223
|
1 /* vi:set ts=8 sts=4 sw=4:
|
|
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.c: code for spell checking
|
226
|
12 *
|
243
|
13 * The basic spell checking mechanism is:
|
|
14 * 1. Isolate a word, up to the next non-word character.
|
|
15 * 2. Find the word in the hashtable of basic words.
|
|
16 * 3. If not found, look in the hashtable with "prewords". These are prefixes
|
|
17 * with a non-word character following a word character, e.g., "de-".
|
|
18 * 4. If still not found, for each matching a prefix try if the word matches
|
|
19 * without the prefix (and with the "chop" string added back).
|
|
20 * 5. If still still not found, for each matching suffix try if the word
|
|
21 * matches without the suffix (and with the "chop" string added back).
|
|
22 *
|
|
23 * Matching involves checking the caps type: Onecap ALLCAP KeepCap.
|
|
24 * After finding a matching word check for a leadstring (non-word characters
|
|
25 * before the word) and addstring (more text following, starting with a
|
|
26 * non-word character).
|
|
27 *
|
236
|
28 * Why doesn't Vim use aspell/ispell/myspell/etc.?
|
|
29 * See ":help develop-spell".
|
|
30 */
|
|
31
|
223
|
32 #if defined(MSDOS) || defined(WIN16) || defined(WIN32) || defined(_WIN64)
|
|
33 # include <io.h> /* for lseek(), must be before vim.h */
|
|
34 #endif
|
|
35
|
|
36 #include "vim.h"
|
|
37
|
|
38 #if defined(FEAT_SYN_HL) || defined(PROTO)
|
|
39
|
|
40 #ifdef HAVE_FCNTL_H
|
|
41 # include <fcntl.h>
|
|
42 #endif
|
|
43
|
226
|
44 #define MAXWLEN 100 /* assume max. word len is this many bytes */
|
|
45
|
223
|
46 /*
|
243
|
47 * Structure that is used to store the structures and strings from the
|
|
48 * language file. This avoids the need to allocate space for each individual
|
|
49 * word. It's allocated in big chunks for speed. It's freed all at once when
|
|
50 * 'encoding' changes.
|
236
|
51 */
|
|
52 #define SBLOCKSIZE 4096 /* default size of sb_data */
|
|
53 typedef struct sblock_S sblock_T;
|
|
54 struct sblock_S
|
|
55 {
|
|
56 sblock_T *sb_next; /* next block in list */
|
|
57 char_u sb_data[1]; /* data, actually longer */
|
|
58 };
|
|
59
|
243
|
60 /* Info from "REP" entries in ".aff" file used in af_rep.
|
|
61 * TODO: This is not used yet. Either use it or remove it. */
|
236
|
62 typedef struct repentry_S
|
|
63 {
|
|
64 char_u *re_from;
|
|
65 char_u *re_to;
|
|
66 } repentry_T;
|
|
67
|
|
68 /*
|
|
69 * Structure to store affix info.
|
|
70 */
|
|
71 typedef struct affitem_S affitem_T;
|
|
72 struct affitem_S
|
|
73 {
|
|
74 affitem_T *ai_next; /* next affix with same ai_add[] or NULL */
|
|
75 short_u ai_nr; /* affix number */
|
243
|
76 char_u ai_flags; /* AFF_ flags */
|
|
77 char_u ai_choplen; /* length of chop string in bytes */
|
236
|
78 char_u ai_addlen; /* length of ai_add in bytes */
|
243
|
79 char_u ai_leadlen; /* for AFF_PREWORD: length of lead string */
|
|
80 char_u ai_taillen; /* for AFF_PREWORD: length of tail string */
|
|
81 char_u ai_add[1]; /* Text added to basic word. This stores:
|
|
82 * 0: word for AFF_PREWORD or whole addition
|
|
83 * ai_addlen + 1: chop string
|
|
84 * + ai_choplen + 1: lead string for AFF_PREWORD
|
|
85 * + ai_leadlen + 1: trail string f. AFF_PREWORD
|
|
86 */
|
236
|
87 };
|
|
88
|
|
89 /* Get affitem_T pointer from hashitem that uses ai_add */
|
|
90 static affitem_T dumai;
|
|
91 #define HI2AI(hi) ((affitem_T *)((hi)->hi_key - (dumai.ai_add - (char_u *)&dumai)))
|
|
92
|
243
|
93 /* ai_flags: Affix item flags */
|
|
94 #define AFF_COMBINE 0x01 /* prefix combines with suffix */
|
|
95 #define AFF_PREWORD 0x02 /* prefix includes word */
|
|
96
|
236
|
97 /*
|
243
|
98 * Structure used to store words and other info for one language, loaded from
|
|
99 * a .spl file.
|
236
|
100 */
|
|
101 typedef struct slang_S slang_T;
|
|
102 struct slang_S
|
|
103 {
|
|
104 slang_T *sl_next; /* next language */
|
|
105 char_u *sl_name; /* language name "en", "en.rare", "nl", etc. */
|
|
106 hashtab_T sl_words; /* main word table, fword_T */
|
|
107 int sl_prefcnt; /* number of prefix NRs */
|
|
108 garray_T sl_preftab; /* list of hashtables to lookup prefixes */
|
|
109 affitem_T *sl_prefzero; /* list of prefixes with zero add length */
|
243
|
110 hashtab_T sl_prewords; /* prefixes that include a word */
|
236
|
111 int sl_suffcnt; /* number of suffix NRs */
|
|
112 garray_T sl_sufftab; /* list of hashtables to lookup suffixes */
|
|
113 affitem_T *sl_suffzero; /* list of suffixes with zero add length */
|
243
|
114 char_u *sl_try; /* "TRY" from .aff file TODO: not used */
|
|
115 garray_T sl_rep; /* list of repentry_T entries from REP lines
|
|
116 * TODO not used */
|
236
|
117 char_u sl_regions[17]; /* table with up to 8 region names plus NUL */
|
|
118 sblock_T *sl_block; /* list with allocated memory blocks */
|
|
119 int sl_error; /* error while loading */
|
|
120 };
|
|
121
|
243
|
122 /* First language that is loaded, start of the linked list of loaded
|
|
123 * languages. */
|
236
|
124 static slang_T *first_lang = NULL;
|
|
125
|
|
126 /*
|
|
127 * Structure to store an addition to a basic word.
|
243
|
128 * There are many of these, keep it small!
|
236
|
129 */
|
|
130 typedef struct addword_S addword_T;
|
|
131 struct addword_S
|
|
132 {
|
|
133 addword_T *aw_next; /* next addition */
|
|
134 char_u aw_flags; /* ADD_ flags */
|
|
135 char_u aw_region; /* region for word with this addition */
|
243
|
136 char_u aw_leadlen; /* byte length of lead in aw_word */
|
|
137 char_u aw_wordlen; /* byte length of first word in aw_word */
|
|
138 char_u aw_saveb; /* saved byte where aw_word[] is truncated at
|
|
139 end of hashtable key; NUL when not using
|
|
140 hashtable */
|
236
|
141 char_u aw_word[1]; /* text, actually longer: case-folded addition
|
|
142 plus, with ADD_KEEPCAP: keep-case addition */
|
|
143 };
|
|
144
|
243
|
145 /* Get addword_T pointer from hashitem that uses aw_word */
|
|
146 static addword_T dumaw;
|
|
147 #define HI2ADDWORD(hi) ((addword_T *)((hi)->hi_key - (dumaw.aw_word - (char_u *)&dumaw)))
|
|
148
|
236
|
149 /*
|
|
150 * Structure to store a basic word.
|
243
|
151 * There are many of these, keep it small!
|
236
|
152 */
|
|
153 typedef struct fword_S fword_T;
|
|
154 struct fword_S
|
|
155 {
|
243
|
156 fword_T *fw_next; /* same basic word with different caps and/or
|
|
157 * affixes */
|
|
158 addword_T *fw_adds; /* first addword_T entry */
|
236
|
159 short_u fw_flags; /* BWF_ flags */
|
243
|
160 char_u fw_region; /* region bits */
|
|
161 char_u fw_prefixcnt; /* number of prefix NRs */
|
|
162 char_u fw_suffixcnt; /* number of suffix NRs */
|
|
163 char_u fw_word[1]; /* actually longer:
|
|
164 * 0: case folded word or keep-case word when
|
|
165 * (flags & BWF_KEEPCAP)
|
|
166 * + word length + 1: list of prefix NRs
|
|
167 * + fw_prefixcnt [* 2]: list of suffix NRs
|
|
168 */
|
236
|
169 };
|
|
170
|
|
171 /* Get fword_T pointer from hashitem that uses fw_word */
|
|
172 static fword_T dumfw;
|
|
173 #define HI2FWORD(hi) ((fword_T *)((hi)->hi_key - (dumfw.fw_word - (char_u *)&dumfw)))
|
|
174
|
|
175 #define REGION_ALL 0xff
|
|
176
|
|
177
|
|
178 /*
|
|
179 * Structure used in "b_langp", filled from 'spelllang'.
|
|
180 */
|
|
181 typedef struct langp_S
|
|
182 {
|
|
183 slang_T *lp_slang; /* info for this language (NULL for last one) */
|
|
184 int lp_region; /* bitmask for region or REGION_ALL */
|
|
185 } langp_T;
|
|
186
|
|
187 #define LANGP_ENTRY(ga, i) (((langp_T *)(ga).ga_data) + (i))
|
|
188
|
|
189 #define SP_OK 0
|
|
190 #define SP_BAD 1
|
|
191 #define SP_RARE 2
|
|
192 #define SP_LOCAL 3
|
|
193
|
|
194 /* flags used for basic words in the spell file */
|
|
195 #define BWF_VALID 0x01 /* word is valid without additions */
|
|
196 #define BWF_REGION 0x02 /* region byte follows */
|
|
197 #define BWF_ONECAP 0x04 /* first letter must be capital */
|
|
198 #define BWF_SUFFIX 0x08 /* has suffix NR list */
|
|
199 #define BWF_SECOND 0x10 /* second flags byte follows */
|
|
200
|
|
201 #define BWF_ADDS 0x0100 /* there are additions */
|
|
202 #define BWF_PREFIX 0x0200 /* has prefix NR list */
|
|
203 #define BWF_ALLCAP 0x0400 /* all letters must be capital (not used
|
|
204 for single-letter words) */
|
|
205 #define BWF_KEEPCAP 0x0800 /* Keep case as-is */
|
|
206
|
243
|
207 #define BWF_ADDHASH 0x8000 /* Internal: use hashtab for additions */
|
|
208
|
|
209 #define NOWC_KEY (char_u *)"x" /* hashtab key used for additions without
|
|
210 any word character */
|
|
211
|
236
|
212 /* flags used for addition in the spell file */
|
|
213 #define ADD_REGION 0x02 /* region byte follows */
|
|
214 #define ADD_ONECAP 0x04 /* first letter must be capital */
|
|
215 #define ADD_ALLCAP 0x40 /* all letters must be capital (not used
|
|
216 for single-letter words) */
|
|
217 #define ADD_KEEPCAP 0x80 /* fixed case */
|
|
218
|
|
219 /* Translate ADD_ flags to BWF_ flags.
|
|
220 * (Needed to keep ADD_ flags in one byte.) */
|
|
221 #define ADD2BWF(x) (((x) & 0x0f) | (((x) & 0xf0) << 4))
|
|
222
|
243
|
223 #define VIMSPELLMAGIC "VIMspell02" /* string at start of Vim spell file */
|
236
|
224 #define VIMSPELLMAGICL 10
|
|
225
|
|
226 /*
|
|
227 * Structure to store info for word matching.
