1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
4 ********************************************************************************
6 * Copyright (C) 1998-2015, International Business Machines
7 * Corporation and others. All Rights Reserved.
9 ********************************************************************************
13 * tool creating a binary (compressed) representation of the conversion mapping
14 * table (IBM NLTC ucmap format).
16 * 05/04/2000 helena Added fallback mapping into the picture...
17 * 06/29/2000 helena Major rewrite of the callback APIs.
21 #include "unicode/putil.h"
22 #include "unicode/ucnv_err.h"
33 #include "unicode/udata.h"
42 typedef struct ConvData
{
44 NewConverter
*cnvData
, *extData
;
45 UConverterSharedData sharedData
;
46 UConverterStaticData staticData
;
50 initConvData(ConvData
*data
) {
51 uprv_memset(data
, 0, sizeof(ConvData
));
52 data
->sharedData
.structSize
=sizeof(UConverterSharedData
);
53 data
->staticData
.structSize
=sizeof(UConverterStaticData
);
54 data
->sharedData
.staticData
=&data
->staticData
;
58 cleanupConvData(ConvData
*data
) {
60 if(data
->cnvData
!=NULL
) {
61 data
->cnvData
->close(data
->cnvData
);
64 if(data
->extData
!=NULL
) {
65 data
->extData
->close(data
->extData
);
74 * from ucnvstat.c - static prototypes of data-based converters
76 U_CAPI
const UConverterStaticData
* ucnv_converterStaticData
[UCNV_NUMBER_OF_SUPPORTED_CONVERTER_TYPES
];
81 UBool VERBOSE
= FALSE
;
84 UBool IGNORE_SISO_CHECK
= FALSE
;
87 createConverter(ConvData
*data
, const char* converterName
, UErrorCode
*pErrorCode
);
90 * Set up the UNewData and write the converter..
93 writeConverterData(ConvData
*data
, const char *cnvName
, const char *cnvDir
, UErrorCode
*status
);
95 UBool haveCopyright
=TRUE
;
97 static UDataInfo dataInfo
={
106 {0x63, 0x6e, 0x76, 0x74}, /* dataFormat="cnvt" */
107 {6, 2, 0, 0}, /* formatVersion */
108 {0, 0, 0, 0} /* dataVersion (calculated at runtime) */
112 writeConverterData(ConvData
*data
, const char *cnvName
, const char *cnvDir
, UErrorCode
*status
)
114 UNewDataMemory
*mem
= NULL
;
119 if(U_FAILURE(*status
))
124 tableType
=TABLE_NONE
;
125 if(data
->cnvData
!=NULL
) {
126 tableType
|=TABLE_BASE
;
128 if(data
->extData
!=NULL
) {
129 tableType
|=TABLE_EXT
;
132 mem
= udata_create(cnvDir
, "cnv", cnvName
, &dataInfo
, haveCopyright
? U_COPYRIGHT_STRING
: NULL
, status
);
134 if(U_FAILURE(*status
))
136 fprintf(stderr
, "Couldn't create the udata %s.%s: %s\n",
139 u_errorName(*status
));
145 printf("- Opened udata %s.%s\n", cnvName
, "cnv");
149 /* all read only, clean, platform independent data. Mmmm. :) */
150 udata_writeBlock(mem
, &data
->staticData
, sizeof(UConverterStaticData
));
151 size
+= sizeof(UConverterStaticData
); /* Is 4-aligned - by size */
152 /* Now, write the table */
153 if(tableType
&TABLE_BASE
) {
154 size
+= data
->cnvData
->write(data
->cnvData
, &data
->staticData
, mem
, tableType
);
156 if(tableType
&TABLE_EXT
) {
157 size
+= data
->extData
->write(data
->extData
, &data
->staticData
, mem
, tableType
);
160 sz2
= udata_finish(mem
, status
);
163 fprintf(stderr
, "error: wrote %u bytes to the .cnv file but counted %u bytes\n", (int)sz2
, (int)size
);
164 *status
=U_INTERNAL_PROGRAM_ERROR
;
168 printf("- Wrote %u bytes to the udata.\n", (int)sz2
);
174 OPT_HELP_QUESTION_MARK
,
180 OPT_IGNORE_SISO_CHECK
,
186 static UOption options
[]={
188 UOPTION_HELP_QUESTION_MARK
,
193 { "small", NULL
, NULL
, NULL
, '\1', UOPT_NO_ARG
, 0 },
194 { "ignore-siso-check", NULL
, NULL
, NULL
, '\1', UOPT_NO_ARG
, 0 },
198 int main(int argc
, char* argv
[])
201 char cnvName
[UCNV_MAX_FULL_FILE_NAME_LENGTH
];
203 U_MAIN_INIT_ARGS(argc
, argv
);
205 /* Set up the ICU version number */
206 UVersionInfo icuVersion
;
207 u_getVersion(icuVersion
);
208 uprv_memcpy(&dataInfo
