/ext/hunspell/affixmgr.cxx
C++ | 3976 lines | 3145 code | 418 blank | 413 comment | 1291 complexity | 1cd26bf8aa3cdc888a4490b61d8d99f2 MD5 | raw file
Possible License(s): GPL-2.0, LGPL-2.0, GPL-3.0, LGPL-3.0, MPL-2.0-no-copyleft-exception
Large files files are truncated, but you can click here to view the full file
- #include "license.hunspell"
- #include "license.myspell"
-
- #ifndef MOZILLA_CLIENT
- #include <cstdlib>
- #include <cstring>
- #include <cctype>
- #include <cstdio>
- #else
- #include <stdlib.h>
- #include <string.h>
- #include <stdio.h>
- #include <ctype.h>
- #endif
-
- #include "affixmgr.hxx"
- #include "affentry.hxx"
- #include "langnum.hxx"
-
- #include "csutil.hxx"
-
- #ifndef MOZILLA_CLIENT
- #ifndef W32
- using namespace std;
- #endif
- #endif
-
- AffixMgr::AffixMgr(const char * affpath, HashMgr* ptr)
- {
- // register hash manager and load affix data from aff file
- pHMgr = ptr;
- trystring = NULL;
- encoding=NULL;
- utf8 = 0;
- complexprefixes = 0;
- maptable = NULL;
- nummap = 0;
- breaktable = NULL;
- numbreak = 0;
- reptable = NULL;
- numrep = 0;
- checkcpdtable = NULL;
- numcheckcpd = 0;
- defcpdtable = NULL;
- numdefcpd = 0;
- compoundflag = FLAG_NULL; // permits word in compound forms
- compoundbegin = FLAG_NULL; // may be first word in compound forms
- compoundmiddle = FLAG_NULL; // may be middle word in compound forms
- compoundend = FLAG_NULL; // may be last word in compound forms
- compoundroot = FLAG_NULL; // compound word signing flag
- compoundpermitflag = FLAG_NULL; // compound permitting flag for suffixed word
- compoundforbidflag = FLAG_NULL; // compound fordidden flag for suffixed word
- checkcompounddup = 0; // forbid double words in compounds
- checkcompoundrep = 0; // forbid bad compounds (may be non compound word with a REP substitution)
- checkcompoundcase = 0; // forbid upper and lowercase combinations at word bounds
- checkcompoundtriple = 0; // forbid compounds with triple letters
- forbiddenword = FLAG_NULL; // forbidden word signing flag
- nosuggest = FLAG_NULL; // don't suggest words signed with NOSUGGEST flag
- lang = NULL; // language
- langnum = 0; // language code (see http://l10n.openoffice.org/languages.html)
- pseudoroot = FLAG_NULL; // forbidden root, allowed only with suffixes
- cpdwordmax = -1; // default: unlimited wordcount in compound words
- cpdmin = -1; // undefined
- cpdmaxsyllable = 0; // default: unlimited syllablecount in compound words
- cpdvowels=NULL; // vowels (for calculating of Hungarian compounding limit, O(n) search! XXX)
- cpdvowels_utf16=NULL; // vowels for UTF-8 encoding (bsearch instead of O(n) search)
- cpdvowels_utf16_len=0; // vowels
- pfxappnd=NULL; // previous prefix for counting the syllables of prefix BUG
- sfxappnd=NULL; // previous suffix for counting a special syllables BUG
- cpdsyllablenum=NULL; // syllable count incrementing flag
- checknum=0; // checking numbers, and word with numbers
- wordchars=NULL; // letters + spec. word characters
- wordchars_utf16=NULL; // letters + spec. word characters
- wordchars_utf16_len=0; // letters + spec. word characters
- ignorechars=NULL; // letters + spec. word characters
- ignorechars_utf16=NULL; // letters + spec. word characters
- ignorechars_utf16_len=0; // letters + spec. word characters
- version=NULL; // affix and dictionary file version string
- havecontclass=0; // flags of possible continuing classes (double affix)
- // LEMMA_PRESENT: not put root into the morphological output. Lemma presents
- // in morhological description in dictionary file. It's often combined with PSEUDOROOT.
- lemma_present = FLAG_NULL;
- circumfix = FLAG_NULL;
- onlyincompound = FLAG_NULL;
- flag_mode = FLAG_CHAR; // default one-character flags in affix and dic file
- maxngramsugs = -1; // undefined
- nosplitsugs = 0;
- sugswithdots = 0;
- keepcase = 0;
- checksharps = 0;
-
- derived = NULL; // XXX not threadsafe variable for experimental stemming
- sfx = NULL;
- pfx = NULL;
-
- for (int i=0; i < SETSIZE; i++) {
- pStart[i] = NULL;
- sStart[i] = NULL;
- pFlag[i] = NULL;
- sFlag[i] = NULL;
- }
-
- for (int j=0; j < CONTSIZE; j++) {
- contclasses[j] = 0;
- }
-
- if (parse_file(affpath)) {
- HUNSPELL_WARNING(stderr, "Failure loading aff file %s\n",affpath);
- wordchars = mystrdup("qwertzuiopasdfghjklyxcvbnmQWERTZUIOPASDFGHJKLYXCVBNM");
- }
-
- if (cpdmin == -1) cpdmin = MINCPDLEN;
-
- }
-
-
- AffixMgr::~AffixMgr()
- {
-
- // pass through linked prefix entries and clean up
- for (int i=0; i < SETSIZE ;i++) {
- pFlag[i] = NULL;
- PfxEntry * ptr = (PfxEntry *)pStart[i];
- PfxEntry * nptr = NULL;
- while (ptr) {
- nptr = ptr->getNext();
- delete(ptr);
- ptr = nptr;
- nptr = NULL;
- }
- }
-
- // pass through linked suffix entries and clean up
- for (int j=0; j < SETSIZE ; j++) {
- sFlag[j] = NULL;
- SfxEntry * ptr = (SfxEntry *)sStart[j];
- SfxEntry * nptr = NULL;
- while (ptr) {
- nptr = ptr->getNext();
- delete(ptr);
- ptr = nptr;
- nptr = NULL;
- }
- sStart[j] = NULL;
- }
-
- if (trystring) free(trystring);
- trystring=NULL;
- if (encoding) free(encoding);
- encoding=NULL;
- if (maptable) {
- for (int j=0; j < nummap; j++) {
- if (maptable[j].set) free(maptable[j].set);
- if (maptable[j].set_utf16) free(maptable[j].set_utf16);
- maptable[j].set = NULL;
- maptable[j].len = 0;
- }
- free(maptable);
- maptable = NULL;
- }
- nummap = 0;
- if (breaktable) {
- for (int j=0; j < numbreak; j++) {
- if (breaktable[j]) free(breaktable[j]);
- breaktable[j] = NULL;
- }
- free(breaktable);
- breaktable = NULL;
- }
- numbreak = 0;
- if (reptable) {
- for (int j=0; j < numrep; j++) {
