static void SplashLayout() {
unichar_t *start, *pt, *lastspace;
extern const char *source_modtime_str;
extern const char *source_version_str;
uc_strcpy(msg, "When my father finished his book on Renaissance printing (The Craft of Printing and the Publication of Shakespeare's Works) he told me that I would have to write the chapter on computer typography. This is my attempt to do so.");
GDrawSetFont(splashw,splash_font);
linecnt = 0;
lines[linecnt++] = msg-1;
for ( start = msg; *start!='\0'; start = pt ) {
lastspace = NULL;
for ( pt=start; ; ++pt ) {
if ( *pt==' ' || *pt=='\0' ) {
if ( GDrawGetTextWidth(splashw,start,pt-start)<splashimage.u.image->width-10 )
lastspace = pt;
else
break;
if ( *pt=='\0' )
break;
}
}
if ( lastspace!=NULL )
pt = lastspace;
lines[linecnt++] = pt;
if ( *pt ) ++pt;
}
uc_strcpy(pt, " FontForge used to be named PfaEdit.");
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcpy(pt," git hash: ");;
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcat(pt, FONTFORGE_GIT_VERSION);
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcpy(pt," Version: ");;
uc_strcat(pt,FONTFORGE_MODTIME_STR);
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcat(pt," (");
uc_strcat(pt,FONTFORGE_MODTIME_STR);
uc_strcat(pt,"-ML");
#ifdef FREETYPE_HAS_DEBUGGER
uc_strcat(pt,"-TtfDb");
#endif
uc_strcat(pt,")");
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcpy(pt," Lib Version: ");
uc_strcat(pt,FONTFORGE_MODTIME_STR);
lines[linecnt++] = pt+u_strlen(pt);
lines[linecnt] = NULL;
is = u_strchr(msg,'(');
ie = u_strchr(msg,')');
}
char *_GIO_decomposeURL(const unichar_t *url,char **host, int *port, char **username,
char **password) {
unichar_t *pt, *pt2, *upt, *ppt;
char *path;
char proto[40];
/* ftp://[user[:password]@]ftpserver[:port]/url-path */
*username = NULL; *password = NULL; *port = -1;
pt = uc_strstr(url,"://");
if ( pt==NULL ) {
*host = NULL;
return( cu_copy(url));
}
cu_strncpy(proto,url,(size_t)(pt-url)<sizeof(proto)?(size_t)(pt-url):sizeof(proto));
pt += 3;
pt2 = u_strchr(pt,'/');
if ( pt2==NULL ) {
pt2 = pt+u_strlen(pt);
path = copy("/");
} else {
path = cu_copy(pt2);
}
upt = u_strchr(pt,'@');
if ( upt!=NULL && upt<pt2 ) {
ppt = u_strchr(pt,':');
if ( ppt==NULL )
*username = cu_copyn(pt,upt-pt);
else {
*username = cu_copyn(pt,ppt-pt);
*password = cu_copyn(ppt+1,upt-ppt-1);
}
pt = upt+1;
}
ppt = u_strchr(pt,':');
if ( ppt!=NULL && ppt<pt2 ) {
char *temp = cu_copyn(ppt+1,pt2-ppt-1), *end;
*port = strtol(temp,&end,10);
if ( *end!='\0' )
*port = -2;
free(temp);
pt2 = ppt;
}
*host = cu_copyn(pt,pt2-pt);
if ( *username )
*password = GIO_PasswordCache(proto,*host,*username,*password);
return( path );
}
unichar_t *u_GFileNormalize(unichar_t *name) {
unichar_t *pt, *base, *ppt;
if ( (pt = uc_strstr(name,"://"))!=NULL ) {
base = u_strchr(pt+3,'/');
if ( base==NULL )
return( name );
++base;
} else if ( *name=='/' )
base = name+1;
else
base = name;
for ( pt=base; *pt!='\0'; ) {
if ( *pt=='/' )
u_strcpy(pt,pt+1);
else if ( uc_strncmp(pt,"./",2)==0 )
u_strcpy(pt,pt+2);
else if ( uc_strncmp(pt,"../",2)==0 ) {
for ( ppt=pt-2; ppt>=base && *ppt!='/'; --ppt );
++ppt;
if ( ppt>=base ) {
u_strcpy(ppt,pt+3);
pt = ppt;
} else
pt += 3;
} else {
while ( *pt!='/' && *pt!='\0' ) ++pt;
if ( *pt == '/' ) ++pt;
}
}
return( name );
}
/**
* This function allocates and returns a token_t structure corresponding to the given
* string.
*/
token_t* new_token_t(unichar* str) {
token_t* tok=(token_t*)malloc(sizeof(token_t));
if (tok==NULL) {
fatal_alloc_error("new_token_t");
}
for (const keyword_t* key=keywords;key->str!=NULL;key++) {
if (!u_strcmp(str, key->str)) {
/* If the token is a keyword */
tok->type=key->val;
tok->str= NULL;
tok->next=NULL;
return tok;
}
}
if (*str=='<') {
/* If we have a '<', we look for the ending '>' */
unichar* p=u_strchr(str,'>');
if (p==NULL || *(p+1)!='\0') {
fatal_error("Invalid token: '%S'\n",str);
}
*p='\0';
tok->type=TOK_ANGLE;
/* We copy the content between the angle brackets */
tok->str=u_strdup(str+1);
tok->next=NULL;
return tok;
}
/* Otherwise, we create a default token with the string */
tok->type=TOK_STR;
tok->str=u_strdup(str);
tok->next=NULL;
return tok;
}
const CompactTrieDictionary *
ICULanguageBreakFactory::loadDictionaryFor(UScriptCode script, int32_t /*breakType*/) {
UErrorCode status = U_ZERO_ERROR;
// Open root from brkitr tree.
char dictnbuff[256];
char ext[4]={'\0'};
UResourceBundle *b = ures_open(U_ICUDATA_BRKITR, "", &status);
b = ures_getByKeyWithFallback(b, "dictionaries", b, &status);
b = ures_getByKeyWithFallback(b, uscript_getShortName(script), b, &status);
int32_t dictnlength = 0;
const UChar *dictfname = ures_getString(b, &dictnlength, &status);
if (U_SUCCESS(status) && (size_t)dictnlength >= sizeof(dictnbuff)) {
dictnlength = 0;
status = U_BUFFER_OVERFLOW_ERROR;
}
if (U_SUCCESS(status) && dictfname) {
UChar* extStart=u_strchr(dictfname, 0x002e);
int len = 0;
if(extStart!=NULL){
len = extStart-dictfname;
u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
u_UCharsToChars(dictfname, dictnbuff, len);
}
dictnbuff[len]=0; // nul terminate
}
ures_close(b);
UDataMemory *file = udata_open(U_ICUDATA_BRKITR, ext, dictnbuff, &status);
if (U_SUCCESS(status)) {
const CompactTrieDictionary *dict = new CompactTrieDictionary(
file, status);
if (U_SUCCESS(status) && dict == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
}
if (U_FAILURE(status)) {
delete dict;
dict = NULL;
}
return dict;
} else if (dictfname != NULL){
//create dummy dict if dictionary filename not valid
UChar c = 0x0020;
status = U_ZERO_ERROR;
MutableTrieDictionary *mtd = new MutableTrieDictionary(c, status, TRUE);
mtd->addWord(&c, 1, status, 1);
return new CompactTrieDictionary(*mtd, status);
}
return NULL;
}
int msgformat_fix_quotes(UChar **spattern, uint32_t *spattern_len, UErrorCode *ec)
{
if(*spattern && *spattern_len && u_strchr(*spattern, (UChar)'\'')) {
UChar *npattern = safe_emalloc(sizeof(UChar)*2, *spattern_len, sizeof(UChar));
uint32_t npattern_len;
npattern_len = umsg_autoQuoteApostrophe(*spattern, *spattern_len, npattern, 2*(*spattern_len)+1, ec);
efree(*spattern);
if( U_FAILURE(*ec) )
{
return FAILURE;
}
npattern = erealloc(npattern, sizeof(UChar)*(npattern_len+1));
*spattern = npattern;
*spattern_len = npattern_len;
}
return SUCCESS;
}
U_CAPI UChar* U_EXPORT2
u_strchr32(const UChar* s, UChar32 c) {
if ((uint32_t) c <= U_BMP_MAX) {
/* find BMP code point */
return u_strchr(s, (UChar) c);
} else if ((uint32_t) c <= UCHAR_MAX_VALUE) {
/* find supplementary code point as surrogate pair */
UChar cs, lead = U16_LEAD(c), trail = U16_TRAIL(c);
while ((cs = *s++) != 0) {
if (cs == lead && *s == trail) {
return (UChar*) (s - 1);
}
}
return NULL;
} else {
/* not a Unicode code point, not findable */
return NULL;
}
}
WordListLine WordlistEscapedInputStringToParsedDataComplex(
SplineFont* sf,
const unichar_t* input_const,
WordlistEscapedInputStringToRealString_getFakeUnicodeOfScFunc getUnicodeFunc,
void* udata )
{
unichar_t* input = u_copy( input_const );
WordListChar* ret = calloc( WordListLineSz, sizeof(WordListChar));
WordListChar* out = ret;
unichar_t* in = input;
unichar_t* in_end = input + u_strlen(input);
// trim comment and beyond from input
{
unichar_t* p = input;
while( p && p < in_end )
{
p = u_strchr( p, '#' );
if( p > input && *(p-1) == '/' )
{
p++;
continue;
}
if( p )
*p = '\0';
break;
}
}
in_end = input + u_strlen(input);
int addingGlyphsToSelected = 0;
int currentGlyphIndex = -1;
for ( ; in < in_end; in++ )
{
unichar_t ch = *in;
TRACE("in:%p end:%p got char %d %c\n", in, in_end, ch, ch );
if( ch == '[' )
{
addingGlyphsToSelected = 1;
continue;
}
if( ch == ']' )
{
addingGlyphsToSelected = 0;
continue;
}
int isSelected = addingGlyphsToSelected;
currentGlyphIndex++;
if( ch == '/' || ch == '\\' )
{
// start of a glyph name
unichar_t glyphname[ PATH_MAX+1 ];
unichar_t* updated_in = 0;
SplineChar* sc = u_WordlistEscapedInputStringToRealString_readGlyphName( sf, in, in_end, &updated_in, glyphname );
if( sc )
{
in = updated_in;
int n = getUnicodeFunc( sc, udata );
if( n == -1 )
{
/*
* Okay, this probably means we've got an unencoded glyph (generally
* used for OpenType substitutions).