.dataVersion
, &icuVersion
, sizeof(UVersionInfo
));
210 /* preset then read command line options */
211 options
[OPT_DESTDIR
].value
=u_getDataDirectory();
212 argc
=u_parseArgs(argc
, argv
, UPRV_LENGTHOF(options
), options
);
214 /* error handling, printing usage message */
217 "error in command line argument \"%s\"\n",
222 if(argc
<0 || options
[OPT_HELP_H
].doesOccur
|| options
[OPT_HELP_QUESTION_MARK
].doesOccur
) {
223 FILE *stdfile
=argc
<0 ? stderr
: stdout
;
225 "usage: %s [-options] files...\n"
226 "\tread .ucm codepage mapping files and write .cnv files\n"
228 "\t-h or -? or --help this usage text\n"
229 "\t-V or --version show a version message\n"
230 "\t-c or --copyright include a copyright notice\n"
231 "\t-d or --destdir destination directory, followed by the path\n"
232 "\t-v or --verbose Turn on verbose output\n"
233 "\t-q or --quiet do not display warnings and progress\n",
236 "\t --small Generate smaller .cnv files. They will be\n"
237 "\t significantly smaller but may not be compatible with\n"
238 "\t older versions of ICU and will require heap memory\n"
239 "\t allocation when loaded.\n"
240 "\t --ignore-siso-check Use SI/SO other than 0xf/0xe.\n");
241 return argc
<0 ? U_ILLEGAL_ARGUMENT_ERROR
: U_ZERO_ERROR
;
244 if(options
[OPT_VERSION
].doesOccur
) {
245 printf("makeconv version %u.%u, ICU tool to read .ucm codepage mapping files and write .cnv files\n",
246 dataInfo
.formatVersion
[0], dataInfo
.formatVersion
[1]);
247 printf("%s\n", U_COPYRIGHT_STRING
);
251 /* get the options values */
252 haveCopyright
= options
[OPT_COPYRIGHT
].doesOccur
;
253 const char *destdir
= options
[OPT_DESTDIR
].value
;
254 VERBOSE
= options
[OPT_VERBOSE
].doesOccur
;
255 QUIET
= options
[OPT_QUIET
].doesOccur
;
256 SMALL
= options
[OPT_SMALL
].doesOccur
;
258 if (options
[OPT_IGNORE_SISO_CHECK
].doesOccur
) {
259 IGNORE_SISO_CHECK
= TRUE
;
262 icu::CharString outFileName
;
263 UErrorCode err
= U_ZERO_ERROR
;
264 if (destdir
!= NULL
&& *destdir
!= 0) {
265 outFileName
.append(destdir
, err
).ensureEndsWithFileSeparator(err
);
266 if (U_FAILURE(err
)) {
270 int32_t outBasenameStart
= outFileName
.length();
275 printf("makeconv: processing %d files...\n", argc
- 1);
276 for(i
=1; i
<argc
; ++i
) {
277 printf("%s ", argv
[i
]);
284 UBool printFilename
= (UBool
) (argc
> 2 || VERBOSE
);
285 for (++argv
; --argc
; ++argv
)
287 UErrorCode localError
= U_ZERO_ERROR
;
288 const char *arg
= getLongPathname(*argv
);
290 /*produces the right destination path for display*/
291 outFileName
.truncate(outBasenameStart
);
292 if (outBasenameStart
!= 0)
294 /* find the last file sepator */
295 const char *basename
= findBasename(arg
);
296 outFileName
.append(basename
, localError
);
300 outFileName
.append(arg
, localError
);
302 if (U_FAILURE(localError
)) {
306 /*removes the extension if any is found*/
307 int32_t lastDotIndex
= outFileName
.lastIndexOf('.');
308 if (lastDotIndex
>= outBasenameStart
) {
309 outFileName
.truncate(lastDotIndex
);
312 /* the basename without extension is the converter name */
313 if ((outFileName
.length() - outBasenameStart
) >= UPRV_LENGTHOF(cnvName
)) {
314 fprintf(stderr
, "converter name %s too long\n", outFileName
.data() + outBasenameStart
);
315 return U_BUFFER_OVERFLOW_ERROR
;
317 uprv_strcpy(cnvName
, outFileName
.data() + outBasenameStart
);
319 /*Adds the target extension*/
320 outFileName
.append(CONVERTER_FILE_EXTENSION
, localError
);
321 if (U_FAILURE(localError
)) {
326 printf("makeconv: processing %s ...\n", arg
);
330 createConverter(&data
, arg
, &localError
);
332 if (U_FAILURE(localError
))
334 /* if an error is found, print out an error msg and keep going */
335 fprintf(stderr
, "Error creating converter for \"%s\" file for \"%s\" (%s)\n",
336 outFileName
.data(), arg
, u_errorName(localError
));
