; |
| 793 |
return 1; |
| 794 |
} |
| 795 |
} |
| 796 |
|
| 797 |
return 0; |
| 798 |
} |
| 799 |
}] |
| 800 |
} |
| 801 |
|
| 802 |
proc print_fold_test {zFunc mappings} { |
| 803 |
global tl_lookup_table |
| 804 |
|
| 805 |
foreach m $mappings { |
| 806 |
set c [lindex $m 1] |
| 807 |
if {$c == ""} { |
| 808 |
set extra([lindex $m 0]) 0 |
| 809 |
} else { |
| 810 |
scan $c %c i |
| 811 |
set extra([lindex $m 0]) $i |
| 812 |
} |
| 813 |
} |
| 814 |
|
| 815 |
puts "static int fold_test(int *piCode)\{" |
| 816 |
puts -nonewline " static int aLookup\[\] = \{" |
| 817 |
for {set i 0} {$i < 70000} {incr i} { |
| 818 |
|
| 819 |
set expected $i |
| 820 |
catch { set expected $tl_lookup_table($i) } |
| 821 |
set expected2 $expected |
| 822 |
catch { set expected2 $extra($expected2) } |
| 823 |
|
| 824 |
if {($i % 4)==0} { puts "" ; puts -nonewline " " } |
| 825 |
puts -nonewline "$expected, $expected2, " |
| 826 |
} |
| 827 |
puts " \};" |
| 828 |
puts " int i;" |
| 829 |
puts " for(i=0; i<sizeof(aLookup)/sizeof(aLookup\[0\]); i++)\{" |
| 830 |
puts " int iCode = (i/2);" |
| 831 |
puts " int bFlag = i & 0x0001;" |
| 832 |
puts " if( ${zFunc}\(iCode, bFlag)!=aLookup\[i\] )\{" |
| 833 |
puts " *piCode = iCode;" |
| 834 |
puts " return 1;" |
| 835 |
puts " \}" |
| 836 |
puts " \}" |
| 837 |
puts " return 0;" |
| 838 |
puts "\}" |
| 839 |
} |
| 840 |
|
| 841 |
|
| 842 |
proc print_fileheader {} { |
| 843 |
puts [string trim { |
| 844 |
/* |
| 845 |
** 2012-05-25 |
| 846 |
** |
| 847 |
** The author disclaims copyright to this source code. In place of |
| 848 |
** a legal notice, here is a blessing: |
| 849 |
** |
| 850 |
** May you do good and not evil. |
| 851 |
** May you find forgiveness for yourself and forgive others. |
| 852 |
** May you share freely, never taking more than you give. |
| 853 |
** |
| 854 |
****************************************************************************** |
| 855 |
*/ |
| 856 |
|
| 857 |
/* |
| 858 |
** DO NOT EDIT THIS MACHINE GENERATED FILE. |
| 859 |
*/ |
| 860 |
}] |
| 861 |
puts "" |
| 862 |
if {$::generate_fts5_code} { |
| 863 |
# no-op |
| 864 |
} else { |
| 865 |
puts "#ifndef SQLITE_DISABLE_FTS3_UNICODE" |
| 866 |
puts "#if defined(SQLITE_ENABLE_FTS3) || defined(SQLITE_ENABLE_FTS4)" |
| 867 |
} |
| 868 |
puts "" |
| 869 |
puts "#include <assert.h>" |
| 870 |
puts "" |
| 871 |
} |
| 872 |
|
| 873 |
proc print_test_main {} { |
| 874 |
puts "" |
| 875 |
puts "#include <stdio.h>" |
| 876 |
puts "" |
| 877 |
puts "int main(int argc, char **argv)\{" |
| 878 |
puts " int r1, r2, r3;" |
| 879 |
puts " int code;" |
| 880 |
puts " r3 = 0;" |
| 881 |
puts " r1 = isalnum_test(&code);" |
| 882 |
puts " if( r1 ) printf(\"isalnum(): Problem with code %d\\n\",code);" |
| 883 |
puts " else printf(\"isalnum(): test passed\\n\");" |
| 884 |
puts " r2 = fold_test(&code);" |
| 885 |
puts " if( r2 ) printf(\"fold(): Problem with code %d\\n\",code);" |
| 886 |
puts " else printf(\"fold(): test passed\\n\");" |
| 887 |
if {$::generate_fts5_code} { |
| 888 |
puts " r3 = categories_test(&code);" |
| 889 |
puts " if( r3 ) printf(\"categories(): Problem with code %d\\n\",code);" |
| 890 |