* Redeem the value from the SplineFont datamap instead of fetching from
* the Unicode identifier.
*/
n = sf->map->backmap[sc->orig_pos];
/*
* Unencoded glyphs have special mappings in the SplineFont that
* start from 65536 (values beyond Unicode, 65535 being the reserved
* "frontier" value).
*/
if ( (sf->map->enc->is_unicodebmp || sf->map->enc->is_unicodefull) && n < 65536 ) {
TRACE("ToRealString: backmapped position does not match Unicode encoding\n");
TRACE("orig_pos: %d, backmap: %d, attached unicode enc: %d\n", sc->orig_pos, n, sc->unicodeenc );
TRACE("ToRealString: INVALID CHAR POSITION, name: %s\n", sc->name );
}
}
out->sc = sc;
out->isSelected = isSelected;
out->currentGlyphIndex = currentGlyphIndex;
out->n = n;
out++;
/* out = utf8_idpb( out, n, 0 ); */
/* if( !out ) */
/* printf("ToRealString error on out\n"); */
continue;
}
}
/* If we reach this point, we're looking based on codepoint. */
SplineChar* sc = SFGetOrMakeChar( sf, (int)ch, 0 );
out->sc = sc;
out->isSelected = isSelected;
out->currentGlyphIndex = currentGlyphIndex;
out++;
}
free(input);
return(ret);
}
/* Instead of having a separate pass for 'special' patterns, reintegrate the two
* so we don't get bitten by preflight bugs again. We can be reasonably efficient
* without two separate code paths, this code isn't that performance-critical.
*
* This code is general enough to deal with patterns that have a prefix or swap the
* language and remainder components, since we gave developers enough rope to do such
* things if they futz with the pattern data. But since we don't give them a way to
* specify a pattern for arbitrary combinations of components, there's not much use in
* that. I don't think our data includes such patterns, the only variable I know if is
* whether there is a space before the open paren, or not. Oh, and zh uses different
* chars than the standard open/close paren (which ja and ko use, btw).
*/
U_CAPI int32_t U_EXPORT2
uloc_getDisplayName(const char *locale,
const char *displayLocale,
UChar *dest, int32_t destCapacity,
UErrorCode *pErrorCode)
{
static const UChar defaultSeparator[9] = { 0x007b, 0x0030, 0x007d, 0x002c, 0x0020, 0x007b, 0x0031, 0x007d, 0x0000 }; /* "{0}, {1}" */
static const UChar sub0[4] = { 0x007b, 0x0030, 0x007d , 0x0000 } ; /* {0} */
static const UChar sub1[4] = { 0x007b, 0x0031, 0x007d , 0x0000 } ; /* {1} */
static const int32_t subLen = 3;
static const UChar defaultPattern[10] = {
0x007b, 0x0030, 0x007d, 0x0020, 0x0028, 0x007b, 0x0031, 0x007d, 0x0029, 0x0000
}; /* {0} ({1}) */
static const int32_t defaultPatLen = 9;
static const int32_t defaultSub0Pos = 0;
static const int32_t defaultSub1Pos = 5;
int32_t length; /* of formatted result */
const UChar *separator;
int32_t sepLen = 0;
const UChar *pattern;
int32_t patLen = 0;
int32_t sub0Pos, sub1Pos;
UChar formatOpenParen = 0x0028; // (
UChar formatReplaceOpenParen = 0x005B; // [
UChar formatCloseParen = 0x0029; // )
UChar formatReplaceCloseParen = 0x005D; // ]
UBool haveLang = TRUE; /* assume true, set false if we find we don't have
a lang component in the locale */
UBool haveRest = TRUE; /* assume true, set false if we find we don't have
any other component in the locale */
UBool retry = FALSE; /* set true if we need to retry, see below */
int32_t langi = 0; /* index of the language substitution (0 or 1), virtually always 0 */
if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
return 0;
}
if(destCapacity<0 || (destCapacity>0 && dest==NULL)) {
*pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
{
UErrorCode status = U_ZERO_ERROR;
UResourceBundle* locbundle=ures_open(U_ICUDATA_LANG, displayLocale, &status);
UResourceBundle* dspbundle=ures_getByKeyWithFallback(locbundle, _kLocaleDisplayPattern,
NULL, &status);
separator=ures_getStringByKeyWithFallback(dspbundle, _kSeparator, &sepLen, &status);
pattern=ures_getStringByKeyWithFallback(dspbundle, _kPattern, &patLen, &status);
ures_close(dspbundle);
ures_close(locbundle);
}
/* If we couldn't find any data, then use the defaults */
if(sepLen == 0) {
separator = defaultSeparator;
}
/* #10244: Even though separator is now a pattern, it is awkward to handle it as such
* here since we are trying to build the display string in place in the dest buffer,
* and to handle it as a pattern would entail having separate storage for the
* substrings that need to be combined (the first of which may be the result of
* previous such combinations). So for now we continue to treat the portion between
* {0} and {1} as a string to be appended when joining substrings, ignoring anything
* that is before {0} or after {1} (no existing separator pattern has any such thing).
* This is similar to how pattern is handled below.