|
|
228 */
|
|
229 typedef struct matchinf_S
|
|
230 {
|
|
231 langp_T *mi_lp; /* info for language and region */
|
|
232 slang_T *mi_slang; /* info for the language */
|
243
|
233
|
|
234 /* pointers to original text to be checked */
|
236
|
235 char_u *mi_line; /* start of line containing word */
|
|
236 char_u *mi_word; /* start of word being checked */
|
|
237 char_u *mi_end; /* first non-word char after mi_word */
|
243
|
238 char_u *mi_wend; /* end of matching word (is mi_end
|
236
|
239 * or further) */
|
243
|
240 char_u *mi_fend; /* next char to be added to mi_fword */
|
|
241
|
|
242 /* case-folded text */
|
|
243 char_u mi_fword[MAXWLEN + 1]; /* mi_word case-folded */
|
|
244 int mi_fendlen; /* byte length of first word in
|
|
245 mi_fword */
|
|
246 int mi_faddlen; /* byte length of text in mi_fword
|
|
247 after first word */
|
|
248 char_u *mi_cword; /* word to check, points in mi_fword */
|
|
249 char_u *mi_awend; /* after next word, to check for
|
|
250 addition (NULL when not done yet) */
|
|
251 int mi_did_awend; /* did compute mi_awend */
|
|
252
|
|
253 /* others */
|
236
|
254 int mi_result; /* result so far: SP_BAD, SP_OK, etc. */
|
|
255 int mi_capflags; /* BWF_ONECAP BWF_ALLCAP BWF_KEEPCAP */
|
|
256 } matchinf_T;
|
|
257
|
|
258 static int word_match __ARGS((matchinf_T *mip));
|
|
259 static int check_adds __ARGS((matchinf_T *mip, fword_T *fw, int req_pref, int req_suf));
|
243
|
260 static void fill_awend __ARGS((matchinf_T *mip));
|
|
261 static void fold_addchars __ARGS((matchinf_T *mip, int addlen));
|
|
262 static int supports_affix __ARGS((int cnt, char_u *afflist, int afflistlen, int nr));
|
236
|
263 static int prefix_match __ARGS((matchinf_T *mip));
|
243
|
264 static int noprefix_match __ARGS((matchinf_T *mip, char_u *pword, char_u *cstart, affitem_T *ai));
|
236
|
265 static int suffix_match __ARGS((matchinf_T *mip));
|
|
266 static int match_caps __ARGS((int flags, char_u *caseword, matchinf_T *mip, char_u *cword, char_u *end));
|
|
267 static slang_T *slang_alloc __ARGS((char_u *lang));
|
|
268 static void slang_free __ARGS((slang_T *lp));
|
|
269 static slang_T *spell_load_lang __ARGS((char_u *lang));
|
|
270 static void spell_load_file __ARGS((char_u *fname, void *cookie));
|
|
271 static void *getroom __ARGS((slang_T *lp, int *bl_used, int len));
|
|
272 static int find_region __ARGS((char_u *rp, char_u *region));
|
|
273 static int captype __ARGS((char_u *word, char_u *end));
|
|
274
|
|
275 /*
|
|
276 * Main spell-checking function.
|
|
277 * "ptr" points to the start of a word.
|
|
278 * "*attrp" is set to the attributes for a badly spelled word. For a non-word
|
|
279 * or when it's OK it remains unchanged.
|
|
280 * This must only be called when 'spelllang' is not empty.
|
|
281 * Returns the length of the word in bytes, also when it's OK, so that the
|
|
282 * caller can skip over the word.
|
|
283 */
|
|
284 int
|
|
285 spell_check(wp, line, ptr, attrp)
|
|
286 win_T *wp; /* current window */
|
|
287 char_u *line; /* start of line where "ptr" points into */
|
|
288 char_u *ptr;
|
|
289 int *attrp;
|
|
290 {
|
|
291 matchinf_T mi; /* Most things are put in "mi" so that it can
|
|
292 be passed to functions quickly. */
|
|
293
|
|
294 /* Find the end of the word. We already know that *ptr is a word char. */
|
|
295 mi.mi_word = ptr;
|
|
296 mi.mi_end = ptr;
|
|
297 do
|
|
298 {
|
|
299 mb_ptr_adv(mi.mi_end);
|
|
300 } while (*mi.mi_end != NUL && spell_iswordc(mi.mi_end));
|
|
301
|
|
302 /* A word starting with a number is always OK. */
|
|
303 if (*ptr >= '0' && *ptr <= '9')
|
|
304 return (int)(mi.mi_end - ptr);
|
|
305
|
243
|
306 /* Make case-folded copy of the word. */
|
236
|
307 (void)str_foldcase(ptr, mi.mi_end - ptr, mi.mi_fword, MAXWLEN + 1);
|
|
308 mi.mi_cword = mi.mi_fword;
|
243
|
309 mi.mi_fendlen = STRLEN(mi.mi_fword);
|
|
310 mi.mi_faddlen = 0;
|
|
311 mi.mi_fend = mi.mi_end;
|
|
312
|
|
313 /* Check the caps type of the word. */
|
|
314 mi.mi_capflags = captype(ptr, mi.mi_end);
|
|
315
|
|
316 /* The word is bad unless we recognize it. */
|
236
|
317 mi.mi_result = SP_BAD;
|
|
318 mi.mi_wend = mi.mi_end;
|
243
|
319
|
|
320 mi.mi_awend = NULL;
|
|
321 mi.mi_did_awend = FALSE;
|
236
|
322 mi.mi_line = line;
|
|
323
|
|
324 /*
|
|
325 * Loop over the languages specified in 'spelllang'.
|
|
326 * We check them all, because a matching word may have additions that are
|
|
327 * longer than an already found matching word.
|
|
328 */
|
|
329 for (mi.mi_lp = LANGP_ENTRY(wp->w_buffer->b_langp, 0);
|
|
330 mi.mi_lp->lp_slang != NULL; ++mi.mi_lp)
|
|
331 {
|
|
332 /*
|
|
333 * Check for a matching word.
|
|
334 * If not found or wrong region try removing prefixes (and then
|
|
335 * suffixes).
|
|
336 * If still not found or wrong region try removing suffixes.
|
|
337 */
|
|
338 mi.mi_slang = mi.mi_lp->lp_slang;
|
|
339 if (!word_match(&mi) || mi.mi_result != SP_OK)
|
|
340 if (!prefix_match(&mi) || mi.mi_result != SP_OK)
|
|
341 suffix_match(&mi);
|
|
342 }
|
|
343
|
|
344 if (mi.mi_result != SP_OK)
|
|
345 {
|
|
346 if (mi.mi_result == SP_BAD)
|
|
347 *attrp = highlight_attr[HLF_SPB];
|
|
348 else if (mi.mi_result == SP_RARE)
|
|
349 *attrp = highlight_attr[HLF_SPR];
|
|
350 else
|
|
351 *attrp = highlight_attr[HLF_SPL];
|
|
352 }
|
|
353
|
|
354 return (int)(mi.mi_wend - ptr);
|
|
355 }
|
|
356
|
|
357 /*
|
243
|
358 * Check if the word "mip->mi_word" matches.
|
|
359 * "mip->mi_fword" is the same word case-folded;
|
|
360 *
|
|
361 * This checks the word as a whole and for prefixes that include a word.
|
|
362 *
|
|
363 * Note that when called mi_fword only contains the word up to mip->mi_end,
|
|
364 * but when checking additions it gets longer.
|
236
|
365 */
|
|
366 static int
|
|
367 word_match(mip)
|
|
368 matchinf_T *mip;
|
|
369 {
|
243
|
370 hash_T fhash = hash_hash(mip->mi_fword);
|
236
|
371 hashitem_T *hi;
|
|
372 fword_T *fw;
|
|
373 int valid = FALSE;
|
243
|
374 char_u *p;
|
|
375 char_u pword[MAXWLEN + 1];
|
|
376 int charlen;
|
|
377 int capflags_save;
|
|
378 affitem_T *ai;
|
|
379 char_u *cstart;
|
|
380 int addlen;
|
|
381 int n;
|
|
382 char_u *save_end;
|
|
383 int cc;
|
|
384
|
|
385 hi = hash_lookup(&mip->mi_slang->sl_words, mip->mi_fword, fhash);
|
|
386 if (!HASHITEM_EMPTY(hi))
|
|
387 {
|
|
388 /*
|
|
389 * Find a basic word for which the case of "mi_word" is correct.
|
|
390 * If it is, check additions and use the longest one.
|
|
391 */
|
|
392 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
393 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
394 mip->mi_word, mip->mi_end))
|
|
395 valid |= check_adds(mip, fw, -1, -1);
|
|
396 }
|
236
|
397
|
|
398 /*
|
243
|
399 * Try finding a matching preword for "mip->mi_word". These are
|
|
400 * prefixes that have a non-word character after a word character:
|
|
401 * "d'", "de-", "'s-", "l'de-". But not "'s".
|
|
402 * Also need to do this when a matching word was already found, because we
|
|
403 * might find a longer match this way (French: "qu" and "qu'a-t-elle").
|
236
|
404 */
|
243
|
405 cc = mip->mi_fword[mip->mi_fendlen];
|
|
406 mip->mi_fword[mip->mi_fendlen] = NUL;
|
|
407 hi = hash_lookup(&mip->mi_slang->sl_prewords, mip->mi_fword, fhash);
|
|
408 mip->mi_fword[mip->mi_fendlen] = cc;
|
|
409 if (!HASHITEM_EMPTY(hi))
|
|
410 {
|
|
411 capflags_save = mip->mi_capflags;
|
|
412
|
|
413 /* Go through the list of matching prewords. */
|
|
414 for (ai = HI2AI(hi); ai != NULL; ai = ai->ai_next)
|
|
415 {
|
|
416 /* Check that the lead string matches before the word. */
|
|
417 p = ai->ai_add + ai->ai_addlen + ai->ai_choplen + 2;
|
|
418 if (ai->ai_leadlen > 0)
|
|
419 {
|
|
420 if (mip->mi_word - mip->mi_line < ai->ai_leadlen
|
|
421 || STRNCMP(mip->mi_word - ai->ai_leadlen, p,
|
|
422 ai->ai_leadlen) != 0)
|
|
423 continue;
|
|
424 p += ai->ai_leadlen + 1; /* advance "p" to tail */
|
|
425 }
|
|
426 else
|
|
427 ++p; /* advance "p" to tail */
|
|
428
|
|
429 /* Check that the tail string matches after the word. Need
|
|
430 * to fold case first. */
|
|
431 if (ai->ai_taillen > 0)
|
|
432 {
|
|
433 if (ai->ai_taillen >= mip->mi_faddlen)
|
|
434 {
|
|
435 fold_addchars(mip, ai->ai_taillen);
|
|
436 if (ai->ai_taillen > mip->mi_faddlen)
|
|
437 continue; /* not enough chars, can't match */
|
|
438 }
|
|
439 if (STRNCMP(mip->mi_fword + mip->mi_fendlen,
|
|
440 p, ai->ai_taillen) != 0)
|
|
441 continue;
|
|
442 }
|
|
443
|
|
444 /*
|
|
445 * This preword matches. Remove the preword and check that
|
|
446 * the resulting word exits.
|
|
447 */
|
|
448
|
|
449 /* Find the place in the original word where the tail ends,
|
|
450 * needed for case checks. */
|
|
451 #ifdef FEAT_MBYTE
|
|
452 charlen = mb_charlen(p);
|
|
453 #else
|
|
454 charlen = ai->ai_taillen;
|
|
455 #endif
|
|
456 cstart = mip->mi_end;
|
|
457 for (n = 0; n < charlen; ++n)
|
|
458 mb_ptr_adv(cstart);
|
|
459
|
|
460 /* The new word starts with the chop. Then add up to the next
|
|
461 * non-word char. */
|
|
462 mch_memmove(pword, ai->ai_add + ai->ai_addlen + 1,
|
|
463 ai->ai_choplen);
|
|
464 p = mip->mi_fword + mip->mi_fendlen + ai->ai_taillen;
|
|
465 addlen = ai->ai_taillen;
|
|
466 while (spell_iswordc(p))
|
|
467 {
|
|
468 ++charlen;
|
|
469 #ifdef FEAT_MBYTE
|
|
470 addlen += (*mb_ptr2len_check)(p);
|
|
471 #else
|
|
472 ++addlen;
|
|
473 #endif
|
|
474 mb_ptr_adv(p);
|
|
475 if (addlen >= mip->mi_faddlen)
|
|
476 {
|
|
477 /* Get more folded characters in mip->mi_fword. */
|
|
478 fold_addchars(mip, addlen);
|
|
479 if (addlen >= mip->mi_faddlen)
|
|
480 break; /* not enough chars, can't match */
|
|
481 }
|
|
482 }
|
|
483 mch_memmove(pword + ai->ai_choplen,
|
|
484 mip->mi_fword + mip->mi_fendlen + ai->ai_taillen,
|
|
485 addlen - ai->ai_taillen);
|
|
486 pword[ai->ai_choplen + addlen - ai->ai_taillen] = NUL;
|
|
487
|
|
488 /* Need to set mi_end to find additions. Also set mi_fendlen
|
|
489 * and mi_faddlen. */
|
|
490 save_end = mip->mi_end;
|
|
491 while (--charlen >= 0)
|
|
492 mb_ptr_adv(mip->mi_end);
|
|
493 mip->mi_fendlen += addlen;
|
|
494 mip->mi_faddlen -= addlen;
|
|
495
|
|
496 /* Find the word "pword", caseword "cstart". */
|
|
497 n = noprefix_match(mip, pword, cstart, ai);
|
|
498 mip->mi_end = save_end;
|
|
499 mip->mi_fendlen -= addlen;
|
|
500 mip->mi_faddlen += addlen;
|
|
501 if (n)
|
|
502 valid = TRUE;
|
|
503
|
|
504 /* If we found a valid word, we still need to try other
|
|
505 * suffixes, because it may have an addition that's longer. */
|
|
506 }
|
|
507
|
|
508 mip->mi_capflags = capflags_save;
|
|
509 }
|
236
|
510
|
|
511 return valid;
|
|
512 }
|
|
513
|
|
514 /*
|
|
515 * Check a matching basic word for additions.
