1 /* vi:set ts=8 sts=4 sw=4 noet: 2 * 3 * VIM - Vi IMproved by Bram Moolenaar 4 * 5 * Do ":help uganda" in Vim to read copying and usage conditions. 6 * Do ":help credits" in Vim to see a list of people who contributed. 7 * See README.txt for an overview of the Vim source code. 8 */ 9 10 /* 11 * spell.h: common code for spell checking, used by spell.c and spellfile.c. 12 */ 13 14 /* Use SPELL_PRINTTREE for debugging: dump the word tree after adding a word. 15 * Only use it for small word lists! */ 16 #if 0 17 # define SPELL_PRINTTREE 18 #endif 19 20 /* Use SPELL_COMPRESS_ALLWAYS for debugging: compress the word tree after 21 * adding a word. Only use it for small word lists! */ 22 #if 0 23 # define SPELL_COMPRESS_ALLWAYS 24 #endif 25 26 /* Use DEBUG_TRIEWALK to print the changes made in suggest_trie_walk() for a 27 * specific word. */ 28 #if 0 29 # define DEBUG_TRIEWALK 30 #endif 31 32 #define MAXWLEN 254 /* Assume max. word len is this many bytes. 33 Some places assume a word length fits in a 34 byte, thus it can't be above 255. 35 Must be >= PFD_NOTSPECIAL. */ 36 37 #define MAXREGIONS 8 /* Number of regions supported. */ 38 39 /* Type used for indexes in the word tree need to be at least 4 bytes. If int 40 * is 8 bytes we could use something smaller, but what? */ 41 typedef int idx_T; 42 43 typedef int salfirst_T; 44 45 /* 46 * Structure used to store words and other info for one language, loaded from 47 * a .spl file. 48 * The main access is through the tree in "sl_fbyts/sl_fidxs", storing the 49 * case-folded words. "sl_kbyts/sl_kidxs" is for keep-case words. 50 * 51 * The "byts" array stores the possible bytes in each tree node, preceded by 52 * the number of possible bytes, sorted on byte value: 53 * <len> <byte1> <byte2> ... 54 * The "idxs" array stores the index of the child node corresponding to the 55 * byte in "byts". 56 * Exception: when the byte is zero, the word may end here and "idxs" holds 57 * the flags, region mask and affixID for the word. There may be several 58 * zeros in sequence for alternative flag/region/affixID combinations. 59 */ 60 typedef struct slang_S slang_T; 61 struct slang_S 62 { 63 slang_T *sl_next; /* next language */ 64 char_u *sl_name; /* language name "en", "en.rare", "nl", etc. */ 65 char_u *sl_fname; /* name of .spl file */ 66 int sl_add; /* TRUE if it's a .add file. */ 67 68 char_u *sl_fbyts; /* case-folded word bytes */ 69 idx_T *sl_fidxs; /* case-folded word indexes */ 70 char_u *sl_kbyts; /* keep-case word bytes */ 71 idx_T *sl_kidxs; /* keep-case word indexes */ 72 char_u *sl_pbyts; /* prefix tree word bytes */ 73 idx_T *sl_pidxs; /* prefix tree word indexes */ 74 75 char_u *sl_info; /* infotext string or NULL */ 76 77 char_u sl_regions[MAXREGIONS * 2 + 1]; 78 /* table with up to 8 region names plus NUL */ 79 80 char_u *sl_midword; /* MIDWORD string or NULL */ 81 82 hashtab_T sl_wordcount; /* hashtable with word count, wordcount_T */ 83 84 int sl_compmax; /* COMPOUNDWORDMAX (default: MAXWLEN) */ 85 int