- free(reptable[j].pattern);
- free(reptable[j].pattern2);
- reptable[j].pattern = NULL;
- reptable[j].pattern2 = NULL;
- }
- free(reptable);
- reptable = NULL;
- }
- if (defcpdtable) {
- for (int j=0; j < numdefcpd; j++) {
- free(defcpdtable[j].def);
- defcpdtable[j].def = NULL;
- }
- free(defcpdtable);
- defcpdtable = NULL;
- }
- numrep = 0;
- if (checkcpdtable) {
- for (int j=0; j < numcheckcpd; j++) {
- free(checkcpdtable[j].pattern);
- free(checkcpdtable[j].pattern2);
- checkcpdtable[j].pattern = NULL;
- checkcpdtable[j].pattern2 = NULL;
- }
- free(checkcpdtable);
- checkcpdtable = NULL;
- }
- numcheckcpd = 0;
- FREE_FLAG(compoundflag);
- FREE_FLAG(compoundbegin);
- FREE_FLAG(compoundmiddle);
- FREE_FLAG(compoundend);
- FREE_FLAG(compoundpermitflag);
- FREE_FLAG(compoundforbidflag);
- FREE_FLAG(compoundroot);
- FREE_FLAG(forbiddenword);
- FREE_FLAG(nosuggest);
- FREE_FLAG(pseudoroot);
- FREE_FLAG(lemma_present);
- FREE_FLAG(circumfix);
- FREE_FLAG(onlyincompound);
-
- cpdwordmax = 0;
- pHMgr = NULL;
- cpdmin = 0;
- cpdmaxsyllable = 0;
- if (cpdvowels) free(cpdvowels);
- if (cpdvowels_utf16) free(cpdvowels_utf16);
- if (cpdsyllablenum) free(cpdsyllablenum);
- free_utf_tbl();
- if (lang) free(lang);
- if (wordchars) free(wordchars);
- if (wordchars_utf16) free(wordchars_utf16);
- if (ignorechars) free(ignorechars);
- if (ignorechars_utf16) free(ignorechars_utf16);
- if (version) free(version);
- if (derived) free(derived);
- checknum=0;
- }
-
-
- // read in aff file and build up prefix and suffix entry objects
- int AffixMgr::parse_file(const char * affpath)
- {
-
- // io buffers
- char line[MAXLNLEN+1];
-
- // affix type
- char ft;
-
- // checking flag duplication
- char dupflags[CONTSIZE];
- char dupflags_ini = 1;
-
- // first line indicator for removing byte order mark
- int firstline = 1;
-
- // open the affix file
- FILE * afflst;
- afflst = fopen(affpath,"r");
- if (!afflst) {
- HUNSPELL_WARNING(stderr, "error: could not open affix description file %s\n",affpath);
- return 1;
- }
-
- // step one is to parse the affix file building up the internal
- // affix data structures
-
-
- // read in each line ignoring any that do not
- // start with a known line type indicator
- while (fgets(line,MAXLNLEN,afflst)) {
- mychomp(line);
-
- /* remove byte order mark */
- if (firstline) {
- firstline = 0;
- if (strncmp(line,"",3) == 0) {
- memmove(line, line+3, strlen(line+3)+1);
- HUNSPELL_WARNING(stderr, "warning: affix file begins with byte order mark: possible incompatibility with old Hunspell versions\n");
- }
- }
-
- /* parse in the try string */
- if (strncmp(line,"TRY",3) == 0) {
- if (parse_string(line, &trystring, "TRY")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the name of the character set used by the .dict and .aff */
- if (strncmp(line,"SET",3) == 0) {
- if (parse_string(line, &encoding, "SET")) {
- fclose(afflst);
- return 1;
- }
- if (strcmp(encoding, "UTF-8") == 0) {
- utf8 = 1;
- #ifndef OPENOFFICEORG
- #ifndef MOZILLA_CLIENT
- if (initialize_utf_tbl()) {
- fclose(afflst);
- return 1;
- }
- #endif
- #endif
- }
- }
-
- /* parse COMPLEXPREFIXES for agglutinative languages with right-to-left writing system */
- if (strncmp(line,"COMPLEXPREFIXES",15) == 0)
- complexprefixes = 1;
-
- /* parse in the flag used by the controlled compound words */
- if (strncmp(line,"COMPOUNDFLAG",12) == 0) {
- if (parse_flag(line, &compoundflag, "COMPOUNDFLAG")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by compound words */
- if (strncmp(line,"COMPOUNDBEGIN",13) == 0) {
- if (complexprefixes) {
- if (parse_flag(line, &compoundend, "COMPOUNDBEGIN")) {
- fclose(afflst);
- return 1;
- }
- } else {
- if (parse_flag(line, &compoundbegin, "COMPOUNDBEGIN")) {
- fclose(afflst);
- return 1;
- }
- }
- }
-
- /* parse in the flag used by compound words */
- if (strncmp(line,"COMPOUNDMIDDLE",14) == 0) {
- if (parse_flag(line, &compoundmiddle, "COMPOUNDMIDDLE")) {
- fclose(afflst);
- return 1;
- }
- }
- /* parse in the flag used by compound words */
- if (strncmp(line,"COMPOUNDEND",11) == 0) {
- if (complexprefixes) {
- if (parse_flag(line, &compoundbegin, "COMPOUNDEND")) {
- fclose(afflst);
- return 1;
- }
- } else {
- if (parse_flag(line, &compoundend, "COMPOUNDEND")) {
- fclose(afflst);
- return 1;
- }
- }
- }
-
- /* parse in the data used by compound_check() method */
- if (strncmp(line,"COMPOUNDWORDMAX",15) == 0) {
- if (parse_num(line, &cpdwordmax, "COMPOUNDWORDMAX")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag sign compounds in dictionary */
- if (strncmp(line,"COMPOUNDROOT",12) == 0) {
- if (parse_flag(line, &compoundroot, "COMPOUNDROOT")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by compound_check() method */
- if (strncmp(line,"COMPOUNDPERMITFLAG",18) == 0) {
- if (parse_flag(line, &compoundpermitflag, "COMPOUNDPERMITFLAG")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by compound_check() method */
- if (strncmp(line,"COMPOUNDFORBIDFLAG",18) == 0) {
- if (parse_flag(line, &compoundforbidflag, "COMPOUNDFORBIDFLAG")) {
- fclose(afflst);
- return 1;
- }
- }
-
- if (strncmp(line,"CHECKCOMPOUNDDUP",16) == 0) {
- checkcompounddup = 1;
- }
-
- if (strncmp(line,"CHECKCOMPOUNDREP",16) == 0) {
- checkcompoundrep = 1;
- }
-
- if (strncmp(line,"CHECKCOMPOUNDTRIPLE",19) == 0) {
- checkcompoundtriple = 1;
- }
-
- if (strncmp(line,"CHECKCOMPOUNDCASE",17) == 0) {
- checkcompoundcase = 1;
- }
-
- if (strncmp(line,"NOSUGGEST",9) == 0) {
- if (parse_flag(line, &nosuggest, "NOSUGGEST")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by forbidden words */