*/
{
UChar *p0=u_strstr(separator, sub0);
UChar *p1=u_strstr(separator, sub1);
if (p0==NULL || p1==NULL || p1<p0) {
*pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
separator = (const UChar *)p0 + subLen;
sepLen = p1 - separator;
}
if(patLen==0 || (patLen==defaultPatLen && !u_strncmp(pattern, defaultPattern, patLen))) {
pattern=defaultPattern;
patLen=defaultPatLen;
sub0Pos=defaultSub0Pos;
sub1Pos=defaultSub1Pos;
// use default formatOpenParen etc. set above
} else { /* non-default pattern */
UChar *p0=u_strstr(pattern, sub0);
UChar *p1=u_strstr(pattern, sub1);
if (p0==NULL || p1==NULL) {
*pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
sub0Pos=p0-pattern;
sub1Pos=p1-pattern;
if (sub1Pos < sub0Pos) { /* a very odd pattern */
int32_t t=sub0Pos; sub0Pos=sub1Pos; sub1Pos=t;
langi=1;
}
if (u_strchr(pattern, 0xFF08) != NULL) {
formatOpenParen = 0xFF08; // fullwidth (
formatReplaceOpenParen = 0xFF3B; // fullwidth [
formatCloseParen = 0xFF09; // fullwidth )
formatReplaceCloseParen = 0xFF3D; // fullwidth ]
}
}
/* We loop here because there is one case in which after the first pass we could need to
* reextract the data. If there's initial padding before the first element, we put in
* the padding and then write that element. If it turns out there's no second element,
* we didn't need the padding. If we do need the data (no preflight), and the first element
* would have fit but for the padding, we need to reextract. In this case (only) we
* adjust the parameters so padding is not added, and repeat.
*/
do {
UChar* p=dest;
int32_t patPos=0; /* position in the pattern, used for non-substitution portions */
int32_t langLen=0; /* length of language substitution */
int32_t langPos=0; /* position in output of language substitution */
int32_t restLen=0; /* length of 'everything else' substitution */
int32_t restPos=0; /* position in output of 'everything else' substitution */
UEnumeration* kenum = NULL; /* keyword enumeration */
/* prefix of pattern, extremely likely to be empty */
if(sub0Pos) {
if(destCapacity >= sub0Pos) {
while (patPos < sub0Pos) {
*p++ = pattern[patPos++];
}
} else {
patPos=sub0Pos;
}
length=sub0Pos;
} else {
length=0;
}
for(int32_t subi=0,resti=0;subi<2;) { /* iterate through patterns 0 and 1*/
UBool subdone = FALSE; /* set true when ready to move to next substitution */
/* prep p and cap for calls to get display components, pin cap to 0 since
they complain if cap is negative */
int32_t cap=destCapacity-length;
if (cap <= 0) {
cap=0;
} else {
p=dest+length;
}
if (subi == langi) { /* {0}*/
if(haveLang) {
langPos=length;
langLen=uloc_getDisplayLanguage(locale, displayLocale, p, cap, pErrorCode);
length+=langLen;
haveLang=langLen>0;
}
subdone=TRUE;
} else { /* {1} */
if(!haveRest) {
subdone=TRUE;
} else {
int32_t len; /* length of component (plus other stuff) we just fetched */
switch(resti++) {
case 0:
restPos=length;
len=uloc_getDisplayScriptInContext(locale, displayLocale, p, cap, pErrorCode);
break;
case 1:
len=uloc_getDisplayCountry(locale, displayLocale, p, cap, pErrorCode);
break;
case 2:
len=uloc_getDisplayVariant(locale, displayLocale, p, cap, pErrorCode);
break;
case 3:
kenum = uloc_openKeywords(locale, pErrorCode);
/* fall through */
default: {
const char* kw=uenum_next(kenum, &len, pErrorCode);
if (kw == NULL) {
uenum_close(kenum);
len=0; /* mark that we didn't add a component */
subdone=TRUE;
} else {
/* incorporating this behavior into the loop made it even more complex,
so just special case it here */
len = uloc_getDisplayKeyword(kw, displayLocale, p, cap, pErrorCode);
if(len) {
if(len < cap) {
p[len]=0x3d; /* '=', assume we'll need it */
}
len+=1;
/* adjust for call to get keyword */
cap-=len;
if(cap <= 0) {
cap=0;
} else {
p+=len;
}
}
/* reset for call below */
if(*pErrorCode == U_BUFFER_OVERFLOW_ERROR) {
*pErrorCode=U_ZERO_ERROR;
}
int32_t vlen = uloc_getDisplayKeywordValue(locale, kw, displayLocale,
p, cap, pErrorCode);
if(len) {
if(vlen==0) {
--len; /* remove unneeded '=' */
}
/* restore cap and p to what they were at start */
cap=destCapacity-length;
if(cap <= 0) {
cap=0;
} else {
p=dest+length;
}
}
len+=vlen; /* total we added for key + '=' + value */
}
} break;
} /* end switch */
if (len>0) {
/* we addeed a component, so add separator and write it if there's room. */
if(len+sepLen<=cap) {
const UChar * plimit = p + len;
for (; p < plimit; p++) {
if (*p == formatOpenParen) {
*p = formatReplaceOpenParen;
} else if (*p == formatCloseParen) {
*p = formatReplaceCloseParen;
}
}
for(int32_t i=0;i<sepLen;++i) {
*p++=separator[i];
}
}
length+=len+sepLen;
} else if(subdone) {
/* remove separator if we added it */
if (length!=restPos) {
length-=sepLen;
}
restLen=length-restPos;
haveRest=restLen>0;
}
}
}
if(*pErrorCode == U_BUFFER_OVERFLOW_ERROR) {
*pErrorCode=U_ZERO_ERROR;
}
if(subdone) {
if(haveLang && haveRest) {
/* append internal portion of pattern, the first time,
or last portion of pattern the second time */
int32_t padLen;
patPos+=subLen;
padLen=(subi==0 ? sub1Pos : patLen)-patPos;
if(length+padLen < destCapacity) {
p=dest+length;
for(int32_t i=0;i<padLen;++i) {
*p++=pattern[patPos++];
}
} else {
patPos+=padLen;
}
length+=padLen;
} else if(subi==0) {
/* don't have first component, reset for second component */
sub0Pos=0;
length=0;
} else if(length>0) {
/* true length is the length of just the component we got. */
length=haveLang?langLen:restLen;
if(dest && sub0Pos!=0) {
if (sub0Pos+length<=destCapacity) {
/* first component not at start of result,
but we have full component in buffer. */
u_memmove(dest, dest+(haveLang?langPos:restPos), length);
} else {
/* would have fit, but didn't because of pattern prefix. */
sub0Pos=0; /* stops initial padding (and a second retry,
so we won't end up here again) */
retry=TRUE;
}
}
}
++subi; /* move on to next substitution */
}
}
} while(retry);
return u_terminateUChars(dest, destCapacity, length, pErrorCode);
}
enum charset _GDraw_ParseMapping(unichar_t *setname) {
unichar_t *pt;
int val;
if ( uc_strstrmatch(setname,"iso")!=NULL && uc_strstrmatch(setname,"10646")!=NULL )
return( em_unicode );
else if ( uc_strstrmatch(setname,"UnicodePlane")!=NULL ) {
pt = u_strchr(setname,'-');
if ( pt==NULL )
return( em_uplane0+1 );
return( em_uplane0+u_strtol(pt+1,NULL,10) );
} else if ( uc_strstrmatch(setname,"unicode")!=NULL )
return( em_unicode );
#if 0
if ( uc_strstrmatch(setname,"ascii")!=NULL ||
( uc_strstrmatch(setname,"iso")!=NULL && uc_strstrmatch(setname,"646")!=NULL )) {
char *lang = getenv( "LANG" );
if ( lang==NULL || *lang=='\0' || (*lang=='e' && *lang=='n' ))
return( em_iso8859_1 ); /* ascii can masquarade as iso8859-1 for english speakers (no accents needed) */
}
#endif
if ( uc_strstrmatch(setname,"iso")!=NULL && uc_strstrmatch(setname,"8859")!=NULL ) {
pt = uc_strstrmatch(setname,"8859");
pt += 4;
if ( *pt=='-' ) ++pt;
if ( !isdigit(*pt) )
/* Bad */;
else if ( !isdigit(pt[1]) )
return( em_iso8859_1+*pt-'1' );
else {
val = (pt[0]-'0')*10 + pt[1]-'0';
switch ( val ) {
case 10: case 11:
return( em_iso8859_10+val-10 );
case 13: case 14: case 15:
return( em_iso8859_13+val-13 );
}
}
}
if ( uc_strstrmatch(setname,"latin1")!=NULL )
return( em_iso8859_1 );
else if ( uc_strstrmatch(setname,"latin2")!=NULL )
return( em_iso8859_2 );
else if ( uc_strstrmatch(setname,"latin3")!=NULL )
return( em_iso8859_3 );
else if ( uc_strstrmatch(setname,"latin4")!=NULL )
return( em_iso8859_4 );
else if ( uc_strstrmatch(setname,"latin5")!=NULL )
return( em_iso8859_9 );
else if ( uc_strstrmatch(setname,"latin6")!=NULL )