|
|
516 * Return TRUE if we have a valid match.
|
|
517 */
|
|
518 static int
|
|
519 check_adds(mip, fw, req_pref, req_suf)
|
|
520 matchinf_T *mip;
|
|
521 fword_T *fw;
|
|
522 int req_pref; /* required prefix nr, -1 if none */
|
|
523 int req_suf; /* required suffix nr, -1 if none */
|
|
524 {
|
|
525 int valid = FALSE;
|
|
526 addword_T *aw;
|
243
|
527 addword_T *naw = NULL;
|
236
|
528 char_u *p;
|
|
529 int addlen;
|
243
|
530 int cc;
|
|
531 hashitem_T *hi;
|
|
532 char_u *cp = NULL;
|
|
533 int n;
|
|
534
|
|
535 /* Check if required prefixes and suffixes are supported. These are on
|
|
536 * the basic word, not on each addition. */
|
|
537 if (req_pref >= 0 || req_suf >= 0)
|
|
538 {
|
|
539 /* Prefix NRs are stored just after the word in fw_word. */
|
|
540 cp = fw->fw_word + STRLEN(fw->fw_word) + 1;
|
|
541 if (req_pref >= 0 && !supports_affix(mip->mi_slang->sl_prefcnt,
|
|
542 cp, fw->fw_prefixcnt, req_pref))
|
|
543 return FALSE;
|
|
544 if (req_suf >= 0)
|
|
545 {
|
|
546 /* Suffix NRs are stored just after the Prefix NRs. */
|
|
547 if (fw->fw_prefixcnt > 0)
|
|
548 {
|
|
549 if (mip->mi_slang->sl_prefcnt > 256)
|
|
550 cp += fw->fw_prefixcnt * 2;
|
|
551 else
|
|
552 cp += fw->fw_prefixcnt;
|
|
553 }
|
|
554 if (!supports_affix(mip->mi_slang->sl_suffcnt,
|
|
555 cp, fw->fw_suffixcnt, req_suf))
|
|
556 return FALSE;
|
|
557 }
|
|
558 }
|
|
559
|
|
560 /* A word may be valid without an addition. */
|
|
561 if (fw->fw_flags & BWF_VALID)
|
236
|
562 {
|
|
563 valid = TRUE;
|
|
564 if (mip->mi_result != SP_OK)
|
|
565 {
|
|
566 if ((fw->fw_region & mip->mi_lp->lp_region) == 0)
|
|
567 mip->mi_result = SP_LOCAL;
|
|
568 else
|
|
569 mip->mi_result = SP_OK;
|
|
570 }
|
243
|
571 /* Set word end, required when matching a word after a preword. */
|
|
572 if (mip->mi_wend < mip->mi_end)
|
|
573 mip->mi_wend = mip->mi_end;
|
236
|
574 }
|
|
575
|
|
576 /*
|
|
577 * Check additions, both before and after the word.
|
|
578 * This may make the word longer, thus we also need to check
|
|
579 * when we already found a matching word.
|
243
|
580 * When the BWF_ADDHASH flag is present then fw_adds points to a hashtable
|
|
581 * for quick lookup. Otherwise it points to the list of all possible
|
|
582 * additions.
|
236
|
583 */
|
243
|
584 if (fw->fw_flags & BWF_ADDHASH)
|
236
|
585 {
|
243
|
586 /* Locate the text up to the next end-of-word. */
|
|
587 if (!mip->mi_did_awend)
|
|
588 fill_awend(mip);
|
|
589 if (mip->mi_awend == NULL)
|
|
590 return valid; /* there is no next word */
|
|
591
|
|
592 cc = *mip->mi_awend;
|
|
593 *mip->mi_awend = NUL;
|
|
594 hi = hash_find((hashtab_T *)fw->fw_adds,
|
|
595 mip->mi_fword + mip->mi_fendlen);
|
|
596 *mip->mi_awend = cc;
|
|
597 if (HASHITEM_EMPTY(hi))
|
|
598 return valid; /* no matching addition */
|
|
599 aw = HI2ADDWORD(hi);
|
|
600
|
|
601 /* Also check additions without word characters. If they are there,
|
|
602 * skip the first dummy entry. */
|
|
603 hi = hash_find((hashtab_T *)fw->fw_adds, NOWC_KEY);
|
|
604 if (!HASHITEM_EMPTY(hi))
|
|
605 naw = HI2ADDWORD(hi)->aw_next;
|
|
606 }
|
|
607 else
|
|
608 aw = fw->fw_adds;
|
|
609
|
|
610 for ( ; ; aw = aw->aw_next)
|
|
611 {
|
|
612 if (aw == NULL)
|
|
613 {
|
|
614 /* At end of list: may also try additions without word chars. */
|
|
615 if (naw == NULL)
|
|
616 break;
|
|
617 aw = naw;
|
|
618 naw = NULL;
|
|
619 }
|
|
620
|
236
|
621 if (aw->aw_leadlen > 0)
|
|
622 {
|
|
623 /* There is a leader, verify that it matches. */
|
|
624 if (aw->aw_leadlen > mip->mi_word - mip->mi_line
|
|
625 || STRNCMP(mip->mi_word - aw->aw_leadlen,
|
|
626 aw->aw_word, aw->aw_leadlen) != 0)
|
|
627 continue;
|
|
628 if (mip->mi_word - aw->aw_leadlen > mip->mi_line)
|
|
629 {
|
|
630 /* There must not be a word character just before the
|
|
631 * leader. */
|
|
632 p = mip->mi_word - aw->aw_leadlen;
|
|
633 mb_ptr_back(mip->mi_line, p);
|
|
634 if (spell_iswordc(p))
|
|
635 continue;
|
|
636 }
|
|
637 /* Leader matches. Addition is rest of "aw_word". */
|
|
638 p = aw->aw_word + aw->aw_leadlen;
|
|
639 }
|
|
640 else
|
|
641 /* No leader, use whole of "aw_word" for addition. */
|
|
642 p = aw->aw_word;
|
|
643
|
|
644 addlen = aw->aw_wordlen - aw->aw_leadlen;
|
|
645 if (addlen > 0)
|
|
646 {
|
|
647 /* Check for matching addition and no word character after it.
|
|
648 * First make sure we have enough case-folded chars to compare
|
|
649 * with. */
|
243
|
650 if (addlen >= mip->mi_faddlen)
|
|
651 fold_addchars(mip, addlen);
|
|
652
|
|
653 /* Put back the saved char, if needed. */
|
|
654 if (aw->aw_saveb != NUL)
|
236
|
655 {
|
243
|
656 cp = p + STRLEN(p);
|
|
657 *cp = aw->aw_saveb;
|
236
|
658 }
|
243
|
659 n = STRNCMP(mip->mi_fword + mip->mi_fendlen, p, addlen);
|
|
660 if (aw->aw_saveb != NUL)
|
|
661 *cp = NUL;
|
|
662
|
|
663 if (n != 0 || (mip->mi_fword[mip->mi_fendlen + addlen] != NUL
|
|
664 && spell_iswordc(mip->mi_fword + mip->mi_fendlen + addlen)))
|
236
|
665 continue;
|
|
666
|
|
667 /* Compute the length in the original word, before case folding. */
|
|
668 #ifdef FEAT_MBYTE
|
|
669 if (has_mbyte)
|
|
670 {
|
|
671 int l;
|
|
672
|
|
673 p = mip->mi_end;
|
243
|
674 for (l = 0; l < addlen; l += (*mb_ptr2len_check)(mip->mi_fword
|
|
675 + mip->mi_fendlen + l))
|
236
|
676 mb_ptr_adv(p);
|
|
677 addlen = p - mip->mi_end;
|
|
678 }
|
|
679 #endif
|
|
680
|
|
681 /* Check case of the addition. */
|
|
682 if (!match_caps(ADD2BWF(aw->aw_flags),
|
|
683 aw->aw_word + aw->aw_wordlen + 1, mip,
|
|
684 mip->mi_end, mip->mi_end + addlen))
|
|
685 continue;
|
|
686 }
|
|
687
|
|
688 /* Match! Use the new length if it's longer. */
|
|
689 if (mip->mi_wend < mip->mi_end + addlen)
|
|
690 mip->mi_wend = mip->mi_end + addlen;
|
|
691
|
|
692 valid = TRUE;
|
|
693 if (mip->mi_result != SP_OK)
|
|
694 {
|
|
695 if ((aw->aw_region & mip->mi_lp->lp_region) == 0)
|
|
696 mip->mi_result = SP_LOCAL;
|
|
697 else
|
|
698 mip->mi_result = SP_OK;
|
|
699 }
|
|
700 }
|
|
701
|
|
702 return valid;
|
|
703 }
|
|
704
|
|
705 /*
|
243
|
706 * Locate the text up to the next end-of-word after mip->mi_end.
|
|
707 */
|
|
708 static void
|
|
709 fill_awend(mip)
|
|
710 matchinf_T *mip;
|
|
711 {
|
|
712 char_u *p = mip->mi_end;
|
|
713 int addlen = 0;
|
|
714 int find_word = TRUE;
|
|
715
|
|
716 mip->mi_did_awend = TRUE;
|
|
717 if (mip->mi_faddlen == 0)
|
|
718 fold_addchars(mip, 0); /* need to fold first char */
|
|
719
|
|
720 /* 1: find_word == TRUE: skip over non-word characters after mi_end.
|
|
721 * 2: find_word == FALSE: skip over following word characters. */
|
|
722 for (p = mip->mi_fword + mip->mi_fendlen; *p != NUL; mb_ptr_adv(p))
|
|
723 {
|
|
724 if (spell_iswordc(p) == find_word)
|
|
725 {
|
|
726 if (!find_word)
|
|
727 break; /* done */
|
|
728 find_word = !find_word;
|
|
729 }
|
|
730 #ifdef FEAT_MBYTE
|
|
731 addlen += (*mb_ptr2len_check)(p);
|
|
732 #else
|
|
733 ++addlen;
|
|
734 #endif
|
|
735 if (addlen >= mip->mi_faddlen)
|
|
736 fold_addchars(mip, addlen); /* need to fold more chars */
|
|
737 }
|
|
738
|
|
739 /* If there are extra chars store the result. */
|
|
740 if (addlen != 0)
|
|
741 mip->mi_awend = p;
|
|
742 }
|
|
743
|
|
744 /*
|
|
745 * Fold enough characters of the checked text to be able to compare with an
|
|
746 * addition of length "addlen" plus one character (to be able to check the
|
|
747 * next character to be a non-word char).
|
|
748 * When there are not enough characters (end of line) mip->mi_faddlen will be
|
|
749 * smaller than "addlen".
|
|
750 */
|
|
751 static void
|
|
752 fold_addchars(mip, addlen)
|
|
753 matchinf_T *mip;
|
|
754 int addlen;
|
|
755 {
|
|
756 int l;
|
|
757 char_u *p = mip->mi_fword + mip->mi_fendlen;
|
|
758
|
|
759 while (mip->mi_faddlen <= addlen)
|
|
760 {
|
|
761 if (*mip->mi_fend == NUL) /* end of the line */
|
|
762 {
|
|
763 p[mip->mi_faddlen] = NUL;
|
|
764 break;
|
|
765 }
|
|
766 #ifdef FEAT_MBYTE
|
|
767 if (has_mbyte)
|
|
768 l = (*mb_ptr2len_check)(mip->mi_fend);
|
|
769 else
|
|
770 #endif
|
|
771 l = 1;
|
|
772 (void)str_foldcase(mip->mi_fend, l, p + mip->mi_faddlen,
|
|
773 MAXWLEN - mip->mi_fendlen - mip->mi_faddlen);
|
|
774 mip->mi_fend += l;
|
|
775 mip->mi_faddlen += STRLEN(p + mip->mi_faddlen);
|
|
776 }
|
|
777 }
|
|
778
|
|
779 /*
|
|
780 * Return TRUE if affix "nr" appears in affix list "afflist[afflistlen]".
|
236
|
781 */
|
|
782 static int
|
243
|
783 supports_affix(cnt, afflist, afflistlen, nr)
|
|
784 int cnt; /* total affix NR count */
|
|
785 char_u *afflist;
|
|
786 int afflistlen; /* affix count in "afflist" */
|
236
|
787 int nr;
|
|
788 {
|
243
|
789 char_u *pc = afflist;
|
236
|
790 int i;
|
243
|
791 int nr_msb, nr_lsb;
|
236
|
792
|
|
793 if (cnt <= 256)
|
|
794 {
|
243
|
795 /* one byte affix numbers */
|
|
796 for (i = afflistlen; --i >= 0; )
|
236
|
797 if (*pc++ == nr)
|
|
798 return TRUE;
|
|
799 }
|
|
800 else
|
|
801 {
|
243
|
802 /* two byte affix numbers, MSB first */
|
|
803 nr_msb = (unsigned)nr >> 8;
|
|
804 nr_lsb = nr & 0xff;
|
|
805 for (i = afflistlen; --i >= 0; )
|
|
806 {
|
|
807 if (*pc++ == nr_msb && *pc == nr_lsb)
|
236
|
808 return TRUE;
|
243
|
809 ++pc;
|
|
810 }
|
236
|
811 }
|
|
812 return FALSE;
|
|
813 }
|
|
814
|
|
815 /*
|
|
816 * Try finding a match for "mip->mi_cword" by removing prefixes.