- if (strncmp(line,"FORBIDDENWORD",13) == 0) {
- if (parse_flag(line, &forbiddenword, "FORBIDDENWORD")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by forbidden words */
- if (strncmp(line,"LEMMA_PRESENT",13) == 0) {
- if (parse_flag(line, &lemma_present, "LEMMA_PRESENT")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by circumfixes */
- if (strncmp(line,"CIRCUMFIX",9) == 0) {
- if (parse_flag(line, &circumfix, "CIRCUMFIX")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by fogemorphemes */
- if (strncmp(line,"ONLYINCOMPOUND",14) == 0) {
- if (parse_flag(line, &onlyincompound, "ONLYINCOMPOUND")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by `pseudoroots' */
- if (strncmp(line,"PSEUDOROOT",10) == 0) {
- if (parse_flag(line, &pseudoroot, "PSEUDOROOT")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by `pseudoroots' */
- if (strncmp(line,"NEEDAFFIX",9) == 0) {
- if (parse_flag(line, &pseudoroot, "NEEDAFFIX")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the minimal length for words in compounds */
- if (strncmp(line,"COMPOUNDMIN",11) == 0) {
- if (parse_num(line, &cpdmin, "COMPOUNDMIN")) {
- fclose(afflst);
- return 1;
- }
- if (cpdmin < 1) cpdmin = 1;
- }
-
- /* parse in the max. words and syllables in compounds */
- if (strncmp(line,"COMPOUNDSYLLABLE",16) == 0) {
- if (parse_cpdsyllable(line)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by compound_check() method */
- if (strncmp(line,"SYLLABLENUM",11) == 0) {
- if (parse_string(line, &cpdsyllablenum, "SYLLABLENUM")) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the flag used by the controlled compound words */
- if (strncmp(line,"CHECKNUM",8) == 0) {
- checknum=1;
- }
-
- /* parse in the extra word characters */
- if (strncmp(line,"WORDCHARS",9) == 0) {
- if (parse_array(line, &wordchars, &wordchars_utf16, &wordchars_utf16_len, "WORDCHARS", utf8)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the ignored characters (for example, Arabic optional diacretics charachters */
- if (strncmp(line,"IGNORE",6) == 0) {
- if (parse_array(line, &ignorechars, &ignorechars_utf16, &ignorechars_utf16_len, "IGNORE", utf8)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the typical fault correcting table */
- if (strncmp(line,"REP",3) == 0) {
- if (parse_reptable(line, afflst)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the checkcompoundpattern table */
- if (strncmp(line,"CHECKCOMPOUNDPATTERN",20) == 0) {
- if (parse_checkcpdtable(line, afflst)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the defcompound table */
- if (strncmp(line,"COMPOUNDRULE",12) == 0) {
- if (parse_defcpdtable(line, afflst)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the related character map table */
- if (strncmp(line,"MAP",3) == 0) {
- if (parse_maptable(line, afflst)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the word breakpoints table */
- if (strncmp(line,"BREAK",5) == 0) {
- if (parse_breaktable(line, afflst)) {
- fclose(afflst);
- return 1;
- }
- }
-
- /* parse in the language for language specific codes */
- if (strncmp(line,"LANG",4) == 0) {
- if (parse_string(line, &lang, "LANG")) {
- fclose(afflst);
- return 1;
- }
- langnum = get_lang_num(lang);
- }
-
- if (strncmp(line,"VERSION",7) == 0) {
- if (parse_string(line, &version, "VERSION")) {
- fclose(afflst);
- return 1;
- }
- }
-
- if (strncmp(line,"MAXNGRAMSUGS",12) == 0) {
- if (parse_num(line, &maxngramsugs, "MAXNGRAMSUGS")) {
- fclose(afflst);
- return 1;
- }
- }
-
- if (strncmp(line,"NOSPLITSUGS",11) == 0) {
- nosplitsugs=1;
- }
-
- if (strncmp(line,"SUGSWITHDOTS",12) == 0) {
- sugswithdots=1;
- }
-
- /* parse in the flag used by forbidden words */
- if (strncmp(line,"KEEPCASE",8) == 0) {
- if (parse_flag(line, &keepcase, "KEEPCASE")) {
- fclose(afflst);
- return 1;
- }
- }
-
- if (strncmp(line,"CHECKSHARPS",11) == 0) {
- checksharps=1;
- }
-
- /* parse this affix: P - prefix, S - suffix */
- ft = ' ';
- if (strncmp(line,"PFX",3) == 0) ft = complexprefixes ? 'S' : 'P';
- if (strncmp(line,"SFX",3) == 0) ft = complexprefixes ? 'P' : 'S';
- if (ft != ' ') {
- if (dupflags_ini) {
- for (int i = 0; i < CONTSIZE; i++) dupflags[i] = 0;
- dupflags_ini = 0;
- }
- if (parse_affix(line, ft, afflst, dupflags)) {
- fclose(afflst);
- process_pfx_tree_to_list();
- process_sfx_tree_to_list();
- return 1;
- }
- }
-
- }
- fclose(afflst);
-
- // convert affix trees to sorted list
- process_pfx_tree_to_list();
- process_sfx_tree_to_list();
-
- // now we can speed up performance greatly taking advantage of the
- // relationship between the affixes and the idea of "subsets".
-
- // View each prefix as a potential leading subset of another and view
- // each suffix (reversed) as a potential trailing subset of another.
-
- // To illustrate this relationship if we know the prefix "ab" is found in the
- // word to examine, only prefixes that "ab" is a leading subset of need be examined.
- // Furthermore is "ab" is not present then none of the prefixes that "ab" is
- // is a subset need be examined.
- // The same argument goes for suffix string that are reversed.
-
- // Then to top this off why not examine the first char of the word to quickly
- // limit the set of prefixes to examine (i.e. the prefixes to examine must
- // be leading supersets of the first character of the word (if they exist)
-
- // To take advantage of this "subset" relationship, we need to add two links
- // from entry. One to take next if the current prefix is found (call it nexteq)
- // and one to take next if the current prefix is not found (call it nextne).