343 /* Insure the static data name matches the file name */
344 /* Changed to ignore directory and only compare base name
347 p
= strrchr(cnvName
, U_FILE_SEP_CHAR
); /* Find last file separator */
349 if(p
== NULL
) /* OK, try alternate */
351 p
= strrchr(cnvName
, U_FILE_ALT_SEP_CHAR
);
354 p
=cnvName
; /* If no separators, no problem */
359 p
++; /* If found separator, don't include it in compare */
361 if(uprv_stricmp(p
,data
.staticData
.name
) && !QUIET
)
363 fprintf(stderr
, "Warning: %s%s claims to be '%s'\n",
364 cnvName
, CONVERTER_FILE_EXTENSION
,
365 data
.staticData
.name
);
368 uprv_strcpy((char*)data
.staticData
.name
, cnvName
);
370 if(!uprv_isInvariantString((char*)data
.staticData
.name
, -1)) {
372 "Error: A converter name must contain only invariant characters.\n"
373 "%s is not a valid converter name.\n",
374 data
.staticData
.name
);
376 err
= U_INVALID_TABLE_FORMAT
;
380 localError
= U_ZERO_ERROR
;
381 writeConverterData(&data
, cnvName
, destdir
, &localError
);
383 if(U_FAILURE(localError
))
385 /* if an error is found, print out an error msg and keep going*/
386 fprintf(stderr
, "Error writing \"%s\" file for \"%s\" (%s)\n", outFileName
.data(), arg
,
387 u_errorName(localError
));
392 else if (printFilename
)
394 puts(outFileName
.data() + outBasenameStart
);
400 cleanupConvData(&data
);
407 getPlatformAndCCSIDFromName(const char *name
, int8_t *pPlatform
, int32_t *pCCSID
) {
408 if( (name
[0]=='i' || name
[0]=='I') &&
409 (name
[1]=='b' || name
[1]=='B') &&
410 (name
[2]=='m' || name
[2]=='M')
417 *pCCSID
=(int32_t)uprv_strtoul(name
, NULL
, 10);
419 *pPlatform
=UCNV_UNKNOWN
;
425 readHeader(ConvData
*data
,
426 FileStream
* convFile
,
427 UErrorCode
*pErrorCode
) {
429 char *s
, *key
, *value
;
430 const UConverterStaticData
*prototype
;
431 UConverterStaticData
*staticData
;
433 if(U_FAILURE(*pErrorCode
)) {
437 staticData
=&data
->staticData
;
438 staticData
->platform
=UCNV_IBM
;
439 staticData
->subCharLen
=0;
441 while(T_FileStream_readLine(convFile
, line
, sizeof(line
))) {
442 /* basic parsing and handling of state-related items */
443 if(ucm_parseHeaderLine(data
->ucm
, line
, &key
, &value
)) {
447 /* stop at the beginning of the mapping section */
448 if(uprv_strcmp(line
, "CHARMAP")==0) {
452 /* collect the information from the header field, ignore unknown keys */
453 if(uprv_strcmp(key
, "code_set_name")==0) {
455 uprv_strcpy((char *)staticData
->name
, value
);
456 getPlatformAndCCSIDFromName(value
, &staticData
->platform
, &staticData
->codepage
);
458 } else if(uprv_strcmp(key
, "subchar")==0) {
459 uint8_t bytes
[UCNV_EXT_MAX_BYTES
];
463 length
=ucm_parseBytes(bytes
, line
, (const char **)&s
);
464 if(1<=length
&& length
<=4 && *s
==0) {
465 staticData
->subCharLen
=length
;
466 uprv_memcpy(staticData
->subChar
, bytes
, length
);
468 fprintf(stderr
, "error: illegal <subchar> %s\n", value
);
469 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
472 } else if(uprv_strcmp(key
, "subchar1")==0) {
473 uint8_t bytes
[UCNV_EXT_MAX_BYTES
];
476 if(1==ucm_parseBytes(bytes
, line
, (const char **)&s
) && *s
==0) {
477 staticData
->subChar1
=bytes
[0];
479 fprintf(stderr
, "error: illegal <subchar1> %s\n", value
);
480 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
486 /* copy values from the UCMFile to the static data */
487 staticData
->maxBytesPerChar
=(int8_t)data
->ucm
->states
.maxCharLength
;
488 staticData
->minBytesPerChar
=(int8_t)data
->ucm
->states
.minCharLength
;
489 staticData
->conversionType
=data
->ucm
->states
.conversionType
;
491 if(staticData
->conversionType
==UCNV_UNSUPPORTED_CONVERTER
) {
492 fprintf(stderr
, "ucm error: missing conversion type (<uconv_class>)\n");
493 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
498 * Now that we know the type, copy any 'default' values from the table.