return( em_iso8859_10 );
else if ( uc_strstrmatch(setname,"latin7")!=NULL )
return( em_iso8859_13 );
else if ( uc_strstrmatch(setname,"latin8")!=NULL )
return( em_iso8859_14 );
else if ( uc_strstrmatch(setname,"latin0")!=NULL || uc_strstrmatch(setname,"latin9")!=NULL )
return( em_iso8859_15 );
if ( uc_strstrmatch(setname,"koi8")!=NULL )
return( em_koi8_r );
if ( uc_strstrmatch(setname,"cyrillic")!=NULL )
return( em_iso8859_5 ); /* This is grasping at straws */
else if ( uc_strstrmatch(setname,"greek")!=NULL )
return( em_iso8859_7 ); /* This is grasping at straws */
else if ( uc_strstrmatch(setname,"arabic")!=NULL )
return( em_iso8859_6 ); /* This is grasping at straws */
else if ( uc_strstrmatch(setname,"hebrew")!=NULL )
return( em_iso8859_8 ); /* This is grasping at straws */
else if ( uc_strstrmatch(setname,"thai")!=NULL || uc_strstrmatch(setname,"tis")!=NULL )
return( em_iso8859_11 );
if ( uc_strstrmatch(setname,"jis")!=NULL ) {
if ( uc_strstrmatch(setname,"201")!=NULL )
return( em_jis201 );
if ( uc_strstrmatch(setname,"208")!=NULL )
return( em_jis208 );
if ( uc_strstrmatch(setname,"212")!=NULL )
return( em_jis212 );
if ( uc_strstrmatch(setname,"213")!=NULL ) /* I don't support 213 */
return( em_none );
return( em_jis208 );
}
if ( uc_strstrmatch(setname,"ksc")!=NULL && uc_strstrmatch(setname,"5601")!=NULL )
return( em_ksc5601 ); /* Seem to be several versions of 5601, we want 94x94 */
if ( uc_strstrmatch(setname,"gb")!=NULL && uc_strstrmatch(setname,"2312")!=NULL )
return( em_gb2312 );
if ( uc_strstrmatch(setname,"big5")!=NULL )
return( em_big5 );
if ( uc_strstrmatch(setname,"mac")!=NULL )
return( em_mac );
if ( uc_strstrmatch(setname,"win")!=NULL )
return( em_win );
if ( IsUserMap(setname))
return( em_user );
/* !!! Encodings used for postscript japanese fonts, which I don't understand */
#if 0
if ( uc_strstrmatch(setname,"RJSJ")!=NULL )
return( em_sjis );
if ( uc_strstrmatch(setname,"EUC")!=NULL )
return( em_euc );
#endif
return( em_none );
}
U_CAPI UChar* U_EXPORT2
u_strFindFirst(const UChar* s, int32_t length,
const UChar* sub, int32_t subLength) {
const UChar* start, * p, * q, * subLimit;
UChar c, cs, cq;
if (sub == NULL || subLength < -1) {
return (UChar*) s;
}
if (s == NULL || length < -1) {
return NULL;
}
start = s;
if (length < 0 && subLength < 0) {
/* both strings are NUL-terminated */
if ((cs = *sub++) == 0) {
return (UChar*) s;
}
if (*sub == 0 && !U16_IS_SURROGATE(cs)) {
/* the substring consists of a single, non-surrogate BMP code point */
return u_strchr(s, cs);
}
while ((c = *s++) != 0) {
if (c == cs) {
/* found first substring UChar, compare rest */
p = s;
q = sub;
for (; ;) {
if ((cq = *q) == 0) {
if (isMatchAtCPBoundary(start, s - 1, p, NULL)) {
return (UChar*) (s - 1); /* well-formed match */
} else {
break; /* no match because surrogate pair is split */
}
}
if ((c = *p) == 0) {
return NULL; /* no match, and none possible after s */
}
if (c != cq) {
break; /* no match */
}
++p;
++q;
}
}
}
/* not found */
return NULL;
}
if (subLength < 0) {
subLength = u_strlen(sub);
}
if (subLength == 0) {
return (UChar*) s;
}
/* get sub[0] to search for it fast */
cs = *sub++;
--subLength;
subLimit = sub + subLength;
if (subLength == 0 && !U16_IS_SURROGATE(cs)) {
/* the substring consists of a single, non-surrogate BMP code point */
return length < 0 ? u_strchr(s, cs) : u_memchr(s, cs, length);
}
if (length < 0) {
/* s is NUL-terminated */
while ((c = *s++) != 0) {
if (c == cs) {
/* found first substring UChar, compare rest */
p = s;
q = sub;
for (; ;) {
if (q == subLimit) {
if (isMatchAtCPBoundary(start, s - 1, p, NULL)) {
return (UChar*) (s - 1); /* well-formed match */
} else {
break; /* no match because surrogate pair is split */
}
}
if ((c = *p) == 0) {
return NULL; /* no match, and none possible after s */
}
if (c != *q) {
break; /* no match */
}
++p;
++q;
}
}
}
} else {
const UChar* limit, * preLimit;
/* subLength was decremented above */
if (length <= subLength) {
return NULL; /* s is shorter than sub */
}
limit = s + length;
/* the substring must start before preLimit */
preLimit = limit - subLength;
while (s != preLimit) {
c = *s++;
if (c == cs) {
/* found first substring UChar, compare rest */
p = s;
q = sub;
for (; ;) {
if (q == subLimit) {
if (isMatchAtCPBoundary(start, s - 1, p, limit)) {
return (UChar*) (s - 1); /* well-formed match */
} else {
break; /* no match because surrogate pair is split */
}
}
if (*p != *q) {
break; /* no match */
}
++p;
++q;
}
}
}
}
/* not found */
return NULL;
}
/**
* Testing the discontigous contractions
*/
static void TestDiscontiguos() {
const char *rulestr =
"&z < AB < X\\u0300 < ABC < X\\u0300\\u0315";
UChar rule[50];
int rulelen = u_unescape(rulestr, rule, 50);
const char *src[] = {
"ADB", "ADBC", "A\\u0315B", "A\\u0315BC",
/* base character blocked */
"XD\\u0300", "XD\\u0300\\u0315",
/* non blocking combining character */
"X\\u0319\\u0300", "X\\u0319\\u0300\\u0315",
/* blocking combining character */
"X\\u0314\\u0300", "X\\u0314\\u0300\\u0315",
/* contraction prefix */
"ABDC", "AB\\u0315C","X\\u0300D\\u0315", "X\\u0300\\u0319\\u0315",
"X\\u0300\\u031A\\u0315",
/* ends not with a contraction character */
"X\\u0319\\u0300D", "X\\u0319\\u0300\\u0315D", "X\\u0300D\\u0315D",
"X\\u0300\\u0319\\u0315D", "X\\u0300\\u031A\\u0315D"
};
const char *tgt[] = {
/* non blocking combining character */
"A D B", "A D BC", "A \\u0315 B", "A \\u0315 BC",
/* base character blocked */
"X D \\u0300", "X D \\u0300\\u0315",
/* non blocking combining character */
"X\\u0300 \\u0319", "X\\u0300\\u0315 \\u0319",
/* blocking combining character */
"X \\u0314 \\u0300", "X \\u0314 \\u0300\\u0315",
/* contraction prefix */
"AB DC", "AB \\u0315 C","X\\u0300 D \\u0315", "X\\u0300\\u0315 \\u0319",
"X\\u0300 \\u031A \\u0315",
/* ends not with a contraction character */
"X\\u0300 \\u0319D", "X\\u0300\\u0315 \\u0319D", "X\\u0300 D\\u0315D",
"X\\u0300\\u0315 \\u0319D", "X\\u0300 \\u031A\\u0315D"
};
int size = 20;
UCollator *coll;
UErrorCode status = U_ZERO_ERROR;
int count = 0;
UCollationElements *iter;
UCollationElements *resultiter;
coll = ucol_openRules(rule, rulelen, UCOL_OFF, UCOL_DEFAULT_STRENGTH,NULL, &status);
iter = ucol_openElements(coll, rule, 1, &status);
resultiter = ucol_openElements(coll, rule, 1, &status);
if (U_FAILURE(status)) {
log_err_status(status, "Error opening collation rules -> %s\n", u_errorName(status));
return;
}
while (count < size) {
UChar str[20];
UChar tstr[20];
int strLen = u_unescape(src[count], str, 20);
UChar *s;
ucol_setText(iter, str, strLen, &status);
if (U_FAILURE(status)) {
log_err("Error opening collation iterator\n");
return;
}
u_unescape(tgt[count], tstr, 20);
s = tstr;
log_verbose("count %d\n", count);
for (;;) {
uint32_t ce;
UChar *e = u_strchr(s, 0x20);
if (e == 0) {
e = u_strchr(s, 0);
}
ucol_setText(resultiter, s, (int32_t)(e - s), &status);
ce = ucol_next(resultiter, &status);
if (U_FAILURE(status)) {
log_err("Error manipulating collation iterator\n");
return;
}
while (ce != UCOL_NULLORDER) {
if (ce != (uint32_t)ucol_next(iter, &status) ||
U_FAILURE(status)) {
log_err("Discontiguos contraction test mismatch\n");
return;
}
ce = ucol_next(resultiter, &status);
if (U_FAILURE(status)) {
log_err("Error getting next collation element\n");
return;
}
}
s = e + 1;
if (*e == 0) {
break;
}
}
ucol_reset(iter);
backAndForth(iter);
count ++;
}
ucol_closeElements(resultiter);
ucol_closeElements(iter);
ucol_close(coll);
}
/**
* Computes training by extracting statistics from a tagged corpus file.