-
- // Since we have built ordered lists, all that remains is to properly intialize
- // the nextne and nexteq pointers that relate them
-
- process_pfx_order();
- process_sfx_order();
-
- // expand wordchars string, based on csutil (for external tokenization)
-
- char * enc = get_encoding();
- csconv = get_current_cs(enc);
- free(enc);
- enc = NULL;
-
- char expw[MAXLNLEN];
- if (wordchars) {
- strcpy(expw, wordchars);
- free(wordchars);
- } else *expw = '\0';
-
- for (int i = 0; i <= 255; i++) {
- if ( (csconv[i].cupper != csconv[i].clower) &&
- (! strchr(expw, (char) i))) {
- *(expw + strlen(expw) + 1) = '\0';
- *(expw + strlen(expw)) = (char) i;
- }
- }
-
- wordchars = mystrdup(expw);
-
- // temporary BREAK definition for German dash handling (OOo issue 64400)
- if ((langnum == LANG_de) && (!breaktable)) {
- breaktable = (char **) malloc(sizeof(char *));
- if (!breaktable) return 1;
- breaktable[0] = mystrdup("-");
- numbreak = 1;
- }
- return 0;
- }
-
-
- // we want to be able to quickly access prefix information
- // both by prefix flag, and sorted by prefix string itself
- // so we need to set up two indexes
-
- int AffixMgr::build_pfxtree(AffEntry* pfxptr)
- {
- PfxEntry * ptr;
- PfxEntry * pptr;
- PfxEntry * ep = (PfxEntry*) pfxptr;
-
- // get the right starting points
- const char * key = ep->getKey();
- const unsigned char flg = (unsigned char) (ep->getFlag() & 0x00FF);
-
- // first index by flag which must exist
- ptr = (PfxEntry*)pFlag[flg];
- ep->setFlgNxt(ptr);
- pFlag[flg] = (AffEntry *) ep;
-
-
- // handle the special case of null affix string
- if (strlen(key) == 0) {
- // always inset them at head of list at element 0
- ptr = (PfxEntry*)pStart[0];
- ep->setNext(ptr);
- pStart[0] = (AffEntry*)ep;
- return 0;
- }
-
- // now handle the normal case
- ep->setNextEQ(NULL);
- ep->setNextNE(NULL);
-
- unsigned char sp = *((const unsigned char *)key);
- ptr = (PfxEntry*)pStart[sp];
-
- // handle the first insert
- if (!ptr) {
- pStart[sp] = (AffEntry*)ep;
- return 0;
- }
-
-
- // otherwise use binary tree insertion so that a sorted
- // list can easily be generated later
- pptr = NULL;
- for (;;) {
- pptr = ptr;
- if (strcmp(ep->getKey(), ptr->getKey() ) <= 0) {
- ptr = ptr->getNextEQ();
- if (!ptr) {
- pptr->setNextEQ(ep);
- break;
- }
- } else {
- ptr = ptr->getNextNE();
- if (!ptr) {
- pptr->setNextNE(ep);
- break;
- }
- }
- }
- return 0;
- }
-
- // we want to be able to quickly access suffix information
- // both by suffix flag, and sorted by the reverse of the
- // suffix string itself; so we need to set up two indexes
- int AffixMgr::build_sfxtree(AffEntry* sfxptr)
- {
- SfxEntry * ptr;
- SfxEntry * pptr;
- SfxEntry * ep = (SfxEntry *) sfxptr;
-
- /* get the right starting point */
- const char * key = ep->getKey();
- const unsigned char flg = (unsigned char) (ep->getFlag() & 0x00FF);
-
- // first index by flag which must exist
- ptr = (SfxEntry*)sFlag[flg];
- ep->setFlgNxt(ptr);
- sFlag[flg] = (AffEntry *) ep;
-
- // next index by affix string
-
- // handle the special case of null affix string
- if (strlen(key) == 0) {
- // always inset them at head of list at element 0
- ptr = (SfxEntry*)sStart[0];
- ep->setNext(ptr);
- sStart[0] = (AffEntry*)ep;
- return 0;
- }
-
- // now handle the normal case
- ep->setNextEQ(NULL);
- ep->setNextNE(NULL);
-
- unsigned char sp = *((const unsigned char *)key);
- ptr = (SfxEntry*)sStart[sp];
-
- // handle the first insert
- if (!ptr) {
- sStart[sp] = (AffEntry*)ep;
- return 0;
- }
-
- // otherwise use binary tree insertion so that a sorted
- // list can easily be generated later
- pptr = NULL;
- for (;;) {
- pptr = ptr;
- if (strcmp(ep->getKey(), ptr->getKey() ) <= 0) {
- ptr = ptr->getNextEQ();
- if (!ptr) {
- pptr->setNextEQ(ep);
- break;
- }
- } else {
- ptr = ptr->getNextNE();
- if (!ptr) {
- pptr->setNextNE(ep);
- break;
- }
- }
- }
- return 0;
- }
-
- // convert from binary tree to sorted list
- int AffixMgr::process_pfx_tree_to_list()
- {
- for (int i=1; i< SETSIZE; i++) {
- pStart[i] = process_pfx_in_order(pStart[i],NULL);
- }
- return 0;
- }
-
-
- AffEntry* AffixMgr::process_pfx_in_order(AffEntry* ptr, AffEntry* nptr)
- {
- if (ptr) {
- nptr = process_pfx_in_order(((PfxEntry*) ptr)->getNextNE(), nptr);
- ((PfxEntry*) ptr)->setNext((PfxEntry*) nptr);
- nptr = process_pfx_in_order(((PfxEntry*) ptr)->getNextEQ(), ptr);
- }
- return nptr;
- }
-
-
- // convert from binary tree to sorted list
- int AffixMgr:: process_sfx_tree_to_list()
- {
- for (int i=1; i< SETSIZE; i++) {
- sStart[i] = process_sfx_in_order(sStart[i],NULL);
- }
- return 0;
- }
-
- AffEntry* AffixMgr::process_sfx_in_order(AffEntry* ptr, AffEntry* nptr)
- {
- if (ptr) {
- nptr = process_sfx_in_order(((SfxEntry*) ptr)->getNextNE(), nptr);
- ((SfxEntry*) ptr)->setNext((SfxEntry*) nptr);
- nptr = process_sfx_in_order(((SfxEntry*) ptr)->getNextEQ(), ptr);
- }
- return nptr;
- }
-
-
- // reinitialize the PfxEntry links NextEQ and NextNE to speed searching
- // using the idea of leading subsets this time
- int AffixMgr::process_pfx_order()
- {
- PfxEntry* ptr;
-
- // loop through each prefix list starting point
- for (int i=1; i < SETSIZE; i++) {
-
- ptr = (PfxEntry*)pStart[i];
-
- // look through the remainder of the list
- // and find next entry with affix that
- // the current one is not a subset of
- // mark that as destination for NextNE
- // use next in list that you are a subset
- // of as NextEQ
-
- for (; ptr != NULL; ptr = ptr->getNext()) {
-
- PfxEntry * nptr = ptr->getNext();
- for (; nptr != NULL; nptr = nptr->getNext()) {
- if (! isSubset( ptr->getKey() , nptr->getKey() )) break;
- }
- ptr->setNextNE(nptr);
- ptr->setNextEQ(NULL);
- if ((ptr->getNext()) && isSubset(ptr->getKey() , (ptr->getNext())->getKey()))