499 * We need not check the type any further because the parser only
500 * recognizes what we have prototypes for.
502 * For delta (extension-only) tables, copy values from the base file
503 * instead, see createConverter().
505 if(data
->ucm
->baseName
[0]==0) {
506 prototype
=ucnv_converterStaticData
[staticData
->conversionType
];
507 if(prototype
!=NULL
) {
508 if(staticData
->name
[0]==0) {
509 uprv_strcpy((char *)staticData
->name
, prototype
->name
);
512 if(staticData
->codepage
==0) {
513 staticData
->codepage
=prototype
->codepage
;
516 if(staticData
->platform
==0) {
517 staticData
->platform
=prototype
->platform
;
520 if(staticData
->minBytesPerChar
==0) {
521 staticData
->minBytesPerChar
=prototype
->minBytesPerChar
;
524 if(staticData
->maxBytesPerChar
==0) {
525 staticData
->maxBytesPerChar
=prototype
->maxBytesPerChar
;
528 if(staticData
->subCharLen
==0) {
529 staticData
->subCharLen
=prototype
->subCharLen
;
530 if(prototype
->subCharLen
>0) {
531 uprv_memcpy(staticData
->subChar
, prototype
->subChar
, prototype
->subCharLen
);
537 if(data
->ucm
->states
.outputType
<0) {
538 data
->ucm
->states
.outputType
=(int8_t)data
->ucm
->states
.maxCharLength
-1;
541 if( staticData
->subChar1
!=0 &&
542 (staticData
->minBytesPerChar
>1 ||
543 (staticData
->conversionType
!=UCNV_MBCS
&&
544 staticData
->conversionType
!=UCNV_EBCDIC_STATEFUL
))
546 fprintf(stderr
, "error: <subchar1> defined for a type other than MBCS or EBCDIC_STATEFUL\n");
547 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
551 /* return TRUE if a base table was read, FALSE for an extension table */
553 readFile(ConvData
*data
, const char* converterName
,
554 UErrorCode
*pErrorCode
) {
557 FileStream
*convFile
;
559 UCMStates
*baseStates
;
562 if(U_FAILURE(*pErrorCode
)) {
566 data
->ucm
=ucm_open();
568 convFile
=T_FileStream_open(converterName
, "r");
570 *pErrorCode
=U_FILE_ACCESS_ERROR
;
574 readHeader(data
, convFile
, pErrorCode
);
575 if(U_FAILURE(*pErrorCode
)) {
579 if(data
->ucm
->baseName
[0]==0) {
581 baseStates
=&data
->ucm
->states
;
582 ucm_processStates(baseStates
, IGNORE_SISO_CHECK
);
588 /* read the base table */
589 ucm_readTable(data
->ucm
, convFile
, dataIsBase
, baseStates
, pErrorCode
);
590 if(U_FAILURE(*pErrorCode
)) {
594 /* read an extension table if there is one */
595 while(T_FileStream_readLine(convFile
, line
, sizeof(line
))) {
596 end
=uprv_strchr(line
, 0);
598 (*(end
-1)=='\n' || *(end
-1)=='\r' || *(end
-1)==' ' || *(end
-1)=='\t')) {
603 if(line
[0]=='#' || u_skipWhitespace(line
)==end
) {
604 continue; /* ignore empty and comment lines */
607 if(0==uprv_strcmp(line
, "CHARMAP")) {
608 /* read the extension table */
609 ucm_readTable(data
->ucm
, convFile
, FALSE
, baseStates
, pErrorCode
);
611 fprintf(stderr
, "unexpected text after the base mapping table\n");
616 T_FileStream_close(convFile
);
618 if(data
->ucm
->base
->flagsType
==UCM_FLAGS_MIXED
|| data
->ucm
->ext
->flagsType
==UCM_FLAGS_MIXED
) {
619 fprintf(stderr
, "error: some entries have the mapping precision (with '|'), some do not\n");
620 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
627 createConverter(ConvData
*data
, const char *converterName
, UErrorCode
*pErrorCode
) {
631 UConverterStaticData
*staticData
;
632 UCMStates
*states
, *baseStates
;
634 if(U_FAILURE(*pErrorCode
)) {
640 dataIsBase
=readFile(data
, converterName
, pErrorCode
);
641 if(U_FAILURE(*pErrorCode
)) {
645 staticData
=&data
->staticData
;
646 states
=&data
->ucm
->states
;
650 * Build a normal .cnv file with a base table
651 * and an optional extension table.