puts " else printf(\"categories(): test passed\\n\");" |
| 891 |
} |
| 892 |
puts " return (r1 || r2 || r3);" |
| 893 |
puts "\}" |
| 894 |
} |
| 895 |
|
| 896 |
# Process the command line arguments. Exit early if they are not to |
| 897 |
# our liking. |
| 898 |
# |
| 899 |
proc usage {} { |
| 900 |
puts -nonewline stderr "Usage: $::argv0 ?-test? ?-fts5? " |
| 901 |
puts stderr "<CaseFolding.txt file> <UnicodeData.txt file>" |
| 902 |
exit 1 |
| 903 |
} |
| 904 |
if {[llength $argv]<2} usage |
| 905 |
set unicodedata.txt [lindex $argv end] |
| 906 |
set casefolding.txt [lindex $argv end-1] |
| 907 |
|
| 908 |
set remove_diacritic remove_diacritic |
| 909 |
set generate_test_code 0 |
| 910 |
set generate_fts5_code 0 |
| 911 |
set function_prefix "sqlite3Fts" |
| 912 |
for {set i 0} {$i < [llength $argv]-2} {incr i} { |
| 913 |
switch -- [lindex $argv $i] { |
| 914 |
-test { |
| 915 |
set generate_test_code 1 |
| 916 |
} |
| 917 |
-fts5 { |
| 918 |
set function_prefix sqlite3Fts5 |
| 919 |
set generate_fts5_code 1 |
| 920 |
set remove_diacritic fts5_remove_diacritic |
| 921 |
} |
| 922 |
default { |
| 923 |
usage |
| 924 |
} |
| 925 |
} |
| 926 |
} |
| 927 |
|
| 928 |
print_fileheader |
| 929 |
|
| 930 |
if {$::generate_test_code} { |
| 931 |
puts "typedef unsigned short int u16;" |
| 932 |
puts "typedef unsigned char u8;" |
| 933 |
puts "#include <string.h>" |
| 934 |
} |
| 935 |
|
| 936 |
# Print the isalnum() function to stdout. |
| 937 |
# |
| 938 |
set lRange [an_load_separator_ranges] |
| 939 |
if {$generate_fts5_code==0} { |
| 940 |
print_isalnum ${function_prefix}UnicodeIsalnum $lRange |
| 941 |
} |
| 942 |
|
| 943 |
# Leave a gap between the two generated C functions. |
| 944 |
# |
| 945 |
puts "" |
| 946 |
puts "" |
| 947 |
|
| 948 |
# Load the fold data. This is used by the [rd_XXX] commands |
| 949 |
# as well as [print_fold]. |
| 950 |
tl_load_casefolding_txt ${casefolding.txt} |
| 951 |
|
| 952 |
set mappings [rd_load_unicodedata_text ${unicodedata.txt}] |
| 953 |
print_rd $mappings |
| 954 |
puts "" |
| 955 |
puts "" |
| 956 |
print_isdiacritic ${function_prefix}UnicodeIsdiacritic $mappings |
| 957 |
puts "" |
| 958 |
puts "" |
| 959 |
|
| 960 |
# Print the fold() function to stdout. |
| 961 |
# |
| 962 |
print_fold ${function_prefix}UnicodeFold |
| 963 |
|
| 964 |
if {$generate_fts5_code} { |
| 965 |
puts "" |
| 966 |
puts "" |
| 967 |
print_categories [cc_load_unicodedata_text ${unicodedata.txt}] |
| 968 |
} |
| 969 |
|
| 970 |
# Print the test routines and main() function to stdout, if -test |
| 971 |
# was specified. |
| 972 |
# |
| 973 |
if {$::generate_test_code} { |
| 974 |
if {$generate_fts5_code==0} { |
| 975 |
print_test_isalnum ${function_prefix}UnicodeIsalnum $lRange |
| 976 |
} |
| 977 |
print_fold_test ${function_prefix}UnicodeFold $mappings |
| 978 |
print_test_categories [cc_load_unicodedata_text ${unicodedata.txt}] |
| 979 |
print_test_main |
| 980 |
} |
| 981 |
|
| 982 |
if {$generate_fts5_code} { |
| 983 |
# no-op |
| 984 |
} else { |
| 985 |
puts "#endif /* defined(SQLITE_ENABLE_FTS3) || defined(SQLITE_ENABLE_FTS4) */" |
| 986 |
puts "#endif /* !defined(SQLITE_DISABLE_FTS3_UNICODE) */" |
| 987 |
} |