|
|
817 */
|
|
818 static int
|
|
819 prefix_match(mip)
|
|
820 matchinf_T *mip;
|
|
821 {
|
|
822 int len = 0;
|
|
823 int charlen = 0;
|
|
824 int cc;
|
|
825 affitem_T *ai;
|
|
826 char_u pword[MAXWLEN + 1];
|
|
827 hashtab_T *ht;
|
|
828 hashitem_T *hi;
|
|
829 int found_valid = FALSE;
|
|
830 int cstart_charlen = 0;
|
|
831 char_u *cstart = mip->mi_word;
|
|
832 int capflags_save = mip->mi_capflags;
|
|
833
|
|
834 /*
|
|
835 * Check for prefixes with different character lengths.
|
|
836 * Start with zero length (only chop off).
|
|
837 */
|
|
838 for (charlen = 0; charlen <= mip->mi_slang->sl_preftab.ga_len; ++charlen)
|
|
839 {
|
|
840 if (charlen > 0)
|
|
841 {
|
|
842 #ifdef FEAT_MBYTE
|
|
843 if (has_mbyte)
|
243
|
844 len += (*mb_ptr2len_check)(mip->mi_cword + len);
|
236
|
845 else
|
|
846 #endif
|
|
847 len += 1;
|
|
848 }
|
|
849 if (mip->mi_cword[len] == NUL) /* end of word, no prefix possible */
|
|
850 break;
|
|
851
|
|
852 if (charlen == 0)
|
|
853 ai = mip->mi_slang->sl_prefzero;
|
|
854 else
|
|
855 {
|
|
856 /* Get pointer to hashtab for prefix of this many chars. */
|
|
857 ht = ((hashtab_T *)mip->mi_slang->sl_preftab.ga_data) + charlen - 1;
|
|
858 if (ht->ht_used == 0)
|
|
859 continue;
|
|
860
|
|
861 cc = mip->mi_cword[len];
|
|
862 mip->mi_cword[len] = NUL;
|
|
863 hi = hash_find(ht, mip->mi_cword);
|
|
864 mip->mi_cword[len] = cc;
|
|
865
|
|
866 if (HASHITEM_EMPTY(hi))
|
|
867 ai = NULL;
|
|
868 else
|
|
869 ai = HI2AI(hi);
|
|
870 }
|
|
871
|
|
872 /* Loop over all matching prefixes. */
|
|
873 for ( ; ai != NULL; ai = ai->ai_next)
|
|
874 {
|
243
|
875 /* Create the basic word from the chop string and the word after
|
|
876 * the matching add string. */
|
|
877 mch_memmove(pword, ai->ai_add + ai->ai_addlen + 1, ai->ai_choplen);
|
|
878 mch_memmove(pword + ai->ai_choplen, mip->mi_cword + ai->ai_addlen,
|
|
879 mip->mi_fendlen - ai->ai_addlen);
|
|
880 pword[mip->mi_fendlen - ai->ai_addlen] = NUL;
|
236
|
881
|
|
882 /* Adjust the word start for case checks, we only check the
|
|
883 * part after the prefix. */
|
|
884 while (cstart_charlen < charlen)
|
|
885 {
|
|
886 mb_ptr_adv(cstart);
|
|
887 ++cstart_charlen;
|
|
888 }
|
|
889
|
243
|
890 /* Find the word "pword", caseword "cstart". */
|
|
891 found_valid |= noprefix_match(mip, pword, cstart, ai);
|
|
892
|
|
893 if (found_valid && mip->mi_result == SP_OK)
|
236
|
894 {
|
243
|
895 /* Found a valid word, no need to try other suffixes. */
|
|
896 mip->mi_capflags = capflags_save;
|
|
897 return TRUE;
|
236
|
898 }
|
|
899 }
|
|
900 }
|
|
901
|
|
902 mip->mi_capflags = capflags_save;
|
|
903 return FALSE;
|
|
904 }
|
|
905
|
|
906 /*
|
243
|
907 * Check for matching word after removing a prefix.
|
|
908 * Return TRUE if found.
|
|
909 */
|
|
910 static int
|
|
911 noprefix_match(mip, pword, cstart, ai)
|
|
912 matchinf_T *mip;
|
|
913 char_u *pword; /* case-folded word */
|
|
914 char_u *cstart; /* original word after removed prefix */
|
|
915 affitem_T *ai; /* the prefix item */
|
|
916 {
|
|
917 hashitem_T *hi;
|
|
918 fword_T *fw;
|
|
919 int found_valid = FALSE;
|
|
920 char_u *word;
|
|
921 int i;
|
|
922 int fendlen;
|
|
923
|
|
924 /* Removing the prefix may change the caps, e.g. for
|
|
925 * "deAlf" removing "de" makes it ONECAP. */
|
|
926 mip->mi_capflags = captype(cstart, mip->mi_end);
|
|
927
|
|
928 /* Find the basic word. */
|
|
929 hi = hash_find(&mip->mi_slang->sl_words, pword);
|
|
930 if (!HASHITEM_EMPTY(hi))
|
|
931 {
|
|
932 /* Check if the word supports this prefix. */
|
|
933 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
934 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
935 cstart, mip->mi_end))
|
|
936 found_valid |= check_adds(mip, fw, ai->ai_nr, -1);
|
|
937
|
|
938 if (found_valid && mip->mi_result == SP_OK)
|
|
939 /* Found a valid word, no need to try other suffixes. */
|
|
940 return TRUE;
|
|
941 }
|
|
942
|
|
943 /* No matching basic word without prefix. When combining is
|
|
944 * allowed try with suffixes. */
|
|
945 if (ai->ai_flags & AFF_COMBINE)
|
|
946 {
|
|
947 /* Pass the word with prefix removed to suffix_match(). */
|
|
948 mip->mi_cword = pword;
|
|
949 word = mip->mi_word;
|
|
950 mip->mi_word = cstart;
|
|
951 fendlen = mip->mi_fendlen;
|
|
952 mip->mi_fendlen = STRLEN(pword);
|
|
953 i = suffix_match(mip);
|
|
954 mip->mi_cword = mip->mi_fword;
|
|
955 mip->mi_word = word;
|
|
956 mip->mi_fendlen = fendlen;
|
|
957 if (i)
|
|
958 return TRUE;
|
|
959 }
|
|
960
|
|
961 return FALSE;
|
|
962 }
|
|
963
|
|
964 /*
|
236
|
965 * Try finding a match for "mip->mi_cword" by removing suffixes.
|
|
966 */
|
|
967 static int
|
|
968 suffix_match(mip)
|
|
969 matchinf_T *mip;
|
|
970 {
|
|
971 char_u *sufp;
|
243
|
972 char_u *endw = mip->mi_cword + mip->mi_fendlen;
|
|
973 int endw_c = *endw;
|
236
|
974 int charlen;
|
|
975 affitem_T *ai;
|
|
976 char_u pword[MAXWLEN + 1];
|
|
977 fword_T *fw;
|
|
978 hashtab_T *ht;
|
|
979 hashitem_T *hi;
|
|
980 int tlen;
|
|
981 int cend_charlen = 0;
|
|
982 char_u *cend = mip->mi_end;
|
|
983 int found_valid = FALSE;
|
|
984 int capflags_save = mip->mi_capflags;
|
|
985
|
|
986 /*
|
|
987 * Try suffixes of different length, starting with an empty suffix (chop
|
|
988 * only, thus adds something).
|
|
989 * Stop checking if there are no suffixes with so many characters.
|
|
990 */
|
243
|
991 sufp = endw;
|
236
|
992 for (charlen = 0; charlen <= mip->mi_slang->sl_sufftab.ga_len; ++charlen)
|
|
993 {
|
|
994 /* Move the pointer to the possible suffix back one character, unless
|
|
995 * doing the first round (empty suffix). */
|
|
996 if (charlen > 0)
|
|
997 {
|
|
998 mb_ptr_back(mip->mi_cword, sufp);
|
|
999 if (sufp <= mip->mi_cword) /* start of word, no suffix possible */
|
|
1000 break;
|
|
1001 }
|
|
1002
|
|
1003 if (charlen == 0)
|
|
1004 ai = mip->mi_slang->sl_suffzero;
|
|
1005 else
|
|
1006 {
|
|
1007 /* Get pointer to hashtab for suffix of this many chars. */
|
|
1008 ht = ((hashtab_T *)mip->mi_slang->sl_sufftab.ga_data) + charlen - 1;
|
|
1009 if (ht->ht_used == 0)
|
|
1010 continue;
|
|
1011
|
243
|
1012 *endw = NUL; /* truncate after possible suffix */
|
236
|
1013 hi = hash_find(ht, sufp);
|
|
1014 if (HASHITEM_EMPTY(hi))
|
|
1015 ai = NULL;
|
|
1016 else
|
|
1017 ai = HI2AI(hi);
|
243
|
1018 *endw = endw_c;
|
236
|
1019 }
|
|
1020
|
|
1021 if (ai != NULL)
|
|
1022 {
|
|
1023 /* Found a list of matching suffixes. Now check that there is one
|
|
1024 * we can use. */
|
|
1025 tlen = sufp - mip->mi_cword; /* length of word without suffix */
|
|
1026 mch_memmove(pword, mip->mi_cword, tlen);
|
|
1027
|
|
1028 for ( ; ai != NULL; ai = ai->ai_next)
|
|
1029 {
|
|
1030 /* Found a matching suffix. Create the basic word by removing
|
|
1031 * the suffix and adding the chop string. */
|
|
1032 if (ai->ai_choplen == 0)
|
|
1033 pword[tlen] = NUL;
|
|
1034 else
|
243
|
1035 mch_memmove(pword + tlen, ai->ai_add + ai->ai_addlen + 1,
|
|
1036 ai->ai_choplen + 1);
|
236
|
1037
|
|
1038 /* Find the basic word. */
|
|
1039 hi = hash_find(&mip->mi_slang->sl_words, pword);
|
|
1040 if (!HASHITEM_EMPTY(hi))
|
|
1041 {
|
|
1042 /* Adjust the end for case checks, we only check the part
|
|
1043 * before the suffix. */
|
|
1044 while (cend_charlen < charlen)
|
|
1045 {
|
|
1046 mb_ptr_back(mip->mi_word, cend);
|
|
1047 ++cend_charlen;
|
|
1048 }
|
|
1049
|
|
1050 /* Removing the suffix may change the caps, e.g. for
|
|
1051 * "UFOs" removing 's' makes it ALLCAP. */
|
|
1052 mip->mi_capflags = captype(mip->mi_word, cend);
|
|
1053
|
|
1054 /* Check if the word supports this suffix. */
|
|
1055 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
1056 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
1057 mip->mi_word, cend))
|
|
1058 found_valid |= check_adds(mip, fw, -1, ai->ai_nr);
|
|
1059
|
|
1060 if (found_valid && mip->mi_result == SP_OK)
|
|
1061 {
|
|
1062 /* Found a valid word, no need to try other suffixes. */
|
|
1063 mip->mi_capflags = capflags_save;
|
|
1064 return TRUE;
|
|
1065 }
|
|
1066 }
|
|
1067 }
|
|
1068 }
|
|
1069 }
|
|
1070
|
|
1071 mip->mi_capflags = capflags_save;
|
|
1072 return FALSE;
|
|
1073 }
|
|
1074
|
|
1075 /*
|
|
1076 * Return TRUE if case of "cword" meets the requirements of case flags
|
|
1077 * "flags".