*/
void do_training(U_FILE* input_text,U_FILE* rforms_file,U_FILE* iforms_file){
/* these two hash tables are respectively for simple and compound entries */
struct string_hash_ptr* rforms_table = NULL, *iforms_table = NULL;
if(rforms_file != NULL){
rforms_table = new_string_hash_ptr(200000);
}
if(iforms_file != NULL){
iforms_table = new_string_hash_ptr(200000);
}
/* we initialize a contextual matrix */
struct corpus_entry** context = new_context_matrix();
initialize_context_matrix(context);
unichar line[MAX_TAGGED_CORPUS_LINE];
/* check the format of the corpus */
long previous_file_position = ftell(input_text);
if(u_fgets(line,input_text) == EOF){
fatal_error("File is empty");
}
fseek(input_text,previous_file_position,SEEK_SET);
int format_corpus = check_corpus_entry(line);
if(format_corpus == 0){
// the corpus is in the Tagger format, one word per line where line=word/tag
while(u_fgets(line,input_text) !=EOF){
if(u_strlen(line) == 0){
initialize_context_matrix(context);
}
else{
corpus_entry* entry = new_corpus_entry(line);
if(u_strchr(line,'_')!=NULL && line[0]!='_'){
corpus_entry** entries = extract_simple_words(entry);
free_corpus_entry(entry);
for(int i=0;entries[i]!=NULL;i++){
push_corpus_entry(entries[i],context);
add_statistics(context,rforms_table,iforms_table);
}
free(entries);
}
else {
push_corpus_entry(entry,context);
add_statistics(context,rforms_table,iforms_table);
}
}
}
}
else {
// the corpus is in the Unitex tagged format, one sentence per line where token={word,lemma.tag}
unichar *tmp,*s = (unichar*)malloc(sizeof(unichar)*(MAX_TAGGED_CORPUS_LINE));
int current_len,len;
unsigned int i;
while(u_fgets(line,input_text) != EOF){
current_len = 0, len = 0;
/* extract each token of the sentence */
for (;;) {
len = 1+u_strlen(line+current_len)-u_strlen(u_strchr(line+current_len,'}'));
tmp = u_strcpy_sized(s,len-1,line+current_len+1);
u_strcat(tmp,"\0");
if(u_strcmp(s,"S") == 0)
break;
//particular case: '\},\}.PONCT'
if(line[current_len+2] == '}'){
int start = current_len+3;
do{
tmp = u_strchr(line+start,'}');
start += 1+u_strlen(line+start)-u_strlen(tmp);
}
while(*(tmp+1) != ' ');
tmp = u_strcpy_sized(s,start-current_len-1,line+current_len+1);
u_strcat(tmp,"\0");
len += start-current_len-3;
}
/* format the {XX.YY} into standard tagger format, XX/YY */
unichar* newline = (unichar*)malloc(sizeof(unichar)*(8096));
if(u_strchr(s,',')[1] == ','){
u_strcpy(newline,",");
}
else
u_strcpy_sized(newline,1+u_strlen(s)-u_strlen(u_strchr(s,',')),s);
u_sprintf(newline,"%S/%S\0",newline,s+u_strrchr(s,'.')+1);
for(i=0;i<u_strlen(newline);i++){
if(newline[i] == ' ')
newline[i] = '_';
}
//create corpus entry
corpus_entry* entry = new_corpus_entry(newline);
if(u_strchr(newline,'_') != NULL && newline[0] != '_'){
corpus_entry** entries = extract_simple_words(entry);
free_corpus_entry(entry);
for(int j=0;entries[j]!=NULL;j++){
push_corpus_entry(entries[j],context);
add_statistics(context,rforms_table,iforms_table);
}
free(entries);
}
else {
push_corpus_entry(entry,context);
add_statistics(context,rforms_table,iforms_table);
}
free(newline);
current_len += len+1;
}
initialize_context_matrix(context);
}
free(s);
}
free_context_matrix(context);
/* we fill dictionary files with pairs (tuple,value) and then
* we add a special line "CODE\tFEATURES,.value" in order to
* specify whether the dictionary contains inflected or raw form tuples*/
unichar* str = u_strdup("");
if(rforms_table != NULL){
write_keys_values(rforms_table,rforms_table->hash->root,str,rforms_file);
u_fprintf(rforms_file,"%s,.%d\n","CODE\tFEATURES",0);
free_string_hash_ptr(rforms_table,NULL);
}
if(iforms_table != NULL){
write_keys_values(iforms_table,iforms_table->hash->root,str,iforms_file);
u_fprintf(iforms_file,"%s,.%d\n","CODE\tFEATURES",1);
free_string_hash_ptr(iforms_table,NULL);
}
free(str);
}
static jobjectArray getContentImpl(JNIEnv* env, jclass clazz,
jstring locale, jboolean needsTZ) {
UErrorCode status = U_ZERO_ERROR;
const char *loc = env->GetStringUTFChars(locale, NULL);
UResourceBundle *root = ures_openU(NULL, loc, &status);
env->ReleaseStringUTFChars(locale, loc);
if(U_FAILURE(status)) {
LOGI("Error getting resources");
status = U_ZERO_ERROR;
return NULL;
}
jclass obj_class = env->FindClass("java/lang/Object");
jclass integer_class = env->FindClass("java/lang/Integer");
jmethodID integerInit = env->GetMethodID(integer_class, "<init>", "(I)V");
jobjectArray result;
jobject firstDayOfWeek = NULL;
jobject minimalDaysInFirstWeek = NULL;
jobjectArray amPmMarkers = NULL;
jobjectArray eras = NULL;
jstring localPatternChars = NULL;
jobjectArray weekdays = NULL;
jobjectArray shortWeekdays = NULL;
jobjectArray months = NULL;
jobjectArray shortMonths = NULL;
jstring time_SHORT = NULL;
jstring time_MEDIUM = NULL;
jstring time_LONG = NULL;
jstring time_FULL = NULL;
jstring date_SHORT = NULL;
jstring date_MEDIUM = NULL;
jstring date_LONG = NULL;
jstring date_FULL = NULL;
jstring decimalPatternChars = NULL;
jstring naN = NULL;
jstring infinity = NULL;
jstring currencySymbol = NULL;
jstring intCurrencySymbol = NULL;
jstring numberPattern = NULL;
jstring integerPattern = NULL;
jstring currencyPattern = NULL;
jstring percentPattern = NULL;
jobjectArray zones = NULL;
int counter = 0;
int firstDayVals[2] = {-1, -1};
const jchar* nan = (const jchar *)NULL;
const jchar* inf = (const jchar *)NULL;
int nanL, infL;
UResourceBundle *gregorian;
UResourceBundle *gregorianElems;
UResourceBundle *rootElems;
// get the resources needed
rootElems = ures_getByKey(root, "calendar", NULL, &status);
if(U_FAILURE(status)) {
return NULL;
}
gregorian = ures_getByKey(rootElems, "gregorian", NULL, &status);
if(U_FAILURE(status)) {
ures_close(rootElems);
return NULL;
}
// adding the first day of week and minimal days in first week values
getDayInitVector(env, gregorian, firstDayVals);
if((firstDayVals[0] != -1) && (firstDayVals[1] != -1)) {
firstDayOfWeek = env->NewObject(integer_class, integerInit, firstDayVals[0]);
minimalDaysInFirstWeek = env->NewObject(integer_class, integerInit, firstDayVals[1]);
// adding First_Day and Minimal_Days integer to the result
counter += 2;
}
// adding ampm string array to the result");
amPmMarkers = getAmPmMarkers(env, gregorian);
if(amPmMarkers != NULL) {
counter++;
}
// adding eras string array to the result
eras = getEras(env, gregorian);
if(eras != NULL) {
counter++;
}
// local pattern chars are initially always the same
localPatternChars = env->NewStringUTF("GyMdkHmsSEDFwWahKzZ");
// adding local pattern chars string to the result
counter++;
// adding month names string array to the result
months = getMonthNames(env, gregorian);
if(months != NULL) {
counter++;
}