sl_compminlen; /* COMPOUNDMIN (default: 0) */ 86 int sl_compsylmax; /* COMPOUNDSYLMAX (default: MAXWLEN) */ 87 int sl_compoptions; /* COMP_* flags */ 88 garray_T sl_comppat; /* CHECKCOMPOUNDPATTERN items */ 89 regprog_T *sl_compprog; /* COMPOUNDRULE turned into a regexp progrm 90 * (NULL when no compounding) */ 91 char_u *sl_comprules; /* all COMPOUNDRULE concatenated (or NULL) */ 92 char_u *sl_compstartflags; /* flags for first compound word */ 93 char_u *sl_compallflags; /* all flags for compound words */ 94 char_u sl_nobreak; /* When TRUE: no spaces between words */ 95 char_u *sl_syllable; /* SYLLABLE repeatable chars or NULL */ 96 garray_T sl_syl_items; /* syllable items */ 97 98 int sl_prefixcnt; /* number of items in "sl_prefprog" */ 99 regprog_T **sl_prefprog; /* table with regprogs for prefixes */ 100 101 garray_T sl_rep; /* list of fromto_T entries from REP lines */ 102 short sl_rep_first[256]; /* indexes where byte first appears, -1 if 103 there is none */ 104 garray_T sl_sal; /* list of salitem_T entries from SAL lines */ 105 salfirst_T sl_sal_first[256]; /* indexes where byte first appears, -1 if 106 there is none */ 107 int sl_followup; /* SAL followup */ 108 int sl_collapse; /* SAL collapse_result */ 109 int sl_rem_accents; /* SAL remove_accents */ 110 int sl_sofo; /* SOFOFROM and SOFOTO instead of SAL items: 111 * "sl_sal_first" maps chars, when has_mbyte 112 * "sl_sal" is a list of wide char lists. */ 113 garray_T sl_repsal; /* list of fromto_T entries from REPSAL lines */ 114 short sl_repsal_first[256]; /* sl_rep_first for REPSAL lines */ 115 int sl_nosplitsugs; /* don't suggest splitting a word */ 116 int sl_nocompoundsugs; /* don't suggest compounding */ 117 118 /* Info from the .sug file. Loaded on demand. */ 119 time_t sl_sugtime; /* timestamp for .sug file */ 120 char_u *sl_sbyts; /* soundfolded word bytes */ 121 idx_T *sl_sidxs; /* soundfolded word indexes */ 122 buf_T *sl_sugbuf; /* buffer with word number table */ 123 int sl_sugloaded; /* TRUE when .sug file was loaded or failed to 124 load */ 125 126 int sl_has_map; /* TRUE if there is a MAP line */ 127 hashtab_T sl_map_hash; /* MAP for multi-byte chars */ 128 int sl_map_array[256]; /* MAP for first 256 chars */ 129 hashtab_T sl_sounddone; /* table with soundfolded words that have 130 handled, see add_sound_suggest() */ 131 }; 132 133 #ifdef VMS 134 # define SPL_FNAME_TMPL "%s_%s.spl" 135 # define SPL_FNAME_ADD "_add." 136 # define SPL_FNAME_ASCII "_ascii." 137 #else 138 # define SPL_FNAME_TMPL "%s.%s.spl" 139 # define SPL_FNAME_ADD ".add." 140 # define SPL_FNAME_ASCII ".ascii." 