|
|
1078 */
|
|
1079 static int
|
|
1080 match_caps(flags, caseword, mip, cword, end)
|
|
1081 int flags; /* flags required by basic word or addition */
|
|
1082 char_u *caseword; /* word with case as required */
|
|
1083 matchinf_T *mip;
|
|
1084 char_u *cword; /* word to compare against "caseword" */
|
|
1085 char_u *end; /* end of "cword" */
|
|
1086 {
|
|
1087 char_u *p;
|
|
1088 int c;
|
|
1089 int len;
|
|
1090 int capflags = mip->mi_capflags; /* flags of checked word */
|
|
1091 int past_second;
|
|
1092
|
|
1093 if ((capflags & BWF_KEEPCAP) == 0 && end > mip->mi_end)
|
|
1094 {
|
243
|
1095 /* If "end" is past "mip->mi_end" we need to adjust the caps type for
|
|
1096 * characters after the basic word. */
|
236
|
1097 #ifdef FEAT_MBYTE
|
|
1098 past_second = (mip->mi_word + (*mb_ptr2len_check)(mip->mi_word)
|
|
1099 < mip->mi_end);
|
|
1100 #else
|
|
1101 past_second = mip->mi_word + 1 < mip->mi_end;
|
|
1102 #endif
|
|
1103 for (p = mip->mi_end; p < end; )
|
|
1104 {
|
|
1105 if (!spell_iswordc(p))
|
|
1106 mb_ptr_adv(p);
|
|
1107 else
|
|
1108 {
|
|
1109 #ifdef FEAT_MBYTE
|
|
1110 if (has_mbyte)
|
|
1111 c = mb_ptr2char_adv(&p);
|
|
1112 else
|
|
1113 #endif
|
|
1114 c = *p++;
|
|
1115 if (MB_ISUPPER(c))
|
|
1116 {
|
|
1117 if (capflags == 0 || (capflags & BWF_ONECAP))
|
|
1118 {
|
|
1119 capflags = BWF_KEEPCAP; /* lU or UlU */
|
|
1120 break;
|
|
1121 }
|
|
1122 }
|
|
1123 else
|
|
1124 {
|
|
1125 if (capflags & BWF_ALLCAP)
|
|
1126 {
|
|
1127 if (past_second)
|
|
1128 {
|
|
1129 capflags = BWF_KEEPCAP; /* UUl */
|
|
1130 break;
|
|
1131 }
|
|
1132 capflags = BWF_ONECAP; /* Uu */
|
|
1133 }
|
|
1134 }
|
|
1135 past_second = TRUE;
|
|
1136 }
|
|
1137 }
|
|
1138 }
|
|
1139
|
|
1140 if (capflags == BWF_ALLCAP)
|
|
1141 return TRUE; /* All caps is always OK. */
|
|
1142
|
|
1143 if (flags & BWF_KEEPCAP)
|
|
1144 {
|
|
1145 len = STRLEN(caseword);
|
|
1146 return (len == end - cword && STRNCMP(caseword, cword, len) == 0);
|
|
1147 }
|
|
1148
|
|
1149 if (flags & BWF_ALLCAP)
|
|
1150 return FALSE; /* need ALLCAP, already checked above */
|
|
1151
|
|
1152 if (flags & BWF_ONECAP)
|
|
1153 return capflags == BWF_ONECAP;
|
|
1154
|
|
1155 return capflags != BWF_KEEPCAP; /* no case check, only KEEPCAP is bad */
|
|
1156 }
|
|
1157
|
|
1158 /*
|
|
1159 * Move to next spell error.
|
|
1160 * Return OK if found, FAIL otherwise.
|
|
1161 */
|
|
1162 int
|
|
1163 spell_move_to(dir, allwords)
|
|
1164 int dir; /* FORWARD or BACKWARD */
|
|
1165 int allwords; /* TRUE for "[s" and "]s" */
|
|
1166 {
|
|
1167 pos_T pos;
|
|
1168 char_u *line;
|
|
1169 char_u *p;
|
|
1170 int wc;
|
|
1171 int nwc;
|
|
1172 int attr = 0;
|
|
1173 int len;
|
|
1174
|
|
1175 if (!curwin->w_p_spell || *curwin->w_buffer->b_p_spl == NUL)
|
|
1176 {
|
|
1177 EMSG(_("E756: Spell checking not enabled"));
|
|
1178 return FAIL;
|
|
1179 }
|
|
1180
|
|
1181 /* TODO: moving backwards */
|
|
1182
|
|
1183 /* Start looking for bad word at the start of the line, because we can't
|
|
1184 * start halfway a word and know where it ends. */
|
|
1185 pos = curwin->w_cursor;
|
|
1186 pos.col = 0;
|
|
1187 wc = FALSE;
|
|
1188
|
|
1189 while (!got_int)
|
|
1190 {
|
|
1191 line = ml_get(pos.lnum);
|
|
1192 p = line + pos.col;
|
|
1193 while (*p != NUL)
|
|
1194 {
|
|
1195 nwc = spell_iswordc(p);
|
|
1196 if (!wc && nwc)
|
|
1197 {
|
|
1198 /* start of word */
|
|
1199 /* TODO: check for bad word attr */
|
|
1200 len = spell_check(curwin, line, p, &attr);
|
|
1201 if (attr != 0)
|
|
1202 {
|
|
1203 if (curwin->w_cursor.lnum < pos.lnum
|
|
1204 || (curwin->w_cursor.lnum == pos.lnum
|
|
1205 && curwin->w_cursor.col < (colnr_T)(p - line)))
|
|
1206 {
|
|
1207 curwin->w_cursor.lnum = pos.lnum;
|
|
1208 curwin->w_cursor.col = p - line;
|
|
1209 return OK;
|
|
1210 }
|
|
1211 attr = 0; /* bad word is before or at cursor */
|
|
1212 }
|
|
1213 p += len;
|
|
1214 if (*p == NUL)
|
|
1215 break;
|
|
1216 nwc = FALSE;
|
|
1217 }
|
|
1218
|
|
1219 /* advance to next character */
|
|
1220 mb_ptr_adv(p);
|
|
1221 wc = nwc;
|
|
1222 }
|
|
1223
|
|
1224 /* Advance to next line. */
|
|
1225 if (pos.lnum == curbuf->b_ml.ml_line_count)
|
|
1226 return FAIL;
|
|
1227 ++pos.lnum;
|
|
1228 pos.col = 0;
|
|
1229 wc = FALSE;
|
|
1230
|
|
1231 line_breakcheck();
|
|
1232 }
|
|
1233
|
|
1234 return FAIL; /* interrupted */
|
|
1235 }
|
|
1236
|
|
1237 /*
|
|
1238 * Load word list for "lang" from a Vim spell file.
|
|
1239 * "lang" must be the language without the region: "en" or "en-rare".
|
|
1240 */
|
|
1241 static slang_T *
|
|
1242 spell_load_lang(lang)
|
|
1243 char_u *lang;
|
|
1244 {
|
|
1245 slang_T *lp;
|
|
1246 char_u fname_enc[80];
|
|
1247 char_u *p;
|
|
1248 int r;
|
|
1249
|
|
1250 lp = slang_alloc(lang);
|
|
1251 if (lp != NULL)
|
|
1252 {
|
|
1253 /* Find all spell files for "lang" in 'runtimepath' and load them.
|
|
1254 * Use 'encoding', except that we use "latin1" for "latin9". */
|
|
1255 #ifdef FEAT_MBYTE
|
|
1256 if (STRLEN(p_enc) < 60 && STRCMP(p_enc, "iso-8859-15") != 0)
|
|
1257 p = p_enc;
|
|
1258 else
|
|
1259 #endif
|
|
1260 p = (char_u *)"latin1";
|
|
1261 sprintf((char *)fname_enc, "spell/%s.%s.spl", lang, p);
|
|
1262
|
|
1263 r = do_in_runtimepath(fname_enc, TRUE, spell_load_file, lp);
|
240
|
1264 if (r == FAIL && !lp->sl_error)
|
|
1265 {
|
|
1266 /* Try loading the ASCII version. */
|
|
1267 sprintf((char *)fname_enc, "spell/%s.ascii.spl", lang);
|
|
1268
|
|
1269 r = do_in_runtimepath(fname_enc, TRUE, spell_load_file, lp);
|
|
1270 }
|
236
|
1271 if (r == FAIL || lp->sl_error)
|
|
1272 {
|
|
1273 slang_free(lp);
|
|
1274 lp = NULL;
|
|
1275 if (r == FAIL)
|
|
1276 smsg((char_u *)_("Warning: Cannot find word list \"%s\""),
|
|
1277 fname_enc + 6);
|
|
1278 }
|
|
1279 else
|
|
1280 {
|
|
1281 lp->sl_next = first_lang;
|
|
1282 first_lang = lp;
|
|
1283 }
|
|
1284 }
|
|
1285
|
|
1286 return lp;
|
|
1287 }
|
|
1288
|
|
1289 /*
|
|
1290 * Allocate a new slang_T.
|
|
1291 * Caller must fill "sl_next".
|
|
1292 */
|
|
1293 static slang_T *
|
|
1294 slang_alloc(lang)
|
|
1295 char_u *lang;
|
|
1296 {
|
|
1297 slang_T *lp;
|
|
1298
|
|
1299 lp = (slang_T *)alloc(sizeof(slang_T));
|
|
1300 if (lp != NULL)
|
|
1301 {
|
|
1302 lp->sl_name = vim_strsave(lang);
|
|
1303 hash_init(&lp->sl_words);
|
|
1304 ga_init2(&lp->sl_preftab, sizeof(hashtab_T), 4);
|
243
|
1305 hash_init(&lp->sl_prewords);
|
236
|
1306 ga_init2(&lp->sl_sufftab, sizeof(hashtab_T), 4);
|
|
1307 lp->sl_prefzero = NULL;
|
|
1308 lp->sl_suffzero = NULL;
|
|
1309 lp->sl_try = NULL;
|
|
1310 ga_init2(&lp->sl_rep, sizeof(repentry_T), 4);
|
|
1311 lp->sl_regions[0] = NUL;
|
|
1312 lp->sl_block = NULL;
|
|
1313 lp->sl_error = FALSE;
|
|
1314 }
|
|
1315 return lp;
|
|
1316 }
|
|
1317
|
|
1318 /*
|
|
1319 * Free the contents of an slang_T and the structure itself.
|
|
1320 */
|
|
1321 static void
|
|
1322 slang_free(lp)
|
|
1323 slang_T *lp;
|
|
1324 {
|
|
1325 sblock_T *sp;
|
|
1326 int i;
|
243
|
1327 fword_T *fw;
|
|
1328 int todo;
|
|
1329 hashitem_T *hi;
|
236
|
1330
|
|
1331 vim_free(lp->sl_name);
|
243
|
1332
|
|
1333 /* The words themselves are in memory blocks referenced by "sl_block".
|
|
1334 * Only the hashtables for additions need to be cleared. */
|
|
1335 todo = lp->sl_words.ht_used;
|
|
1336 for (hi = lp->sl_words.ht_array; todo > 0; ++hi)
|
|
1337 {
|
|
1338 if (!HASHITEM_EMPTY(hi))
|
|
1339 {
|
|
1340 --todo;
|
|
1341 fw = HI2FWORD(hi);
|
|
1342 if (fw->fw_flags & BWF_ADDHASH)
|
|
1343 hash_clear((hashtab_T *)fw->fw_adds);
|
|
1344 }
|
|
1345 }
|
236
|
1346 hash_clear(&lp->sl_words);
|
243
|
1347
|
236
|
1348 for (i = 0; i < lp->sl_preftab.ga_len; ++i)
|
|
1349 hash_clear(((hashtab_T *)lp->sl_preftab.ga_data) + i);
|
|
1350 ga_clear(&lp->sl_preftab);
|
243
|
1351 hash_clear(&lp->sl_prewords);
|
236
|
1352 for (i = 0; i < lp->sl_sufftab.ga_len; ++i)
|
|
1353 hash_clear(((hashtab_T *)lp->sl_sufftab.ga_data) + i);
|
|
1354 ga_clear(&lp->sl_sufftab);
|
|
1355 ga_clear(&lp->sl_rep);
|
|
1356 vim_free(lp->sl_try);
|
|
1357 while (lp->sl_block != NULL)
|
|
1358 {
|
|
1359 sp = lp->sl_block;
|
|
1360 lp->sl_block = sp->sb_next;
|
|
1361 vim_free(sp);
|
|
1362 }
|
|
1363 vim_free(lp);
|
|
1364 }
|
|
1365
|
|
1366 /*
|
|
1367 * Load one spell file into an slang_T.
|
|
1368 * Invoked through do_in_runtimepath().