- ptr->setNextEQ(ptr->getNext());
- }
-
- // now clean up by adding smart search termination strings:
- // if you are already a superset of the previous prefix
- // but not a subset of the next, search can end here
- // so set NextNE properly
-
- ptr = (PfxEntry *) pStart[i];
- for (; ptr != NULL; ptr = ptr->getNext()) {
- PfxEntry * nptr = ptr->getNext();
- PfxEntry * mptr = NULL;
- for (; nptr != NULL; nptr = nptr->getNext()) {
- if (! isSubset(ptr->getKey(),nptr->getKey())) break;
- mptr = nptr;
- }
- if (mptr) mptr->setNextNE(NULL);
- }
- }
- return 0;
- }
-
- // initialize the SfxEntry links NextEQ and NextNE to speed searching
- // using the idea of leading subsets this time
- int AffixMgr::process_sfx_order()
- {
- SfxEntry* ptr;
-
- // loop through each prefix list starting point
- for (int i=1; i < SETSIZE; i++) {
-
- ptr = (SfxEntry *) sStart[i];
-
- // look through the remainder of the list
- // and find next entry with affix that
- // the current one is not a subset of
- // mark that as destination for NextNE
- // use next in list that you are a subset
- // of as NextEQ
-
- for (; ptr != NULL; ptr = ptr->getNext()) {
- SfxEntry * nptr = ptr->getNext();
- for (; nptr != NULL; nptr = nptr->getNext()) {
- if (! isSubset(ptr->getKey(),nptr->getKey())) break;
- }
- ptr->setNextNE(nptr);
- ptr->setNextEQ(NULL);
- if ((ptr->getNext()) && isSubset(ptr->getKey(),(ptr->getNext())->getKey()))
- ptr->setNextEQ(ptr->getNext());
- }
-
-
- // now clean up by adding smart search termination strings:
- // if you are already a superset of the previous suffix
- // but not a subset of the next, search can end here
- // so set NextNE properly
-
- ptr = (SfxEntry *) sStart[i];
- for (; ptr != NULL; ptr = ptr->getNext()) {
- SfxEntry * nptr = ptr->getNext();
- SfxEntry * mptr = NULL;
- for (; nptr != NULL; nptr = nptr->getNext()) {
- if (! isSubset(ptr->getKey(),nptr->getKey())) break;
- mptr = nptr;
- }
- if (mptr) mptr->setNextNE(NULL);
- }
- }
- return 0;
- }
-
-
-
- // takes aff file condition string and creates the
- // conds array - please see the appendix at the end of the
- // file affentry.cxx which describes what is going on here
- // in much more detail
-
- int AffixMgr::encodeit(struct affentry * ptr, char * cs)
- {
- unsigned char c;
- int i, j, k;
- unsigned char mbr[MAXLNLEN];
- w_char wmbr[MAXLNLEN];
- w_char * wpos = wmbr;
-
- // now clear the conditions array */
- for (i=0;i<SETSIZE;i++) ptr->conds.base[i] = (unsigned char) 0;
-
- // now parse the string to create the conds array */
- int nc = strlen(cs);
- unsigned char neg = 0; // complement indicator
- int grp = 0; // group indicator
- unsigned char n = 0; // number of conditions
- int ec = 0; // end condition indicator
- int nm = 0; // number of member in group
-
- // if no condition just return
- if (strcmp(cs,".")==0) {
- ptr->numconds = 0;
- return 0;
- }
-
- i = 0;
- while (i < nc) {
- c = *((unsigned char *)(cs + i));
-
- // start group indicator
- if (c == '[') {
- grp = 1;
- c = 0;
- }
-
- // complement flag
- if ((grp == 1) && (c == '^')) {
- neg = 1;
- c = 0;
- }
-
- // end goup indicator
- if (c == ']') {
- ec = 1;
- c = 0;
- }
-
- // add character of group to list
- if ((grp == 1) && (c != 0)) {
- *(mbr + nm) = c;
- nm++;
- c = 0;
- }
-
- // end of condition
- if (c != 0) {
- ec = 1;
- }
-
- if (ec) {
- if (!utf8) {
- if (grp == 1) {
- if (neg == 0) {
- // set the proper bits in the condition array vals for those chars
- for (j=0;j<nm;j++) {
- k = (unsigned int) mbr[j];
- ptr->conds.base[k] = ptr->conds.base[k] | ((unsigned char)1 << n);
- }
- } else {
- // complement so set all of them and then unset indicated ones
- for (j=0;j<SETSIZE;j++) ptr->conds.base[j] = ptr->conds.base[j] | ((unsigned char)1 << n);
- for (j=0;j<nm;j++) {
- k = (unsigned int) mbr[j];
- ptr->conds.base[k] = ptr->conds.base[k] & ~((unsigned char)1 << n);
- }
- }
- neg = 0;
- grp = 0;
- nm = 0;
- } else {
- // not a group so just set the proper bit for this char
- // but first handle special case of . inside condition
- if (c == '.') {
- // wild card character so set them all
- for (j=0;j<SETSIZE;j++) ptr->conds.base[j] = ptr->conds.base[j] | ((unsigned char)1 << n);
- } else {
- ptr->conds.base[(unsigned int) c] = ptr->conds.base[(unsigned int)c] | ((unsigned char)1 << n);
- }
- }
- n++;
- ec = 0;
- } else { // UTF-8 character set
- if (grp == 1) {
- ptr->conds.utf8.neg[n] = neg;
- if (neg == 0) {
- // set the proper bits in the condition array vals for those chars
- for (j=0;j<nm;j++) {
- k = (unsigned int) mbr[j];
- if (k >> 7) {
- u8_u16(wpos, 1, (char *) mbr + j);
- wpos++;
- if ((k & 0xe0) == 0xe0) j+=2; else j++; // 3-byte UTF-8 character
- } else {
- ptr->conds.utf8.ascii[k] = ptr->conds.utf8.ascii[k] | ((unsigned char)1 << n);
- }
- }
- } else { // neg == 1
- // complement so set all of them and then unset indicated ones
- for (j=0;j<(SETSIZE/2);j++) ptr->conds.utf8.ascii[j] = ptr->conds.utf8.ascii[j] | ((unsigned char)1 << n);
- for (j=0;j<nm;j++) {
- k = (unsigned int) mbr[j];
- if (k >> 7) {
- u8_u16(wpos, 1, (char *) mbr + j);
- wpos++;
- if ((k & 0xe0) == 0xe0) j+=2; else j++; // 3-byte UTF-8 character
- } else {
- ptr->conds.utf8.ascii[k] = ptr->conds.utf8.ascii[k] & ~((unsigned char)1 << n);
- }
- }
- }
- neg = 0;
- grp = 0;
- nm = 0;
- ptr->conds.utf8.wlen[n] = wpos - wmbr;
- if ((wpos - wmbr) != 0) {
- ptr->conds.utf8.wchars[n] = (w_char *) malloc(sizeof(w_char) * (wpos - wmbr));
- if (!ptr->conds.utf8.wchars[n]) return 1;
- memcpy(ptr->conds.utf8.wchars[n], wmbr, sizeof(w_char) * (wpos - wmbr));
- flag_qsort((unsigned short *) ptr->conds.utf8.wchars[n], 0, ptr->conds.utf8.wlen[n]);
- wpos = wmbr;
- }
- } else { // grp == 0
- // is UTF-8 character?