// adding short month names string array to the result
shortMonths = getShortMonthNames(env, gregorian);
if(shortMonths != NULL) {
counter++;
}
// adding day names string array to the result
weekdays = getWeekdayNames(env, gregorian);
if(weekdays != NULL) {
counter++;
}
// adding short day names string array to the result
shortWeekdays = getShortWeekdayNames(env, gregorian);
if(shortWeekdays != NULL) {
counter++;
}
const UChar *pattern;
jchar check[2] = {0, 0};
u_uastrcpy(check, "v");
jchar replacement[2] = {0, 0};
u_uastrcpy(replacement, "z");
jchar *pos;
jchar *patternCopy;
int patternLength;
// adding date and time format patterns to the result
gregorianElems = ures_getByKey(gregorian, "DateTimePatterns", NULL, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
pattern = ures_getStringByIndex(gregorianElems, 0, &patternLength, &status);
// there are some patterns in icu that use the pattern character 'v'
// java doesn't accept this, so it gets replaced by 'z' which has
// about the same result as 'v', the timezone name.
// 'v' -> "PT", 'z' -> "PST", v is the generic timezone and z the standard tz
// "vvvv" -> "Pacific Time", "zzzz" -> "Pacific Standard Time"
patternCopy = (jchar *) malloc((patternLength + 1) * sizeof(jchar));
u_strcpy(patternCopy, pattern);
if(U_FAILURE(status)) {
free(patternCopy);
status = U_ZERO_ERROR;
goto endOfCalendar;
}
while((pos = u_strchr(patternCopy, check[0])) != NULL) {
u_memset(pos, replacement[0], 1);
}
time_FULL = env->NewString(patternCopy, patternLength);
free(patternCopy);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 1, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
time_LONG = env->NewString(pattern, patternLength);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 2, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
time_MEDIUM = env->NewString(pattern, patternLength);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 3, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
time_SHORT = env->NewString(pattern, patternLength);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 4, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
date_FULL = env->NewString(pattern, patternLength);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 5, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
date_LONG = env->NewString(pattern, patternLength);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 6, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
date_MEDIUM = env->NewString(pattern, patternLength);
counter++;
pattern = ures_getStringByIndex(gregorianElems, 7, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto endOfCalendar;
}
date_SHORT = env->NewString(pattern, patternLength);
counter++;
endOfCalendar:
if(gregorianElems != NULL) {
ures_close(gregorianElems);
}
ures_close(gregorian);
ures_close(rootElems);
rootElems = ures_getByKey(root, "NumberElements", NULL, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
}
if(ures_getSize(rootElems) >= 11) {
// adding decimal pattern chars to the result
decimalPatternChars = getDecimalPatternChars(env, rootElems);
if(decimalPatternChars != NULL) {
counter++;
}
// adding NaN pattern char to the result
nan = ures_getStringByIndex(rootElems, 10, &nanL, &status);
if(U_SUCCESS(status)) {
naN = env->NewString(nan, nanL);
counter++;
}
status = U_ZERO_ERROR;
// adding infinity pattern char to the result
inf = ures_getStringByIndex(rootElems, 9, &infL, &status);
if(U_SUCCESS(status)) {
infinity = env->NewString(inf, infL);
counter++;
}
status = U_ZERO_ERROR;
}
ures_close(rootElems);
// adding intl currency code to result
intCurrencySymbol = getIntCurrencyCode(env, clazz, locale);
if(intCurrencySymbol != NULL) {
// adding currency symbol to result
currencySymbol = getCurrencySymbol(env, clazz, locale, intCurrencySymbol);
} else {
intCurrencySymbol = env->NewStringUTF("XXX");
}
if(currencySymbol == NULL) {
currencySymbol = env->NewStringUTF("\u00a4");
}
counter += 2;
// adding number format patterns to the result
int numOfEntries;
int decSepOffset;
NumberFormat *nf;
jchar *tmpPattern;
rootElems = ures_getByKey(root, "NumberPatterns", NULL, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
goto zones;
}
numOfEntries = ures_getSize(rootElems);
if(numOfEntries < 3) {
ures_close(rootElems);
goto zones;
}
// number pattern
pattern = ures_getStringByIndex(rootElems, 0, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
ures_close(rootElems);
goto zones;
}
numberPattern = env->NewString(pattern, patternLength);
counter++;
// integer pattern derived from number pattern
decSepOffset = u_strcspn(pattern, (jchar *)".\0");
tmpPattern = (jchar *) malloc((decSepOffset + 1) * sizeof(jchar));
u_strncpy(tmpPattern, pattern, decSepOffset);
integerPattern = env->NewString(tmpPattern, decSepOffset);
free(tmpPattern);
counter++;
// currency pattern
pattern = ures_getStringByIndex(rootElems, 1, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
ures_close(rootElems);
goto zones;
}
currencyPattern = env->NewString(pattern, patternLength);
counter++;
// percent pattern
pattern = ures_getStringByIndex(rootElems, 2, &patternLength, &status);
if(U_FAILURE(status)) {
status = U_ZERO_ERROR;
ures_close(rootElems);
goto zones;
}
percentPattern = env->NewString(pattern, patternLength);
counter++;
ures_close(rootElems);
zones:
ures_close(root);
if(needsTZ == JNI_TRUE) {
counter++; //add empty timezone
}
// collect all content and put it into an array
result = env->NewObjectArray(counter, obj_class, NULL);
int index = 0;
if(needsTZ == JNI_TRUE) {
addObject(env, result, "timezones", NULL, index++);
}
if(firstDayOfWeek != NULL && index < counter) {
addObject(env, result, "First_Day", firstDayOfWeek, index++);
}
if(minimalDaysInFirstWeek != NULL && index < counter) {
addObject(env, result, "Minimal_Days", minimalDaysInFirstWeek, index++);
}
if(amPmMarkers != NULL && index < counter) {
addObject(env, result, "ampm", amPmMarkers, index++);
}
if(eras != NULL && index < counter) {
addObject(env, result, "eras", eras, index++);
}
if(localPatternChars != NULL && index < counter) {
addObject(env, result, "LocalPatternChars", localPatternChars, index++);
}
if(weekdays != NULL && index < counter) {
addObject(env, result, "weekdays", weekdays, index++);
}
if(shortWeekdays != NULL && index < counter) {
addObject(env, result, "shortWeekdays", shortWeekdays, index++);
}
if(months != NULL && index < counter) {
addObject(env, result, "months", months, index++);
}
if(shortMonths != NULL && index < counter) {
addObject(env, result, "shortMonths", shortMonths, index++);
}
if(time_SHORT != NULL && index < counter) {
addObject(env, result, "Time_SHORT", time_SHORT, index++);
}
if(time_MEDIUM != NULL && index < counter) {