|
|
1369 */
|
|
1370 static void
|
|
1371 spell_load_file(fname, cookie)
|
|
1372 char_u *fname;
|
|
1373 void *cookie; /* points to the slang_T to be filled */
|
|
1374 {
|
|
1375 slang_T *lp = cookie;
|
|
1376 FILE *fd;
|
|
1377 char_u buf[MAXWLEN + 1];
|
|
1378 char_u cbuf[MAXWLEN + 1];
|
|
1379 char_u fbuf[MAXWLEN + 1];
|
243
|
1380 char_u affixbuf[256 * 2 * 2]; /* max 2 * 256 affix nrs of 2 bytes */
|
236
|
1381 char_u *p;
|
|
1382 int itm;
|
|
1383 int i;
|
|
1384 int affcount;
|
|
1385 int affnr;
|
|
1386 int affflags;
|
|
1387 int affitemcnt;
|
243
|
1388 int prefixcnt, suffixcnt;
|
236
|
1389 int bl_used = SBLOCKSIZE;
|
|
1390 int widx;
|
240
|
1391 int prefm = 0; /* 1 if <= 256 prefixes, sizeof(short_u) otherw. */
|
|
1392 int suffm = 0; /* 1 if <= 256 suffixes, sizeof(short_u) otherw. */
|
236
|
1393 int wlen;
|
|
1394 int flags;
|
|
1395 affitem_T *ai, *ai2, **aip;
|
|
1396 int round;
|
|
1397 char_u *save_sourcing_name = sourcing_name;
|
|
1398 linenr_T save_sourcing_lnum = sourcing_lnum;
|
|
1399 int cnt;
|
|
1400 int choplen;
|
|
1401 int addlen;
|
|
1402 int leadlen;
|
|
1403 int wordcount;
|
|
1404 fword_T *fw, *fw2;
|
|
1405 garray_T *gap;
|
|
1406 hashtab_T *ht;
|
|
1407 hashitem_T *hi;
|
|
1408 hash_T hash;
|
|
1409 int adds;
|
243
|
1410 addword_T *aw, *naw;
|
236
|
1411 int flen;
|
243
|
1412 int xlen;
|
236
|
1413
|
|
1414 fd = fopen((char *)fname, "r");
|
|
1415 if (fd == NULL)
|
|
1416 {
|
|
1417 EMSG2(_(e_notopen), fname);
|
|
1418 goto errorend;
|
|
1419 }
|
|
1420
|
|
1421 /* Set sourcing_name, so that error messages mention the file name. */
|
|
1422 sourcing_name = fname;
|
|
1423 sourcing_lnum = 0;
|
|
1424
|
|
1425 /* <HEADER>: <fileID> <regioncnt> <regionname> ... */
|
|
1426 for (i = 0; i < VIMSPELLMAGICL; ++i)
|
|
1427 buf[i] = getc(fd); /* <fileID> */
|
|
1428 if (STRNCMP(buf, VIMSPELLMAGIC, VIMSPELLMAGICL) != 0)
|
|
1429 {
|
|
1430 EMSG(_("E757: Wrong file ID in spell file"));
|
|
1431 goto errorend;
|
|
1432 }
|
|
1433
|
|
1434 cnt = getc(fd); /* <regioncnt> */
|
|
1435 if (cnt == EOF)
|
|
1436 {
|
|
1437 truncerr:
|
|
1438 EMSG(_("E758: Truncated spell file"));
|
|
1439 goto errorend;
|
|
1440 }
|
|
1441 if (cnt > 8)
|
|
1442 {
|
|
1443 formerr:
|
|
1444 EMSG(_("E759: Format error in spell file"));
|
|
1445 goto errorend;
|
|
1446 }
|
|
1447 for (i = 0; i < cnt; ++i)
|
|
1448 {
|
|
1449 lp->sl_regions[i * 2] = getc(fd); /* <regionname> */
|
|
1450 lp->sl_regions[i * 2 + 1] = getc(fd);
|
|
1451 }
|
|
1452 lp->sl_regions[cnt * 2] = NUL;
|
|
1453
|
|
1454 /* round 1: <PREFIXLIST>: <affcount> <afftotcnt> <affix> ...
|
|
1455 * round 2: <SUFFIXLIST>: <affcount> <afftotcnt> <affix> ... */
|
|
1456 for (round = 1; round <= 2; ++round)
|
|
1457 {
|
|
1458 affcount = (getc(fd) << 8) + getc(fd); /* <affcount> */
|
|
1459 if (affcount < 0)
|
|
1460 goto truncerr;
|
|
1461 if (round == 1)
|
|
1462 {
|
|
1463 gap = &lp->sl_preftab;
|
|
1464 aip = &lp->sl_prefzero;
|
|
1465 lp->sl_prefcnt = affcount;
|
243
|
1466 prefm = affcount > 256 ? 2 : 1;
|
236
|
1467 }
|
|
1468 else
|
|
1469 {
|
|
1470 gap = &lp->sl_sufftab;
|
|
1471 aip = &lp->sl_suffzero;
|
|
1472 lp->sl_suffcnt = affcount;
|
243
|
1473 suffm = affcount > 256 ? 2 : 1;
|
236
|
1474 }
|
|
1475
|
|
1476 i = (getc(fd) << 8) + getc(fd); /* <afftotcnt> */
|
|
1477 /* afftotcnt is not used */
|
|
1478
|
|
1479 /*
|
|
1480 * For each affix NR there can be several affixes.
|
|
1481 */
|
|
1482 for (affnr = 0; affnr < affcount; ++affnr)
|
|
1483 {
|
243
|
1484 /* <affix>: <affitemcnt> <affitem> ... */
|
236
|
1485 affitemcnt = (getc(fd) << 8) + getc(fd); /* <affitemcnt> */
|
|
1486 if (affitemcnt < 0)
|
|
1487 goto truncerr;
|
|
1488 for (itm = 0; itm < affitemcnt; ++itm)
|
|
1489 {
|
243
|
1490 /* <affitem>: <affflags> <affchoplen> <affchop>
|
|
1491 * <affaddlen> <affadd> */
|
|
1492 affflags = getc(fd); /* <affflags> */
|
236
|
1493 choplen = getc(fd); /* <affchoplen> */
|
|
1494 if (choplen == EOF)
|
|
1495 goto truncerr;
|
|
1496 if (choplen >= MAXWLEN)
|
|
1497 goto formerr;
|
|
1498 for (i = 0; i < choplen; ++i) /* <affchop> */
|
|
1499 buf[i] = getc(fd);
|
|
1500 buf[i] = NUL;
|
|
1501 addlen = getc(fd); /* <affaddlen> */
|
|
1502 if (addlen == EOF)
|
|
1503 goto truncerr;
|
243
|
1504 if (affflags & AFF_PREWORD)
|
|
1505 xlen = addlen + 2; /* space for lead and trail string */
|
|
1506 else
|
|
1507 xlen = 0;
|
|
1508
|
236
|
1509 /* Get room to store the affitem_T, chop and add strings. */
|
|
1510 p = (char_u *)getroom(lp, &bl_used,
|
243
|
1511 sizeof(affitem_T) + addlen + choplen + 1 + xlen);
|
236
|
1512 if (p == NULL)
|
|
1513 goto errorend;
|
|
1514
|
|
1515 ai = (affitem_T *)p;
|
|
1516 ai->ai_nr = affnr;
|
243
|
1517 ai->ai_flags = affflags;
|
236
|
1518 ai->ai_choplen = choplen;
|
|
1519 ai->ai_addlen = addlen;
|
|
1520
|
243
|
1521 /* Chop string is at ai_add[ai_addlen + 1]. */
|
|
1522 p = ai->ai_add + addlen + 1;
|
236
|
1523 STRCPY(p, buf);
|
|
1524
|
|
1525 p = ai->ai_add;
|
|
1526 for (i = 0; i < addlen; ++i) /* <affadd> */
|
|
1527 p[i] = getc(fd);
|
|
1528 p[i] = NUL;
|
|
1529
|
243
|
1530 if (affflags & AFF_PREWORD)
|
236
|
1531 {
|
243
|
1532 int l, leadoff, trailoff;
|
|
1533
|
|
1534 /*
|
|
1535 * Separate lead and trail string, put word at ai_add, so
|
|
1536 * that it can be used as hashtable key.
|
|
1537 */
|
|
1538 /* lead string: up to first word char */
|
|
1539 while (*p != NUL && !spell_iswordc(p))
|
|
1540 mb_ptr_adv(p);
|
|
1541 ai->ai_leadlen = p - ai->ai_add;
|
|
1542 leadoff = addlen + choplen + 2;
|
|
1543 mch_memmove(ai->ai_add + leadoff, ai->ai_add,
|
|
1544 ai->ai_leadlen);
|
|
1545 ai->ai_add[leadoff + ai->ai_leadlen] = NUL;
|
|
1546
|
|
1547 /* trail string: after last word char */
|
|
1548 while (*p != NUL && spell_iswordc(p))
|
|
1549 mb_ptr_adv(p);
|
|
1550 trailoff = leadoff + ai->ai_leadlen + 1;
|
|
1551 STRCPY(ai->ai_add + trailoff, p);
|
|
1552 ai->ai_taillen = STRLEN(p);
|
|
1553
|
|
1554 /* word itself */
|
|
1555 l = (p - ai->ai_add) - ai->ai_leadlen;
|
|
1556 mch_memmove(ai->ai_add, ai->ai_add + ai->ai_leadlen, l);
|
|
1557 ai->ai_add[l] = NUL;
|
|
1558 hash = hash_hash(ai->ai_add);
|
|
1559 hi = hash_lookup(&lp->sl_prewords, ai->ai_add, hash);
|
236
|
1560 if (HASHITEM_EMPTY(hi))
|
|
1561 {
|
243
|
1562 /* First affix with this word, add to hashtable. */
|
|
1563 hash_add_item(&lp->sl_prewords, hi, ai->ai_add, hash);
|
236
|
1564 ai->ai_next = NULL;
|
|
1565 }
|
|
1566 else
|
|
1567 {
|
243
|
1568 /* There already is an affix with this word, link in
|
|
1569 * the list. */
|
236
|
1570 ai2 = HI2AI(hi);
|
|
1571 ai->ai_next = ai2->ai_next;
|
|
1572 ai2->ai_next = ai;
|
|
1573 }
|
|
1574 }
|
243
|
1575 else
|
|
1576 {
|
|
1577 /*
|
|
1578 * Add the affix to a hashtable. Which one depends on the
|
|
1579 * length of the added string in characters.
|
|
1580 */
|
|
1581 #ifdef FEAT_MBYTE
|
|
1582 /* Change "addlen" from length in bytes to length in
|
|
1583 * chars. */
|
|
1584 if (has_mbyte)
|
|
1585 addlen = mb_charlen(p);
|
|
1586 #endif
|
|
1587 if (addlen == 0)
|
|
1588 {
|
|
1589 /* Link in list of zero length affixes. */
|
|
1590 ai->ai_next = *aip;
|
|
1591 *aip = ai;
|
|
1592 }
|
|
1593 else
|
|
1594 {
|
|
1595 if (gap->ga_len < addlen)
|
|
1596 {
|
|
1597 /* Longer affix, need more hashtables. */
|
|
1598 if (ga_grow(gap, addlen - gap->ga_len) == FAIL)
|
|
1599 goto errorend;
|
|
1600
|
|
1601 /* Re-allocating ga_data means that an ht_array
|
|
1602 * pointing to ht_smallarray becomes invalid. We
|
|
1603 * can recognize this: ht_mask is at its init
|
|
1604 * value. */
|
|
1605 for (i = 0; i < gap->ga_len; ++i)
|
|
1606 {
|
|
1607 ht = ((hashtab_T *)gap->ga_data) + i;
|
|
1608 if (ht->ht_mask == HT_INIT_SIZE - 1)
|
|
1609 ht->ht_array = ht->ht_smallarray;
|
|
1610 }
|
|
1611
|
|
1612 /* Init the newly used hashtable(s). */
|
|
1613 while (gap->ga_len < addlen)
|
|
1614 {
|
|
1615 hash_init(((hashtab_T *)gap->ga_data)
|
|
1616 + gap->ga_len);
|
|
1617 ++gap->ga_len;
|
|
1618 }
|
|
1619 }
|
|
1620 ht = ((hashtab_T *)gap->ga_data) + addlen - 1;
|
|
1621 hash = hash_hash(p);
|
|
1622 hi = hash_lookup(ht, p, hash);
|
|
1623 if (HASHITEM_EMPTY(hi))
|
|
1624 {
|
|
1625 /* First affix with this "ai_add", add to
|
|
1626 * hashtable. */
|
|
1627 hash_add_item(ht, hi, p, hash);
|
|
1628 ai->ai_next = NULL;
|
|
1629 }
|
|
1630 else
|
|
1631 {
|
|
1632 /* There already is an affix with this "ai_add",
|
|
1633 * link in the list. */
|
|
1634 ai2 = HI2AI(hi);
|
|
1635 ai->ai_next = ai2->ai_next;
|
|
1636 ai2->ai_next = ai;
|
|
1637 }
|
|
1638 }
|
|
1639 }
|
236
|
1640 }
|
|
1641 }
|
|
1642 }
|
|
1643
|
|
1644 /* <SUGGEST> : <suggestlen> <more> ... */
|
|
1645 /* TODO, just skip this for now */
|
|
1646 i = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8) + getc(fd);
|
|
1647 while (i-- > 0)
|
|
1648 if (getc(fd) == EOF) /* <suggestlen> */
|
|
1649 goto truncerr;
|
|
1650
|
|
1651 /* <WORDLIST>: <wordcount> <worditem> ... */ /* <wordcount> */
|
|
1652 wordcount = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8)
|
|
1653 + getc(fd);
|
|
1654 if (wordcount < 0)
|
|
1655 goto truncerr;
|
|
1656
|
|
1657 /* Init hashtable for this number of words, so that it doesn't need to
|
|
1658 * reallocate the table halfway. */
|
|
1659 hash_lock_size(&lp->sl_words, wordcount);
|
|
1660
|
|
1661 for (widx = 0; ; ++widx)
|
|
1662 {
|
|
1663 /* <worditem>: <nr> <string> <flags> [<flags2>]
|
|
1664 * [<caselen> <caseword>]
|
|
1665 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
1666 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
1667 * [<region>]
|
|
1668 * [<addcnt> <add> ...]