- if (c >> 7) {
- ptr->conds.utf8.wchars[n] = (w_char *) malloc(sizeof(w_char));
- if (!ptr->conds.utf8.wchars[n]) return 1;
- ptr->conds.utf8.wlen[n] = 1;
- u8_u16(ptr->conds.utf8.wchars[n], 1, cs + i);
- if ((c & 0xe0) == 0xe0) i+=2; else i++; // 3-byte UFT-8 character
- } else {
- ptr->conds.utf8.wchars[n] = NULL;
- // not a group so just set the proper bit for this char
- // but first handle special case of . inside condition
- if (c == '.') {
- ptr->conds.utf8.all[n] = 1;
- // wild card character so set them all
- for (j=0;j<(SETSIZE/2);j++) ptr->conds.utf8.ascii[j] = ptr->conds.utf8.ascii[j] | ((unsigned char)1 << n);
- } else {
- ptr->conds.utf8.all[n] = 0;
- ptr->conds.utf8.ascii[(unsigned int) c] = ptr->conds.utf8.ascii[(unsigned int)c] | ((unsigned char)1 << n);
- }
- }
- neg = 0;
- }
- n++;
- ec = 0;
- neg = 0;
- }
- }
-
- i++;
- }
- ptr->numconds = n;
- return 0;
- }
-
- // return 1 if s1 is a leading subset of s2
- /* inline int AffixMgr::isSubset(const char * s1, const char * s2)
- {
- while ((*s1 == *s2) && *s1) {
- s1++;
- s2++;
- }
- return (*s1 == '\0');
- }
- */
-
- // return 1 if s1 is a leading subset of s2 (dots are for infixes)
- inline int AffixMgr::isSubset(const char * s1, const char * s2)
- {
- while (((*s1 == *s2) || (*s1 == '.')) && (*s1 != '\0')) {
- s1++;
- s2++;
- }
- return (*s1 == '\0');
- }
-
-
- // check word for prefixes
- struct hentry * AffixMgr::prefix_check(const char * word, int len, char in_compound,
- const FLAG needflag)
- {
- struct hentry * rv= NULL;
-
- pfx = NULL;
- pfxappnd = NULL;
- sfxappnd = NULL;
-
- // first handle the special case of 0 length prefixes
- PfxEntry * pe = (PfxEntry *) pStart[0];
- while (pe) {
- if (
- // fogemorpheme
- ((in_compound != IN_CPD_NOT) || !(pe->getCont() &&
- (TESTAFF(pe->getCont(), onlyincompound, pe->getContLen())))) &&
- // permit prefixes in compounds
- ((in_compound != IN_CPD_END) || (pe->getCont() &&
- (TESTAFF(pe->getCont(), compoundpermitflag, pe->getContLen()))))
- ) {
- // check prefix
- rv = pe->checkword(word, len, in_compound, needflag);
- if (rv) {
- pfx=(AffEntry *)pe; // BUG: pfx not stateless
- return rv;
- }
- }
- pe = pe->getNext();
- }
-
- // now handle the general case
- unsigned char sp = *((const unsigned char *)word);
- PfxEntry * pptr = (PfxEntry *)pStart[sp];
-
- while (pptr) {
- if (isSubset(pptr->getKey(),word)) {
- if (
- // fogemorpheme
- ((in_compound != IN_CPD_NOT) || !(pptr->getCont() &&
- (TESTAFF(pptr->getCont(), onlyincompound, pptr->getContLen())))) &&
- // permit prefixes in compounds
- ((in_compound != IN_CPD_END) || (pptr->getCont() &&
- (TESTAFF(pptr->getCont(), compoundpermitflag, pptr->getContLen()))))
- ) {
- // check prefix
- rv = pptr->checkword(word, len, in_compound, needflag);
- if (rv) {
- pfx=(AffEntry *)pptr; // BUG: pfx not stateless
- return rv;
- }
- }
- pptr = pptr->getNextEQ();
- } else {
- pptr = pptr->getNextNE();
- }
- }
-
- return NULL;
- }
-
- // check word for prefixes
- struct hentry * AffixMgr::prefix_check_twosfx(const char * word, int len,
- char in_compound, const FLAG needflag)
- {
- struct hentry * rv= NULL;
-
- pfx = NULL;
- sfxappnd = NULL;
-
- // first handle the special case of 0 length prefixes
- PfxEntry * pe = (PfxEntry *) pStart[0];
-
- while (pe) {
- rv = pe->check_twosfx(word, len, in_compound, needflag);
- if (rv) return rv;
- pe = pe->getNext();
- }
-
- // now handle the general case
- unsigned char sp = *((const unsigned char *)word);
- PfxEntry * pptr = (PfxEntry *)pStart[sp];
-
- while (pptr) {
- if (isSubset(pptr->getKey(),word)) {
- rv = pptr->check_twosfx(word, len, in_compound, needflag);
- if (rv) {
- pfx = (AffEntry *)pptr;
- return rv;
- }
- pptr = pptr->getNextEQ();
- } else {
- pptr = pptr->getNextNE();
- }
- }
-
- return NULL;
- }
-
- #ifdef HUNSPELL_EXPERIMENTAL
- // check word for prefixes
- char * AffixMgr::prefix_check_morph(const char * word, int len, char in_compound,
- const FLAG needflag)
- {
- char * st;
-
- char result[MAXLNLEN];
- result[0] = '\0';
-
- pfx = NULL;
- sfxappnd = NULL;
-
- // first handle the special case of 0 length prefixes
- PfxEntry * pe = (PfxEntry *) pStart[0];
- while (pe) {
- st = pe->check_morph(word,len,in_compound, needflag);
- if (st) {
- strcat(result, st);
- free(st);
- }
- // if (rv) return rv;
- pe = pe->getNext();
- }
-
- // now handle the general case
- unsigned char sp = *((const unsigned char *)word);
- PfxEntry * pptr = (PfxEntry *)pStart[sp];
-
- while (pptr) {
- if (isSubset(pptr->getKey(),word)) {
- st = pptr->check_morph(word,len,in_compound, needflag);
- if (st) {