653 data
->cnvData
=MBCSOpen(data
->ucm
);
654 if(data
->cnvData
==NULL
) {
655 *pErrorCode
=U_MEMORY_ALLOCATION_ERROR
;
657 } else if(!data
->cnvData
->isValid(data
->cnvData
,
658 staticData
->subChar
, staticData
->subCharLen
)
660 fprintf(stderr
, " the substitution character byte sequence is illegal in this codepage structure!\n");
661 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
663 } else if(staticData
->subChar1
!=0 &&
664 !data
->cnvData
->isValid(data
->cnvData
, &staticData
->subChar1
, 1)
666 fprintf(stderr
, " the subchar1 byte is illegal in this codepage structure!\n");
667 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
670 data
->ucm
->ext
->mappingsLength
>0 &&
671 !ucm_checkBaseExt(states
, data
->ucm
->base
, data
->ucm
->ext
, data
->ucm
->ext
, FALSE
)
673 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
674 } else if(data
->ucm
->base
->flagsType
&UCM_FLAGS_EXPLICIT
) {
675 /* sort the table so that it can be turned into UTF-8-friendly data */
676 ucm_sortTable(data
->ucm
->base
);
679 if(U_SUCCESS(*pErrorCode
)) {
681 /* add the base table after ucm_checkBaseExt()! */
682 !data
->cnvData
->addTable(data
->cnvData
, data
->ucm
->base
, &data
->staticData
)
684 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
687 * addTable() may have requested moving more mappings to the extension table
688 * if they fit into the base toUnicode table but not into the
689 * base fromUnicode table.
690 * (Especially for UTF-8-friendly fromUnicode tables.)
691 * Such mappings will have the MBCS_FROM_U_EXT_FLAG set, which causes them
692 * to be excluded from the extension toUnicode data.
693 * See MBCSOkForBaseFromUnicode() for which mappings do not fit into
694 * the base fromUnicode table.