|
|
1669 */
|
|
1670 /* Use <nr> bytes from the previous word. */
|
|
1671 wlen = getc(fd); /* <nr> */
|
|
1672 if (wlen == EOF)
|
|
1673 {
|
|
1674 if (widx >= wordcount) /* normal way to end the file */
|
|
1675 break;
|
|
1676 goto truncerr;
|
|
1677 }
|
|
1678
|
|
1679 /* Read further word bytes until one below 0x20, that must be the
|
|
1680 * flags. Keep this fast! */
|
|
1681 for (;;)
|
|
1682 {
|
|
1683 if ((buf[wlen] = getc(fd)) < 0x20) /* <string> */
|
|
1684 break;
|
|
1685 if (++wlen == MAXWLEN)
|
|
1686 goto formerr;
|
|
1687 }
|
|
1688 flags = buf[wlen]; /* <flags> */
|
|
1689 buf[wlen] = NUL;
|
|
1690
|
|
1691 /* Get more flags if they're there. */
|
|
1692 if (flags & BWF_SECOND)
|
|
1693 flags += getc(fd) << 8; /* <flags2> */
|
|
1694
|
|
1695 if (flags & BWF_KEEPCAP)
|
|
1696 {
|
|
1697 /* Read <caselen> and <caseword> first, its length may differ from
|
|
1698 * the case-folded word. Note: this should only happen after the
|
|
1699 * basic word! */
|
|
1700 wlen = getc(fd);
|
243
|
1701 if (wlen < 0)
|
236
|
1702 goto truncerr;
|
|
1703 for (i = 0; i < wlen; ++i)
|
|
1704 cbuf[i] = getc(fd);
|
|
1705 cbuf[i] = NUL;
|
|
1706 }
|
|
1707
|
243
|
1708 /* Optional prefixes */
|
|
1709 p = affixbuf;
|
|
1710 if (flags & BWF_PREFIX)
|
|
1711 {
|
|
1712 cnt = getc(fd); /* <affixcnt> */
|
|
1713 if (cnt < 0)
|
|
1714 goto truncerr;
|
|
1715 prefixcnt = cnt;
|
|
1716 for (i = cnt * prefm; --i >= 0; ) /* <affixNR> */
|
|
1717 *p++ = getc(fd);
|
|
1718 }
|
|
1719 else
|
|
1720 prefixcnt = 0;
|
|
1721
|
|
1722 /* Optional suffixes */
|
|
1723 if (flags & BWF_SUFFIX)
|
|
1724 {
|
|
1725 cnt = getc(fd); /* <affixcnt> */
|
|
1726 if (cnt < 0)
|
|
1727 goto truncerr;
|
|
1728 suffixcnt = cnt;
|
|
1729 for (i = cnt * suffm; --i >= 0; ) /* <affixNR> */
|
|
1730 *p++ = getc(fd);
|
|
1731 }
|
|
1732 else
|
|
1733 suffixcnt = 0;
|
|
1734
|
|
1735 /* Find room to store the word in an fword_T. */
|
|
1736 fw = (fword_T *)getroom(lp, &bl_used, (int)sizeof(fword_T) + wlen
|
|
1737 + (p - affixbuf));
|
236
|
1738 if (fw == NULL)
|
|
1739 goto errorend;
|
|
1740 mch_memmove(fw->fw_word, (flags & BWF_KEEPCAP) ? cbuf : buf, wlen + 1);
|
243
|
1741
|
|
1742 /* Put the affix NRs just after the word, if any. */
|
|
1743 if (p > affixbuf)
|
|
1744 mch_memmove(fw->fw_word + wlen + 1, affixbuf, p - affixbuf);
|
|
1745
|
236
|
1746 fw->fw_flags = flags;
|
243
|
1747 fw->fw_prefixcnt = prefixcnt;
|
|
1748 fw->fw_suffixcnt = suffixcnt;
|
236
|
1749
|
|
1750 hash = hash_hash(buf);
|
|
1751 hi = hash_lookup(&lp->sl_words, buf, hash);
|
|
1752 if (HASHITEM_EMPTY(hi))
|
|
1753 {
|
|
1754 if (hash_add_item(&lp->sl_words, hi, fw->fw_word, hash) == FAIL)
|
|
1755 goto errorend;
|
|
1756 fw->fw_next = NULL;
|
|
1757 }
|
|
1758 else
|
|
1759 {
|
|
1760 /* Already have this basic word in the hashtable, this one will
|
243
|
1761 * have different case flags and/or affixes. */
|
236
|
1762 fw2 = HI2FWORD(hi);
|
|
1763 fw->fw_next = fw2->fw_next;
|
|
1764 fw2->fw_next = fw;
|
|
1765 --widx; /* don't count this one */
|
|
1766 }
|
|
1767
|
|
1768 if (flags & BWF_REGION)
|
|
1769 fw->fw_region = getc(fd); /* <region> */
|
|
1770 else
|
|
1771 fw->fw_region = REGION_ALL;
|
|
1772
|
|
1773 fw->fw_adds = NULL;
|
|
1774 if (flags & BWF_ADDS)
|
|
1775 {
|
|
1776 adds = (getc(fd) << 8) + getc(fd); /* <addcnt> */
|
|
1777
|
243
|
1778 if (adds > 30)
|
|
1779 {
|
|
1780 /* Use a hashtable to loopup the part until the next word end.
|
|
1781 * This uses more memory and involves some overhead, thus only
|
|
1782 * do it when there are many additions (e.g., for French). */
|
|
1783 ht = (hashtab_T *)getroom(lp, &bl_used, sizeof(hashtab_T));
|
|
1784 if (ht == NULL)
|
|
1785 goto errorend;
|
|
1786 hash_init(ht);
|
|
1787 fw->fw_adds = (addword_T *)ht;
|
|
1788 fw->fw_flags |= BWF_ADDHASH;
|
|
1789
|
|
1790 /* Preset the size of the hashtable. It's never unlocked. */
|
|
1791 hash_lock_size(ht, adds + 1);
|
|
1792 }
|
|
1793 else
|
|
1794 ht = NULL;
|
|
1795
|
236
|
1796 while (--adds >= 0)
|
|
1797 {
|
|
1798 /* <add>: <addflags> <addlen> [<leadlen> <addstring>]
|
|
1799 * [<region>] */
|
|
1800 flags = getc(fd); /* <addflags> */
|
|
1801 addlen = getc(fd); /* <addlen> */
|
|
1802 if (addlen == EOF)
|
|
1803 goto truncerr;
|
|
1804 if (addlen >= MAXWLEN)
|
|
1805 goto formerr;
|
|
1806
|
|
1807 if (addlen > 0)
|
|
1808 {
|
|
1809 leadlen = getc(fd); /* <leadlen> */
|
|
1810 for (i = 0; i < addlen; ++i) /* <addstring> */
|
|
1811 cbuf[i] = getc(fd);
|
|
1812 cbuf[i] = NUL;
|
|
1813 }
|
|
1814 else
|
|
1815 leadlen = 0;
|
|
1816
|
|
1817 if (flags & ADD_KEEPCAP)
|
|
1818 {
|
|
1819 /* <addstring> is in original case, need to get
|
|
1820 * case-folded word too. */
|
|
1821 (void)str_foldcase(cbuf, addlen, fbuf, MAXWLEN);
|
|
1822 flen = addlen - leadlen + 1;
|
|
1823 addlen = STRLEN(fbuf);
|
|
1824 }
|
|
1825 else
|
|
1826 flen = 0;
|
|
1827
|
|
1828 aw = (addword_T *)getroom(lp, &bl_used,
|
|
1829 sizeof(addword_T) + addlen + flen);
|
|
1830 if (aw == NULL)
|
|
1831 goto errorend;
|
|
1832
|
|
1833 if (flags & ADD_KEEPCAP)
|
|
1834 {
|
|
1835 /* Put the addition in original case after the case-folded
|
|
1836 * string. */
|
|
1837 STRCPY(aw->aw_word, fbuf);
|
|
1838 STRCPY(aw->aw_word + addlen + 1, cbuf + leadlen);
|
|
1839 }
|
|
1840 else
|
|
1841 STRCPY(aw->aw_word, cbuf);
|
|
1842
|
|
1843 aw->aw_flags = flags;
|
|
1844 aw->aw_wordlen = addlen;
|
243
|
1845 aw->aw_leadlen = leadlen;
|
236
|
1846
|
|
1847 if (flags & ADD_REGION)
|
|
1848 aw->aw_region = getc(fd); /* <region> */
|
|
1849 else
|
|
1850 aw->aw_region = REGION_ALL;
|
243
|
1851
|
|
1852 if (ht == NULL)
|
|
1853 {
|
|
1854 /* Using simple linked list, put it in front. */
|
|
1855 aw->aw_next = fw->fw_adds;
|
|
1856 fw->fw_adds = aw;
|
|
1857 aw->aw_saveb = NUL;
|
|
1858 }
|
|
1859 else
|
|
1860 {
|
|
1861 /* Put addition in hashtable. For key we use the part up
|
|
1862 * to the next end-of-word. */
|
|
1863 if (leadlen == 0)
|
|
1864 {
|
|
1865 p = aw->aw_word;
|
|
1866 while (*p != NUL && !spell_iswordc(p))
|
|
1867 mb_ptr_adv(p);
|
|
1868 }
|
|
1869
|
|
1870 if (leadlen != 0 || *p == NUL)
|
|
1871 {
|
|
1872 /* Only non-word characters in addition, add it to the
|
|
1873 * list with the special key NOWC_KEY. Also do this
|
|
1874 * when there is a leadstring, it would get too
|
|
1875 * complicated. */
|
|
1876 hash = hash_hash(NOWC_KEY);
|
|
1877 hi = hash_lookup(ht, NOWC_KEY, hash);
|
|
1878 if (HASHITEM_EMPTY(hi))
|
|
1879 {
|
|
1880 /* we use a dummy item as the list header */
|
|
1881 naw = (addword_T *)getroom(lp, &bl_used,
|
|
1882 sizeof(addword_T) + STRLEN(NOWC_KEY));
|
|
1883 if (naw == NULL)
|
|
1884 goto errorend;
|
|
1885 STRCPY(naw->aw_word, NOWC_KEY);
|
|
1886 hash_add_item(ht, hi, naw->aw_word, hash);
|
|
1887 naw->aw_next = aw;
|
|
1888 aw->aw_next = NULL;
|
|
1889 }
|
|
1890 else
|
|
1891 {
|
|
1892 naw = HI2ADDWORD(hi);
|
|
1893 aw->aw_next = naw->aw_next;
|
|
1894 naw->aw_next = aw;
|
|
1895 }
|
|
1896 aw->aw_saveb = NUL;
|
|
1897 }
|
|
1898 else
|
|
1899 {
|
|
1900 /* Truncate at next non-word character, store that
|
|
1901 * byte in "aw_saveb". */
|
|
1902 while (*p != NUL && spell_iswordc(p))
|
|
1903 mb_ptr_adv(p);
|
|
1904 aw->aw_saveb = *p;
|
|
1905 *p = NUL;
|
|
1906 hash = hash_hash(aw->aw_word);
|
|
1907 hi = hash_lookup(ht, aw->aw_word, hash);
|
|
1908 if (HASHITEM_EMPTY(hi))
|
|
1909 {
|
|
1910 hash_add_item(ht, hi, aw->aw_word, hash);
|
|
1911 aw->aw_next = NULL;
|
|
1912 }
|
|
1913 else
|
|
1914 {
|
|
1915 naw = HI2ADDWORD(hi);
|
|
1916 aw->aw_next = naw->aw_next;
|
|
1917 naw->aw_next = aw;
|
|
1918 }
|
|
1919 }
|
|
1920 }
|
236
|
1921 }
|
|
1922 }
|
|
1923 }
|
|
1924 goto end_OK;
|
|
1925
|
|
1926 errorend:
|
|
1927 lp->sl_error = TRUE;
|
|
1928 end_OK:
|
|
1929 if (fd != NULL)
|
|
1930 fclose(fd);
|
|
1931 hash_unlock(&lp->sl_words);
|
|
1932 sourcing_name = save_sourcing_name;
|
|
1933 sourcing_lnum = save_sourcing_lnum;
|
|
1934 }
|
|
1935
|
|
1936 /*
|
|
1937 * Get part of an sblock_T, at least "len" bytes long.
|
|
1938 * Returns NULL when out of memory.