addObject(env, result, "Time_MEDIUM", time_MEDIUM, index++);
}
if(time_LONG != NULL && index < counter) {
addObject(env, result, "Time_LONG", time_LONG, index++);
}
if(time_FULL != NULL && index < counter) {
addObject(env, result, "Time_FULL", time_FULL, index++);
}
if(date_SHORT != NULL && index < counter) {
addObject(env, result, "Date_SHORT", date_SHORT, index++);
}
if(date_MEDIUM != NULL && index < counter) {
addObject(env, result, "Date_MEDIUM", date_MEDIUM, index++);
}
if(date_LONG != NULL && index < counter) {
addObject(env, result, "Date_LONG", date_LONG, index++);
}
if(date_FULL != NULL && index < counter) {
addObject(env, result, "Date_FULL", date_FULL, index++);
}
if(decimalPatternChars != NULL && index < counter) {
addObject(env, result, "DecimalPatternChars", decimalPatternChars, index++);
}
if(naN != NULL && index < counter) {
addObject(env, result, "NaN", naN, index++);
}
if(infinity != NULL && index < counter) {
addObject(env, result, "Infinity", infinity, index++);
}
if(currencySymbol != NULL && index < counter) {
addObject(env, result, "CurrencySymbol", currencySymbol, index++);
}
if(intCurrencySymbol != NULL && index < counter) {
addObject(env, result, "IntCurrencySymbol", intCurrencySymbol, index++);
}
if(numberPattern != NULL && index < counter) {
addObject(env, result, "Number", numberPattern, index++);
}
if(integerPattern != NULL && index < counter) {
addObject(env, result, "Integer", integerPattern, index++);
}
if(currencyPattern != NULL && index < counter) {
addObject(env, result, "Currency", currencyPattern, index++);
}
if(percentPattern != NULL && index < counter) {
addObject(env, result, "Percent", percentPattern, index++);
}
return result;
}
U_NAMESPACE_BEGIN
// -------------------------------------
BreakIterator*
BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind, UErrorCode &status)
{
char fnbuff[256];
char ext[4]={'\0'};
char actualLocale[ULOC_FULLNAME_CAPACITY];
int32_t size;
const UChar* brkfname = NULL;
UResourceBundle brkRulesStack;
UResourceBundle brkNameStack;
UResourceBundle *brkRules = &brkRulesStack;
UResourceBundle *brkName = &brkNameStack;
RuleBasedBreakIterator *result = NULL;
if (U_FAILURE(status))
return NULL;
ures_initStackObject(brkRules);
ures_initStackObject(brkName);
// Get the locale
UResourceBundle *b = ures_open(U_ICUDATA_BRKITR, loc.getName(), &status);
/* this is a hack for now. Should be fixed when the data is fetched from
brk_index.txt */
if(status==U_USING_DEFAULT_WARNING){
status=U_ZERO_ERROR;
ures_openFillIn(b, U_ICUDATA_BRKITR, "", &status);
}
// Get the "boundaries" array.
if (U_SUCCESS(status)) {
brkRules = ures_getByKeyWithFallback(b, "boundaries", brkRules, &status);
// Get the string object naming the rules file
brkName = ures_getByKeyWithFallback(brkRules, type, brkName, &status);
// Get the actual string
brkfname = ures_getString(brkName, &size, &status);
U_ASSERT((size_t)size<sizeof(fnbuff));
if ((size_t)size>=sizeof(fnbuff)) {
size=0;
if (U_SUCCESS(status)) {
status = U_BUFFER_OVERFLOW_ERROR;
}
}
// Use the string if we found it
if (U_SUCCESS(status) && brkfname) {
uprv_strncpy(actualLocale,
ures_getLocale(brkName, &status),
sizeof(actualLocale)/sizeof(actualLocale[0]));
UChar* extStart=u_strchr(brkfname, 0x002e);
int len = 0;
if(extStart!=NULL){
len = (int)(extStart-brkfname);
u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
u_UCharsToChars(brkfname, fnbuff, len);
}
fnbuff[len]=0; // nul terminate
}
}
ures_close(brkRules);
ures_close(brkName);
UDataMemory* file = udata_open(U_ICUDATA_BRKITR, ext, fnbuff, &status);
if (U_FAILURE(status)) {
ures_close(b);
return NULL;
}
// Create a RuleBasedBreakIterator
result = new RuleBasedBreakIterator(file, status);
// If there is a result, set the valid locale and actual locale, and the kind
if (U_SUCCESS(status) && result != NULL) {
U_LOCALE_BASED(locBased, *(BreakIterator*)result);
locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), actualLocale);
result->setBreakType(kind);
}
ures_close(b);
if (U_FAILURE(status) && result != NULL) { // Sometimes redundant check, but simple
delete result;
return NULL;
}
if (result == NULL) {
udata_close(file);
if (U_SUCCESS(status)) {
status = U_MEMORY_ALLOCATION_ERROR;
}
}
return result;
}
static void GIOdispatch(GIOControl *gc, enum giofuncs gf) {
unichar_t *temp, *pt, *tpt;
int i;
gc->gf = gf;
if ( _GIO_stdfuncs.useragent == NULL )
_GIO_stdfuncs.useragent = copy("*****@*****.**");
temp = _GIO_translateURL(gc->path,gf);
if ( temp!=NULL ) {
if ( gc->origpath==NULL )
gc->origpath = gc->path;
else
free(gc->path);
gc->path = temp;
}
if ( gc->topath!=NULL ) {
temp = _GIO_translateURL(gc->topath,gf);
if ( temp!=NULL ) {
free(gc->topath);
gc->topath = temp;
}
if ( gf==gf_renamefile ) {
if (( pt = uc_strstr(gc->path,"://"))== NULL )
pt = gc->path;
else {
pt=u_strchr(pt+3,'/');
if ( pt==NULL ) pt = gc->path+u_strlen(gc->path);
}
if (( tpt = uc_strstr(gc->topath,"://"))== NULL )
tpt = gc->topath;
else {
tpt=u_strchr(tpt+3,'/');
if ( tpt==NULL ) tpt = gc->topath+u_strlen(gc->topath);
}
if ( tpt-gc->topath!=pt-gc->path ||
u_strnmatch(gc->path,gc->topath,pt-gc->path)!=0 ) {
_GIO_reporterror(gc,EXDEV);
return;
}
}
}
pt = uc_strstr(gc->path,"://");
if ( pt!=NULL ) {
for ( i=0; i<plen; ++i )
if ( u_strnmatch(protocols[i].proto,gc->path,pt-gc->path)==0 )
break;
if ( i>=plen && !AddProtocol(gc->path,pt-gc->path) ) {
gc->protocol_index = -2;
gc->return_code = 501;
gc->error = err501;
uc_strcpy(gc->status,"No support for browsing: ");
u_strncpy(gc->status+u_strlen(gc->status), gc->path, pt-gc->path );
gc->done = true;
(gc->receiveerror)(gc);
return;
}
gc->protocol_index = i;
if ( !protocols[i].dothread )
(protocols[i].dispatcher)(gc);
else {
#ifndef HAVE_PTHREAD_H
gc->return_code = 501;
gc->error = err501;
uc_strcpy(gc->status,"No support for protocol");
gc->done = true;
(gc->receiveerror)(gc);
return;
#else
static pthread_cond_t initcond = PTHREAD_COND_INITIALIZER;
static pthread_mutex_t initmutex = PTHREAD_MUTEX_INITIALIZER;
/* could put stuff here to queue functions if we get too many */
/* threads, or perhaps even a thread pool */
uc_strcpy(gc->status,"Queued");
gc->threaddata = (struct gio_threaddata *) malloc(sizeof(struct gio_threaddata));
gc->threaddata->mutex = initmutex;
gc->threaddata->cond = initcond;
if ( _GIO_stdfuncs.gdraw_sync_thread!=NULL )
(_GIO_stdfuncs.gdraw_sync_thread)(NULL,NULL,NULL);
pthread_create(&gc->threaddata->thread,NULL,
(ptread_startfunc_t *) (protocols[i].dispatcher), gc);
#endif
}
} else {
gc->protocol_index = -1;
_GIO_localDispatch(gc);
}
}
/*
* This function behaves in the same way that a main one, except that it does
* not invoke the setBufferMode function.