141 #endif 142 143 /* Flags used for a word. Only the lowest byte can be used, the region byte 144 * comes above it. */ 145 #define WF_REGION 0x01 /* region byte follows */ 146 #define WF_ONECAP 0x02 /* word with one capital (or all capitals) */ 147 #define WF_ALLCAP 0x04 /* word must be all capitals */ 148 #define WF_RARE 0x08 /* rare word */ 149 #define WF_BANNED 0x10 /* bad word */ 150 #define WF_AFX 0x20 /* affix ID follows */ 151 #define WF_FIXCAP 0x40 /* keep-case word, allcap not allowed */ 152 #define WF_KEEPCAP 0x80 /* keep-case word */ 153 154 /* for <flags2>, shifted up one byte to be used in wn_flags */ 155 #define WF_HAS_AFF 0x0100 /* word includes affix */ 156 #define WF_NEEDCOMP 0x0200 /* word only valid in compound */ 157 #define WF_NOSUGGEST 0x0400 /* word not to be suggested */ 158 #define WF_COMPROOT 0x0800 /* already compounded word, COMPOUNDROOT */ 159 #define WF_NOCOMPBEF 0x1000 /* no compounding before this word */ 160 #define WF_NOCOMPAFT 0x2000 /* no compounding after this word */ 161 162 /* flags for <pflags> */ 163 #define WFP_RARE 0x01 /* rare prefix */ 164 #define WFP_NC 0x02 /* prefix is not combining */ 165 #define WFP_UP 0x04 /* to-upper prefix */ 166 #define WFP_COMPPERMIT 0x08 /* prefix with COMPOUNDPERMITFLAG */ 167 #define WFP_COMPFORBID 0x10 /* prefix with COMPOUNDFORBIDFLAG */ 168 169 /* Flags for postponed prefixes in "sl_pidxs". Must be above affixID (one 170 * byte) and prefcondnr (two bytes). */ 171 #define WF_RAREPFX (WFP_RARE << 24) /* rare postponed prefix */ 172 #define WF_PFX_NC (WFP_NC << 24) /* non-combining postponed prefix */ 173 #define WF_PFX_UP (WFP_UP << 24) /* to-upper postponed prefix */ 174 #define WF_PFX_COMPPERMIT (WFP_COMPPERMIT << 24) /* postponed prefix with 175 * COMPOUNDPERMITFLAG */ 176 #define WF_PFX_COMPFORBID (WFP_COMPFORBID << 24) /* postponed prefix with 177 * COMPOUNDFORBIDFLAG */ 178 179 /* flags for <compoptions> */ 180 #define COMP_CHECKDUP 1 /* CHECKCOMPOUNDDUP */ 181 #define COMP_CHECKREP 2 /* CHECKCOMPOUNDREP */ 182 #define COMP_CHECKCASE 4 /* CHECKCOMPOUNDCASE */ 183 #define COMP_CHECKTRIPLE 8 /* CHECKCOMPOUNDTRIPLE */ 184 185 /* Info from "REP", "REPSAL" and "SAL" entries in ".aff" file used in si_rep, 186 * si_repsal, sl_rep, and si_sal. Not for sl_sal! 187 * One replacement: from "ft_from" to "ft_to". */ 188 typedef struct fromto_S 189 { 190 char_u *ft_from; 191 char_u *ft_to; 192 } fromto_T; 193 194 /* Info from "SAL" entries in ".aff" file used in sl_sal. 195 * The info is split for quick processing by spell_soundfold(). 196 * Note that "sm_oneof" and "sm_rules" point into sm_lead. */ 197 typedef struct salitem_S 198 { 199 char_u *sm_lead; /* leading letters */ 200 int sm_leadlen; /* length of "sm_lead" */ 201 char_u *sm_oneof; /* letters from () or NULL */ 202 char_u *sm_rules; /* rules like ^, $, priority */ 203 char_u *sm_to; /* replacement. */ 204 int *sm_lead_w; /* wide character copy of "sm_lead" */ 205 int *sm_oneof_w; /* wide character copy of "sm_oneof" */ 206 int *sm_to_w; /* wide character copy of "sm_to" */ 207 } salitem_T; 208 209 /* Values for SP_*ERROR are negative, positive values are used by 210 * read_cnt_string(). */ 211 #define SP_TRUNCERROR -1 /* spell file truncated error */ 212 #define SP_FORMERROR -2 /* format error in spell file */ 213 #define SP_OTHERERROR -3 /* other error while reading spell file */ 214 215 /* 216 * Structure used in "b_langp", filled from 'spelllang'. 