- // fogemorpheme
- if ((in_compound != IN_CPD_NOT) || !((pptr->getCont() &&
- (TESTAFF(pptr->getCont(), onlyincompound, pptr->getContLen()))))) {
- strcat(result, st);
- pfx = (AffEntry *)pptr;
- }
- free(st);
- }
- pptr = pptr->getNextEQ();
- } else {
- pptr = pptr->getNextNE();
- }
- }
-
- if (*result) return mystrdup(result);
- return NULL;
- }
-
-
- // check word for prefixes
- char * AffixMgr::prefix_check_twosfx_morph(const char * word, int len,
- char in_compound, const FLAG needflag)
- {
- char * st;
-
- char result[MAXLNLEN];
- result[0] = '\0';
-
- pfx = NULL;
- sfxappnd = NULL;
-
- // first handle the special case of 0 length prefixes
- PfxEntry * pe = (PfxEntry *) pStart[0];
- while (pe) {
- st = pe->check_twosfx_morph(word,len,in_compound, needflag);
- if (st) {
- strcat(result, st);
- free(st);
- }
- pe = pe->getNext();
- }
-
- // now handle the general case
- unsigned char sp = *((const unsigned char *)word);
- PfxEntry * pptr = (PfxEntry *)pStart[sp];
-
- while (pptr) {
- if (isSubset(pptr->getKey(),word)) {
- st = pptr->check_twosfx_morph(word, len, in_compound, needflag);
- if (st) {
- strcat(result, st);
- free(st);
- pfx = (AffEntry *)pptr;
- }
- pptr = pptr->getNextEQ();
- } else {
- pptr = pptr->getNextNE();
- }
- }
-
- if (*result) return mystrdup(result);
- return NULL;
- }
- #endif // END OF HUNSPELL_EXPERIMENTAL CODE
-
-
- // Is word a non compound with a REP substitution (see checkcompoundrep)?
- int AffixMgr::cpdrep_check(const char * word, int wl)
- {
- char candidate[MAXLNLEN];
- const char * r;
- int lenr, lenp;
-
- if ((wl < 2) || !numrep) return 0;
-
- for (int i=0; i < numrep; i++ ) {
- r = word;
- lenr = strlen(reptable[i].pattern2);
- lenp = strlen(reptable[i].pattern);
- // search every occurence of the pattern in the word
- while ((r=strstr(r, reptable[i].pattern)) != NULL) {
- strcpy(candidate, word);
- if (r-word + lenr + strlen(r+lenp) >= MAXLNLEN) break;
- strcpy(candidate+(r-word),reptable[i].pattern2);
- strcpy(candidate+(r-word)+lenr, r+lenp);
- if (candidate_check(candidate,strlen(candidate))) return 1;
- r++; // search for the next letter
- }
- }
- return 0;
- }
-
- // forbid compoundings when there are special patterns at word bound
- int AffixMgr::cpdpat_check(const char * word, int pos)
- {
- int len;
- for (int i = 0; i < numcheckcpd; i++) {
- if (isSubset(checkcpdtable[i].pattern2, word + pos) &&
- (len = strlen(checkcpdtable[i].pattern)) && (pos > len) &&
- (strncmp(word + pos - len, checkcpdtable[i].pattern, len) == 0)) return 1;
- }
- return 0;
- }
-
- // forbid compounding with neighbouring upper and lower case characters at word bounds
- int AffixMgr::cpdcase_check(const char * word, int pos)
- {
- if (utf8) {
- w_char u, w;
- const char * p;
- u8_u16(&u, 1, word + pos);
- for (p = word + pos - 1; (*p & 0xc0) == 0x80; p--);
- u8_u16(&w, 1, p);
- unsigned short a = (u.h << 8) + u.l;
- unsigned short b = (w.h << 8) + w.l;
- if (((unicodetoupper(a, langnum) == a) || (unicodetoupper(b, langnum) == b))) return 1;
- } else {
- unsigned char a = *(word + pos - 1);
- unsigned char b = *(word + pos);
- if ((csconv[a].ccase || csconv[b].ccase) && (a != '-') && (b != '-')) return 1;
- }
- return 0;
- }
-
- // check compound patterns
- int AffixMgr::defcpd_check(hentry *** words, short wnum, hentry * rv, hentry ** def, char all)
- {
- signed short btpp[MAXWORDLEN]; // metacharacter (*, ?) positions for backtracking
- signed short btwp[MAXWORDLEN]; // word positions for metacharacters
- int btnum[MAXWORDLEN]; // number of matched characters in metacharacter positions
- short bt = 0;
- int i;
- int ok;
- int w = 0;
- if (!*words) {
- w = 1;
- *words = def;
- }
- (*words)[wnum] = rv;
-
- for (i = 0; i < numdefcpd; i++) {
- signed short pp = 0; // pattern position
- signed short wp = 0; // "words" position
- int ok2;
- ok = 1;
- ok2 = 1;
- do {
- while ((pp < defcpdtable[i].len) && (wp <= wnum)) {
- if (((pp+1) < defcpdtable[i].len) &&
- ((defcpdtable[i].def[pp+1] == '*') || (defcpdtable[i].def[pp+1] == '?'))) {
- int wend = (defcpdtable[i].def[pp+1] == '?') ? wp : wnum;
- ok2 = 1;
- pp+=2;
- btpp[bt] = pp;
- btwp[bt] = wp;
- while (wp <= wend) {
- if (!(*words)[wp]->alen ||
- !TESTAFF((*words)[wp]->astr, defcpdtable[i].def[pp-2], (*words)[wp]->alen)) {
- ok2 = 0;
- break;
- }
- wp++;
- }
- if (wp <= wnum) ok2 = 0;
- btnum[bt] = wp - btwp[bt];
- if (btnum[bt] > 0) bt++;
- if (ok2) break;
- } else {
- ok2 = 1;
- if (!(*words)[wp] || !(*words)[wp]->alen ||
- !TESTAFF((*words)[wp]->astr, defcpdtable[i].def[pp], (*words)[wp]->alen)) {
- ok = 0;