*/
int main_LocateTfst(int argc,char* const argv[]) {
if (argc==1) {
usage();
return SUCCESS_RETURN_CODE;
}
VersatileEncodingConfig vec=VEC_DEFAULT;
int val,index=-1;
char text[FILENAME_MAX]="";
char alphabet[FILENAME_MAX]="";
int is_korean=0;
int tilde_negation_operator=1;
int selected_negation_operator=0;
int tagging=0;
int single_tags_only=0;
int match_word_boundaries=1;
MatchPolicy match_policy=LONGEST_MATCHES;
OutputPolicy output_policy=IGNORE_OUTPUTS;
AmbiguousOutputPolicy ambiguous_output_policy=ALLOW_AMBIGUOUS_OUTPUTS;
VariableErrorPolicy variable_error_policy=IGNORE_VARIABLE_ERRORS;
int search_limit=NO_MATCH_LIMIT;
char foo;
vector_ptr* injected=new_vector_ptr();
bool only_verify_arguments = false;
UnitexGetOpt options;
while (EOF!=(val=options.parse_long(argc,argv,optstring_LocateTfst,lopts_LocateTfst,&index))) {
switch(val) {
case 't': if (options.vars()->optarg[0]=='\0') {
error("You must specify a non empty .tfst name\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
strcpy(text,options.vars()->optarg);
break;
case 'a': if (options.vars()->optarg[0]=='\0') {
error("You must specify a non empty alphabet name\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
strcpy(alphabet,options.vars()->optarg);
break;
case 'K': is_korean=1;
match_word_boundaries=0;
break;
case 'l': search_limit=NO_MATCH_LIMIT; break;
case 'g': if (options.vars()->optarg[0]=='\0') {
error("You must specify an argument for negation operator\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
selected_negation_operator=1;
if ((strcmp(options.vars()->optarg,"minus")==0) || (strcmp(options.vars()->optarg,"-")==0)) {
tilde_negation_operator=0;
}
else
if ((strcmp(options.vars()->optarg,"tilde")!=0) && (strcmp(options.vars()->optarg,"~")!=0)) {
error("You must specify a valid argument for negation operator\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
break;
case 'n': if (1!=sscanf(options.vars()->optarg,"%d%c",&search_limit,&foo) || search_limit<=0) {
/* foo is used to check that the search limit is not like "45gjh" */
error("Invalid search limit argument: %s\n",options.vars()->optarg);
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
break;
case 'S': match_policy=SHORTEST_MATCHES; break;
case 'L': match_policy=LONGEST_MATCHES; break;
case 'A': match_policy=ALL_MATCHES; break;
case 'I': output_policy=IGNORE_OUTPUTS; break;
case 'M': output_policy=MERGE_OUTPUTS; break;
case 'R': output_policy=REPLACE_OUTPUTS; break;
case 'X': variable_error_policy=EXIT_ON_VARIABLE_ERRORS; break;
case 'Y': variable_error_policy=IGNORE_VARIABLE_ERRORS; break;
case 'Z': variable_error_policy=BACKTRACK_ON_VARIABLE_ERRORS; break;
case 'b': ambiguous_output_policy=ALLOW_AMBIGUOUS_OUTPUTS; break;
case 'z': ambiguous_output_policy=IGNORE_AMBIGUOUS_OUTPUTS; break;
case 'V': only_verify_arguments = true;
break;
case 'h': usage();
return SUCCESS_RETURN_CODE;
case 1: tagging=1; break;
case 2: single_tags_only=1; break;
case 3: match_word_boundaries=0; break;
case 'k': if (options.vars()->optarg[0]=='\0') {
error("Empty input_encoding argument\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
decode_reading_encoding_parameter(&(vec.mask_encoding_compatibility_input),options.vars()->optarg);
break;
case 'q': if (options.vars()->optarg[0]=='\0') {
error("Empty output_encoding argument\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
decode_writing_encoding_parameter(&(vec.encoding_output),&(vec.bom_output),options.vars()->optarg);
break;
case 'v': {
unichar* key=u_strdup(options.vars()->optarg);
unichar* value=u_strchr(key,'=');
if (value==NULL) {
error("Invalid variable injection: %s\n",options.vars()->optarg);
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
(*value)='\0';
value++;
value=u_strdup(value);
vector_ptr_add(injected,key);
vector_ptr_add(injected,value);
break;
}
case ':': index==-1 ? error("Missing argument for option -%c\n",options.vars()->optopt) :
error("Missing argument for option --%s\n",lopts_LocateTfst[index].name);
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
case '?': index==-1 ? error("Invalid option -%c\n",options.vars()->optopt) :
error("Invalid option --%s\n",options.vars()->optarg);
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
break;
}
index=-1;
}
if (options.vars()->optind!=argc-1) {
error("Invalid arguments: rerun with --help\n");
free_vector_ptr(injected);
return USAGE_ERROR_CODE;
}
if (only_verify_arguments) {
// freeing all allocated memory
free_vector_ptr(injected);
return SUCCESS_RETURN_CODE;
}
if (selected_negation_operator==0) {
get_graph_compatibility_mode_by_file(&vec,&tilde_negation_operator);
}
char grammar[FILENAME_MAX];
char output[FILENAME_MAX];
strcpy(grammar,argv[options.vars()->optind]);
get_path(text,output);
strcat(output,"concord.ind");
int OK=locate_tfst(text,
grammar,
alphabet,
output,
&vec,
match_policy,
output_policy,
ambiguous_output_policy,
variable_error_policy,
search_limit,
is_korean,
tilde_negation_operator,
injected,
tagging,
single_tags_only,
match_word_boundaries);
free_vector_ptr(injected);
return (!OK);
}
static void SplashLayout() {
unichar_t *start, *pt, *lastspace;
extern const char *source_modtime_str;
extern const char *source_version_str;
uc_strcpy(msg, "When my father finished his book on Renaissance printing (The Craft of Printing and the Publication of Shakespeare's Works) he told me that I would have to write the chapter on computer typography. This is my attempt to do so.");
GDrawSetFont(splashw,splash_font);
linecnt = 0;
lines[linecnt++] = msg-1;
for ( start = msg; *start!='\0'; start = pt ) {
lastspace = NULL;
for ( pt=start; ; ++pt ) {
if ( *pt==' ' || *pt=='\0' ) {
if ( GDrawGetTextWidth(splashw,start,pt-start,NULL)<splashimage.u.image->width-10 )
lastspace = pt;
else
break;
if ( *pt=='\0' )
break;
}
}
if ( lastspace!=NULL )
pt = lastspace;
lines[linecnt++] = pt;
if ( *pt ) ++pt;
}
uc_strcpy(pt, " FontForge used to be named PfaEdit.");
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcpy(pt," Version: ");;
uc_strcat(pt,source_modtime_str);
uc_strcat(pt," (");
uc_strcat(pt,source_version_str);
#ifdef FONTFORGE_CONFIG_TYPE3
uc_strcat(pt,"-ML");
#endif
#ifdef FREETYPE_HAS_DEBUGGER
uc_strcat(pt,"-TtfDb");
#endif
#ifdef _NO_PYTHON
uc_strcat(pt,"-NoPython");
#endif
#ifdef FONTFORGE_CONFIG_USE_LONGDOUBLE
uc_strcat(pt,"-LD");
#elif defined(FONTFORGE_CONFIG_USE_DOUBLE)
uc_strcat(pt,"-D");
#endif
#ifndef FONTFORGE_CONFIG_DEVICETABLES
uc_strcat(pt,"-NoDevTab");
#endif
uc_strcat(pt,")");
pt += u_strlen(pt);
lines[linecnt++] = pt;
uc_strcpy(pt," Library Version: ");
uc_strcat(pt,library_version_configuration.library_source_modtime_string);
lines[linecnt++] = pt+u_strlen(pt);
lines[linecnt] = NULL;
is = u_strchr(msg,'(');
ie = u_strchr(msg,')');
}