217 */ 218 typedef struct langp_S 219 { 220 slang_T *lp_slang; /* info for this language */ 221 slang_T *lp_sallang; /* language used for sound folding or NULL */ 222 slang_T *lp_replang; /* language used for REP items or NULL */ 223 int lp_region; /* bitmask for region or REGION_ALL */ 224 } langp_T; 225 226 #define LANGP_ENTRY(ga, i) (((langp_T *)(ga).ga_data) + (i)) 227 228 #define VIMSUGMAGIC "VIMsug" /* string at start of Vim .sug file */ 229 #define VIMSUGMAGICL 6 230 #define VIMSUGVERSION 1 231 232 /* 233 * The tables used for recognizing word characters according to spelling. 234 * These are only used for the first 256 characters of 'encoding'. 235 */ 236 typedef struct spelltab_S 237 { 238 char_u st_isw[256]; /* flags: is word char */ 239 char_u st_isu[256]; /* flags: is uppercase char */ 240 char_u st_fold[256]; /* chars: folded case */ 241 char_u st_upper[256]; /* chars: upper case */ 242 } spelltab_T; 243 244 /* 245 * Use our own character-case definitions, because the current locale may 246 * differ from what the .spl file uses. 247 * These must not be called with negative number! 248 */ 249 #if defined(HAVE_WCHAR_H) 250 # include <wchar.h> /* for towupper() and towlower() */ 251 #endif 252 /* Multi-byte implementation. For Unicode we can call utf_*(), but don't do 253 * that for ASCII, because we don't want to use 'casemap' here. Otherwise use 254 * the "w" library function for characters above 255 if available. */ 255 #ifdef HAVE_TOWLOWER 256 # define SPELL_TOFOLD(c) (enc_utf8 && (c) >= 128 ? utf_fold(c) \ 257 : (c) < 256 ? (int)spelltab.st_fold[c] : (int)towlower(c)) 258 #else 259 # define SPELL_TOFOLD(c) (enc_utf8 && (c) >= 128 ? utf_fold(c) \ 260 : (c) < 256 ? (int)spelltab.st_fold[c] : (c)) 261 #endif 262 263 #ifdef HAVE_TOWUPPER 264 # define SPELL_TOUPPER(c) (enc_utf8 && (c) >= 128 ? utf_toupper(c) \ 265 : (c) < 256 ? (int)spelltab.st_upper[c] : (int)towupper(c)) 266 #else 267 # define SPELL_TOUPPER(c) (enc_utf8 && (c) >= 128 ? utf_toupper(c) \ 268 : (c) < 256 ? (int)spelltab.st_upper[c] : (c)) 269 #endif 270 271 #ifdef HAVE_ISWUPPER 272 # define SPELL_ISUPPER(c) (enc_utf8 && (c) >= 128 ? utf_isupper(c) \ 273 : (c) < 256 ? spelltab.st_isu[c] : iswupper(c)) 274 #else 275 # define SPELL_ISUPPER(c) (enc_utf8 && (c) >= 128 ? utf_isupper(c) \ 276 : (c) < 256 ? spelltab.st_isu[c] : (FALSE)) 277 #endif 278 279 #ifdef FEAT_SPELL 280 /* First language that is loaded, start of the linked list of loaded 281 * languages. */ 282 # ifdef IN_SPELL_C 283 # define SPELL_EXTERN 284 # define SPELL_INIT(x) x 285 # else 286 # define SPELL_EXTERN extern 287 # define SPELL_INIT(x) 288 # endif 289 290 SPELL_EXTERN slang_T *first_lang SPELL_INIT(= NULL); 291 292 /* file used for "zG" and "zW" */ 293 SPELL_EXTERN char_u *int_wordlist SPELL_INIT(= NULL); 294 295 296 SPELL_EXTERN char e_format[] SPELL_INIT(= N_("E759: Format error in spell file")); 297 298 SPELL_EXTERN spelltab_T spelltab; 299 SPELL_EXTERN int did_set_spelltab; 300 301 #endif 302