696 ucm_moveMappings(data
->ucm
->base
, data
->ucm
->ext
);
697 ucm_sortTable(data
->ucm
->ext
);
698 if(data
->ucm
->ext
->mappingsLength
>0) {
699 /* prepare the extension table, if there is one */
700 data
->extData
=CnvExtOpen(data
->ucm
);
701 if(data
->extData
==NULL
) {
702 *pErrorCode
=U_MEMORY_ALLOCATION_ERROR
;
704 !data
->extData
->addTable(data
->extData
, data
->ucm
->ext
, &data
->staticData
)
706 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
712 /* Build an extension-only .cnv file. */
713 char baseFilename
[500];
716 initConvData(&baseData
);
718 /* assemble a path/filename for data->ucm->baseName */
719 uprv_strcpy(baseFilename
, converterName
);
720 basename
=(char *)findBasename(baseFilename
);
721 uprv_strcpy(basename
, data
->ucm
->baseName
);
722 uprv_strcat(basename
, ".ucm");
724 /* read the base table */
725 dataIsBase
=readFile(&baseData
, baseFilename
, pErrorCode
);
726 if(U_FAILURE(*pErrorCode
)) {
728 } else if(!dataIsBase
) {
729 fprintf(stderr
, "error: the <icu:base> file \"%s\" is not a base table file\n", baseFilename
);
730 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
732 /* prepare the extension table */
733 data
->extData
=CnvExtOpen(data
->ucm
);
734 if(data
->extData
==NULL
) {
735 *pErrorCode
=U_MEMORY_ALLOCATION_ERROR
;
737 /* fill in gaps in extension file header fields */
738 UCMapping
*m
, *mLimit
;
739 uint8_t fallbackFlags
;
741 baseStates
=&baseData
.ucm
->states
;
742 if(states
->conversionType
==UCNV_DBCS
) {
743 staticData
->minBytesPerChar
=(int8_t)(states
->minCharLength
=2);
744 } else if(states
->minCharLength
==0) {
745 staticData
->minBytesPerChar
=(int8_t)(states
->minCharLength
=baseStates
->minCharLength
);
747 if(states
->maxCharLength
<states
->minCharLength
) {
748 staticData
->maxBytesPerChar
=(int8_t)(states
->maxCharLength
=baseStates
->maxCharLength
);
751 if(staticData
->subCharLen
==0) {
752 uprv_memcpy(staticData
->subChar
, baseData
.staticData
.subChar
, 4);
753 staticData
->subCharLen
=baseData
.staticData
.subCharLen
;
756 * do not copy subChar1 -
757 * only use what is explicitly specified
758 * because it cannot be unset in the extension file header
761 /* get the fallback flags */
763 for(m
=baseData
.ucm
->base
->mappings
, mLimit
=m
+baseData
.ucm
->base
->mappingsLength
;
764 m
<mLimit
&& fallbackFlags
!=3;
774 if(fallbackFlags
&1) {
775 staticData
->hasFromUnicodeFallback
=TRUE
;
777 if(fallbackFlags
&2) {
778 staticData
->hasToUnicodeFallback
=TRUE
;
781 if(1!=ucm_countChars(baseStates
, staticData
->subChar
, staticData
->subCharLen
)) {
782 fprintf(stderr
, " the substitution character byte sequence is illegal in this codepage structure!\n");
783 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
785 } else if(staticData
->subChar1
!=0 && 1!=ucm_countChars(baseStates
, &staticData
->subChar1
, 1)) {
786 fprintf(stderr
, " the subchar1 byte is illegal in this codepage structure!\n");
787 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
790 !ucm_checkValidity(data
->ucm
->ext
, baseStates
) ||
791 !ucm_checkBaseExt(baseStates
, baseData
.ucm
->base
, data
->ucm
->ext
, data
->ucm
->ext
, FALSE
)
793 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
795 if(states
->maxCharLength
>1) {
797 * When building a normal .cnv file with a base table
798 * for an MBCS (not SBCS) table with explicit precision flags,
799 * the MBCSAddTable() function marks some mappings for moving
800 * to the extension table.
801 * They fit into the base toUnicode table but not into the
802 * base fromUnicode table.
803 * (Note: We do have explicit precision flags because they are
804 * required for extension table generation, and
805 * ucm_checkBaseExt() verified it.)
807 * We do not call MBCSAddTable() here (we probably could)
808 * so we need to do the analysis before building the extension table.
809 * We assume that MBCSAddTable() will build a UTF-8-friendly table.
810 * Redundant mappings in the extension table are ok except they cost some size.
812 * Do this after ucm_checkBaseExt().
814 const MBCSData
*mbcsData
=MBCSGetDummy();
816 for(m
=baseData
.ucm
->base
->mappings
, mLimit
=m
+baseData
.ucm
->base
->mappingsLength
;
820 if(!MBCSOkForBaseFromUnicode(mbcsData
, m
->b
.bytes
, m
->bLen
, m
->u
, m
->f
)) {
821 m
->f
|=MBCS_FROM_U_EXT_FLAG
;
822 m
->moveFlag
=UCM_MOVE_TO_EXT
;
828 ucm_moveMappings(baseData
.ucm
->base
, data
->ucm
->ext
);
829 ucm_sortTable(data
->ucm
->ext
);
832 if(!data
->extData
->addTable(data
->extData
, data
->ucm
->ext
, &data
->staticData
)) {
833 *pErrorCode
=U_INVALID_TABLE_FORMAT
;
839 cleanupConvData(&baseData
);
844 * Hey, Emacs, please set the following:
847 * indent-tabs-mode: nil