- break;
- }
- pp++;
- wp++;
- if ((defcpdtable[i].len == pp) && !(wp > wnum)) ok = 0;
- }
- }
- if (ok && ok2) {
- int r = pp;
- while ((defcpdtable[i].len > r) && ((r+1) < defcpdtable[i].len) &&
- ((defcpdtable[i].def[r+1] == '*') || (defcpdtable[i].def[r+1] == '?'))) r+=2;
- if (defcpdtable[i].len <= r) return 1;
- }
- // backtrack
- if (bt) do {
- ok = 1;
- btnum[bt - 1]--;
- pp = btpp[bt - 1];
- wp = btwp[bt - 1] + btnum[bt - 1];
- } while ((btnum[bt - 1] < 0) && --bt);
- } while (bt);
-
- if (ok && ok2 && (!all || (defcpdtable[i].len <= pp))) return 1;
- // check zero ending
- while (ok && ok2 && (defcpdtable[i].len > pp) && ((pp+1) < defcpdtable[i].len) &&
- ((defcpdtable[i].def[pp+1] == '*') || (defcpdtable[i].def[pp+1] == '?'))) pp+=2;
- if (ok && ok2 && (defcpdtable[i].len <= pp)) return 1;
- }
- (*words)[wnum] = NULL;
- if (w) *words = NULL;
- return 0;
- }
-
- inline int AffixMgr::candidate_check(const char * word, int len)
- {
- struct hentry * rv=NULL;
-
- rv = lookup(word);
- if (rv) return 1;
-
- // rv = prefix_check(word,len,1);
- // if (rv) return 1;
-
- rv = affix_check(word,len);
- if (rv) return 1;
- return 0;
- }
-
- // calculate number of syllable for compound-checking
- short AffixMgr::get_syllable(const char * word, int wlen)
- {
- if (cpdmaxsyllable==0) return 0;
-
- short num=0;
-
- if (!utf8) {
- for (int i=0; i<wlen; i++) {
- if (strchr(cpdvowels, word[i])) num++;
- }
- } else if (cpdvowels_utf16) {
- w_char w[MAXWORDUTF8LEN];
- int i = u8_u16(w, MAXWORDUTF8LEN, word);
- for (; i; i--) {
- if (flag_bsearch((unsigned short *) cpdvowels_utf16,
- ((unsigned short *) w)[i - 1], cpdvowels_utf16_len)) num++;
- }
- }
- return num;
- }
-
- // check if compound word is correctly spelled
- // hu_mov_rule = spec. Hungarian rule (XXX)
- struct hentry * AffixMgr::compound_check(const char * word, int len,
- short wordnum, short numsyllable, short maxwordnum, short wnum, hentry ** words = NULL,
- char hu_mov_rule = 0, int * cmpdstemnum = NULL, int * cmpdstem = NULL, char is_sug = 0)
- {
- int i;
- short oldnumsyllable, oldnumsyllable2, oldwordnum, oldwordnum2;
- int oldcmpdstemnum = 0;
- struct hentry * rv = NULL;
- struct hentry * rv_first;
- struct hentry * rwords[MAXWORDLEN]; // buffer for COMPOUND pattern checking
- char st [MAXWORDUTF8LEN + 4];
- char ch;
- int cmin;
- int cmax;
-
- int checked_prefix;
-
- #ifdef HUNSTEM
- if (cmpdstemnum) {
- if (wordnum == 0) {
- *cmpdstemnum = 1;
- } else {
- (*cmpdstemnum)++;
- }
- }
- #endif
- if (utf8) {
- for (cmin = 0, i = 0; (i < cpdmin) && word[cmin]; i++) {
- cmin++;
- for (; (word[cmin] & 0xc0) == 0x80; cmin++);
- }
- for (cmax = len, i = 0; (i < (cpdmin - 1)) && cmax; i++) {
- cmax--;
- for (; (word[cmax] & 0xc0) == 0x80; cmax--);
- }
- } else {
- cmin = cpdmin;
- cmax = len - cpdmin + 1;
- }
-
- strcpy(st, word);
-
- for (i = cmin; i < cmax; i++) {
-
- oldnumsyllable = numsyllable;
- oldwordnum = wordnum;
- checked_prefix = 0;
-
- // go to end of the UTF-8 character
- if (utf8) {
- for (; (st[i] & 0xc0) == 0x80; i++);
- if (i >= cmax) return NULL;
- }
-
-
- ch = st[i];
- st[i] = '\0';
-
- sfx = NULL;
- pfx = NULL;
-
- // FIRST WORD
-
- rv = lookup(st); // perhaps without prefix
-
- // search homonym with compound flag
- while ((rv) && !hu_mov_rule &&
- ((pseudoroot && TESTAFF(rv->astr, pseudoroot, rv->alen)) ||
- !((compoundflag && !words && TESTAFF(rv->astr, compoundflag, rv->alen)) ||
- (compoundbegin && !wordnum &&
- TESTAFF(rv->astr, compoundbegin, rv->alen)) ||
- (compoundmiddle && wordnum && !words &&
- TESTAFF(rv->astr, compoundmiddle, rv->alen)) ||
- (numdefcpd &&
- ((!words && !wordnum && defcpd_check(&words, wnum, rv, (hentry **) &rwords, 0)) ||
- (words && defcpd_check(&words, wnum, rv, (hentry **) &rwords, 0))))
- ))) {
- rv = rv->next_homonym;
- }
-
- if (!rv) {
- if (compoundflag &&
- !(rv = prefix_check(st, i, hu_mov_rule ? IN_CPD_OTHER : IN_CPD_BEGIN, compoundflag))) {
- if ((rv = suffix_check(st, i, 0, NULL, NULL, 0, NULL,
- FLAG_NULL, compoundflag, hu_mov_rule ? IN_CPD_OTHER : IN_CPD_BEGIN)) && !hu_mov_rule &&
- ((SfxEntry*)sfx)->getCont() &&
- ((compoundforbidflag && TESTAFF(((SfxEntry*)sfx)->getCont(), compoundforbidflag,
- ((SfxEntry*)sfx)->getContLen())) || (compoundend &&
- TESTAFF(((SfxEntry*)sfx)->getCont(), compoundend,
- ((SfxEntry*)sfx)->getContLen())))) {
- rv = NULL;
- }
- }
- if (rv ||
- (((wordnum == 0) && compoundbegin &&
- ((rv = suffix_check(st, i, 0, NULL, NULL, 0, NULL, FLAG_NULL, compoundbegin, hu_…
Large files files are truncated, but you can click here to view the full file