|
|
1939 */
|
|
1940 static void *
|
|
1941 getroom(lp, bl_used, len)
|
|
1942 slang_T *lp; /* lp->sl_block is current block or NULL */
|
|
1943 int *bl_used; /* used up from current block */
|
|
1944 int len; /* length needed */
|
|
1945 {
|
|
1946 char_u *p;
|
|
1947 sblock_T *bl = lp->sl_block;
|
|
1948
|
|
1949 if (bl == NULL || *bl_used + len > SBLOCKSIZE)
|
|
1950 {
|
|
1951 /* Allocate a block of memory. This is not freed until spell_reload()
|
|
1952 * is called. */
|
|
1953 bl = (sblock_T *)alloc((unsigned)(sizeof(sblock_T) + SBLOCKSIZE));
|
|
1954 if (bl == NULL)
|
|
1955 return NULL;
|
|
1956 bl->sb_next = lp->sl_block;
|
|
1957 lp->sl_block = bl;
|
|
1958 *bl_used = 0;
|
|
1959 }
|
|
1960
|
|
1961 p = bl->sb_data + *bl_used;
|
|
1962 *bl_used += len;
|
|
1963
|
|
1964 return p;
|
|
1965 }
|
|
1966
|
|
1967 /*
|
|
1968 * Parse 'spelllang' and set buf->b_langp accordingly.
|
|
1969 * Returns an error message or NULL.
|
|
1970 */
|
|
1971 char_u *
|
|
1972 did_set_spelllang(buf)
|
|
1973 buf_T *buf;
|
|
1974 {
|
|
1975 garray_T ga;
|
|
1976 char_u *lang;
|
|
1977 char_u *e;
|
|
1978 char_u *region;
|
|
1979 int region_mask;
|
|
1980 slang_T *lp;
|
|
1981 int c;
|
|
1982 char_u lbuf[MAXWLEN + 1];
|
|
1983
|
|
1984 ga_init2(&ga, sizeof(langp_T), 2);
|
|
1985
|
|
1986 /* loop over comma separated languages. */
|
|
1987 for (lang = buf->b_p_spl; *lang != NUL; lang = e)
|
|
1988 {
|
|
1989 e = vim_strchr(lang, ',');
|
|
1990 if (e == NULL)
|
|
1991 e = lang + STRLEN(lang);
|
240
|
1992 region = NULL;
|
236
|
1993 if (e > lang + 2)
|
|
1994 {
|
|
1995 if (e - lang >= MAXWLEN)
|
|
1996 {
|
|
1997 ga_clear(&ga);
|
|
1998 return e_invarg;
|
|
1999 }
|
|
2000 if (lang[2] == '_')
|
|
2001 region = lang + 3;
|
|
2002 }
|
|
2003
|
|
2004 for (lp = first_lang; lp != NULL; lp = lp->sl_next)
|
|
2005 if (STRNICMP(lp->sl_name, lang, 2) == 0)
|
|
2006 break;
|
|
2007
|
|
2008 if (lp == NULL)
|
|
2009 {
|
|
2010 /* Not found, load the language. */
|
|
2011 STRNCPY(lbuf, lang, e - lang);
|
|
2012 lbuf[e - lang] = NUL;
|
|
2013 if (region != NULL)
|
|
2014 mch_memmove(lbuf + 2, lbuf + 5, e - lang - 4);
|
|
2015 lp = spell_load_lang(lbuf);
|
|
2016 }
|
|
2017
|
|
2018 if (lp != NULL)
|
|
2019 {
|
|
2020 if (region == NULL)
|
|
2021 region_mask = REGION_ALL;
|
|
2022 else
|
|
2023 {
|
|
2024 /* find region in sl_regions */
|
|
2025 c = find_region(lp->sl_regions, region);
|
|
2026 if (c == REGION_ALL)
|
|
2027 {
|
|
2028 c = *e;
|
|
2029 *e = NUL;
|
|
2030 smsg((char_u *)_("Warning: region %s not supported"), lang);
|
|
2031 *e = c;
|
|
2032 region_mask = REGION_ALL;
|
|
2033 }
|
|
2034 else
|
|
2035 region_mask = 1 << c;
|
|
2036 }
|
|
2037
|
|
2038 if (ga_grow(&ga, 1) == FAIL)
|
|
2039 {
|
|
2040 ga_clear(&ga);
|
|
2041 return e_outofmem;
|
|
2042 }
|
|
2043 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = lp;
|
|
2044 LANGP_ENTRY(ga, ga.ga_len)->lp_region = region_mask;
|
|
2045 ++ga.ga_len;
|
|
2046 }
|
|
2047
|
|
2048 if (*e == ',')
|
|
2049 ++e;
|
|
2050 }
|
|
2051
|
|
2052 /* Add a NULL entry to mark the end of the list. */
|
|
2053 if (ga_grow(&ga, 1) == FAIL)
|
|
2054 {
|
|
2055 ga_clear(&ga);
|
|
2056 return e_outofmem;
|
|
2057 }
|
|
2058 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = NULL;
|
|
2059 ++ga.ga_len;
|
|
2060
|
|
2061 /* Everything is fine, store the new b_langp value. */
|
|
2062 ga_clear(&buf->b_langp);
|
|
2063 buf->b_langp = ga;
|
|
2064
|
|
2065 return NULL;
|
|
2066 }
|
|
2067
|
|
2068 /*
|
|
2069 * Find the region "region[2]" in "rp" (points to "sl_regions").
|
|
2070 * Each region is simply stored as the two characters of it's name.
|
|
2071 * Returns the index if found, REGION_ALL if not found.
|
|
2072 */
|
|
2073 static int
|
|
2074 find_region(rp, region)
|
|
2075 char_u *rp;
|
|
2076 char_u *region;
|
|
2077 {
|
|
2078 int i;
|
|
2079
|
|
2080 for (i = 0; ; i += 2)
|
|
2081 {
|
|
2082 if (rp[i] == NUL)
|
|
2083 return REGION_ALL;
|
|
2084 if (rp[i] == region[0] && rp[i + 1] == region[1])
|
|
2085 break;
|
|
2086 }
|
|
2087 return i / 2;
|
|
2088 }
|
|
2089
|
|
2090 /*
|
|
2091 * Return type of word:
|
|
2092 * w word 0
|
|
2093 * Word BWF_ONECAP
|
|
2094 * W WORD BWF_ALLCAP
|
|
2095 * WoRd wOrd BWF_KEEPCAP
|
|
2096 */
|
|
2097 static int
|
|
2098 captype(word, end)
|
|
2099 char_u *word;
|
|
2100 char_u *end;
|
|
2101 {
|
|
2102 char_u *p;
|
|
2103 int c;
|
|
2104 int firstcap;
|
|
2105 int allcap;
|
|
2106 int past_second = FALSE; /* past second word char */
|
|
2107
|
|
2108 /* find first letter */
|
|
2109 for (p = word; !spell_iswordc(p); mb_ptr_adv(p))
|
|
2110 if (p >= end)
|
|
2111 return 0; /* only non-word characters, illegal word */
|
|
2112 #ifdef FEAT_MBYTE
|
|
2113 c = mb_ptr2char_adv(&p);
|
|
2114 #else
|
|
2115 c = *p++;
|
|
2116 #endif
|
|
2117 firstcap = allcap = MB_ISUPPER(c);
|
|
2118
|
|
2119 /*
|
|
2120 * Need to check all letters to find a word with mixed upper/lower.
|
|
2121 * But a word with an upper char only at start is a ONECAP.
|
|
2122 */
|
|
2123 for ( ; p < end; mb_ptr_adv(p))
|
|
2124 if (spell_iswordc(p))
|
|
2125 {
|
|
2126 #ifdef FEAT_MBYTE
|
|
2127 c = mb_ptr2char(p);
|
|
2128 #else
|
|
2129 c = *p;
|
|
2130 #endif
|
|
2131 if (!MB_ISUPPER(c))
|
|
2132 {
|
|
2133 /* UUl -> KEEPCAP */
|
|
2134 if (past_second && allcap)
|
|
2135 return BWF_KEEPCAP;
|
|
2136 allcap = FALSE;
|
|
2137 }
|
|
2138 else if (!allcap)
|
|
2139 /* UlU -> KEEPCAP */
|
|
2140 return BWF_KEEPCAP;
|
|
2141 past_second = TRUE;
|
|
2142 }
|
|
2143
|
|
2144 if (allcap)
|
|
2145 return BWF_ALLCAP;
|
|
2146 if (firstcap)
|
|
2147 return BWF_ONECAP;
|
|
2148 return 0;
|
|
2149 }
|
|
2150
|
|
2151 # if defined(FEAT_MBYTE) || defined(PROTO)
|
|
2152 /*
|
|
2153 * Clear all spelling tables and reload them.
|
|
2154 * Used after 'encoding' is set.
|
|
2155 */
|
|
2156 void
|
|
2157 spell_reload()
|
|
2158 {
|
|
2159 buf_T *buf;
|
|
2160 slang_T *lp;
|
|
2161
|
|
2162 /* Initialize the table for spell_iswordc(). */
|
|
2163 init_spell_chartab();
|
|
2164
|
|
2165 /* Unload all allocated memory. */
|
|
2166 while (first_lang != NULL)
|
|
2167 {
|
|
2168 lp = first_lang;
|
|
2169 first_lang = lp->sl_next;
|
|
2170 slang_free(lp);
|
|
2171 }
|
|
2172
|
|
2173 /* Go through all buffers and handle 'spelllang'. */
|
|
2174 for (buf = firstbuf; buf != NULL; buf = buf->b_next)
|
|
2175 {
|
|
2176 ga_clear(&buf->b_langp);
|
|
2177 if (*buf->b_p_spl != NUL)
|
|
2178 did_set_spelllang(buf);
|
|
2179 }
|
|
2180 }
|
|
2181 # endif
|
|
2182
|
|
2183 /*
|
|
2184 * Recognizing words uses a two-step mechanism:
|
|
2185 * 1. Locate a basic word, made out of word characters only and separated by
|
|
2186 * non-word characters.
|
|
2187 * 2. When a basic word is found, check if (possibly required) additions
|
|
2188 * before and after the word are present.
|
|
2189 *
|
|
2190 * Both mechanisms use affixes (prefixes and suffixes) to reduce the number of
|
|
2191 * words. When no matching word was found in the hashtable the start of the
|
|
2192 * word is checked for matching prefixes and the end of the word for matching
|
|
2193 * suffixes. All matching affixes are removed and then the resulting word is
|
|
2194 * searched for. If found it is checked if it supports the used affix.
|
|
2195 */
|
|
2196
|
|
2197
|
|
2198 #if defined(FEAT_MBYTE) || defined(PROTO)
|
|
2199 /*
|
|
2200 * Functions for ":mkspell".
|
|
2201 * Only possible with the multi-byte feature.
|
|
2202 */
|
|
2203
|
|
2204 #define MAXLINELEN 300 /* Maximum length in bytes of a line in a .aff
|
|
2205 and .dic file. */
|
|
2206 /*
|
|
2207 * Main structure to store the contents of a ".aff" file.
|
|
2208 */
|
|
2209 typedef struct afffile_S
|
|
2210 {
|
|
2211 char_u *af_enc; /* "SET", normalized, alloc'ed string or NULL */
|
|
2212 char_u *af_try; /* "TRY" line in "af_enc" encoding */
|
|
2213 hashtab_T af_pref; /* hashtable for prefixes, affheader_T */
|
|
2214 hashtab_T af_suff; /* hashtable for suffixes, affheader_T */
|
|
2215 garray_T af_rep; /* list of repentry_T entries from REP lines */
|
|
2216 } afffile_T;
|
|
2217
|
|
2218 typedef struct affentry_S affentry_T;
|
|
2219
|
|
2220 /* Affix header from ".aff" file. Used for af_pref and af_suff. */
|
|
2221 typedef struct affheader_S
|
|
2222 {
|
|
2223 char_u ah_key[2]; /* key for hashtable == name of affix entry */
|
|
2224 int ah_combine;
|
|
2225 affentry_T *ah_first; /* first affix entry */
|
|
2226 short_u ah_affnr; /* used in get_new_aff() */
|
|
2227 } affheader_T;
|
|
2228
|
|
2229 #define HI2AH(hi) ((affheader_T *)(hi)->hi_key)
|
|
2230
|
|
2231 /* Affix entry from ".aff" file. Used for prefixes and suffixes. */
|
|
2232 struct affentry_S
|
|
2233 {
|
|
2234 affentry_T *ae_next; /* next affix with same name/number */
|
|
2235 char_u *ae_chop; /* text to chop off basic word (can be NULL) */
|
|
2236 char_u *ae_add; /* text to add to basic word (can be NULL) */
|
240
|
2237 char_u *ae_add_nw; /* For a suffix: first non-word char in
|
|
2238 * "ae_add"; for a prefix with only non-word
|
|
2239 * chars: equal to "ae_add", for a prefix with
|
|
2240 * word and non-word chars: first non-word
|
|
2241 * char after word char. NULL otherwise. */
|
|
2242 char_u *ae_add_pw; /* For a prefix with both word and non-word
|
|
2243 * chars: first word char. NULL otherwise. */
|
243
|
2244 char_u ae_preword; /* TRUE for a prefix with one word */
|
236
|
2245 char_u *ae_cond; /* condition (NULL for ".") */
|
|
2246 regprog_T *ae_prog; /* regexp program for ae_cond or NULL */
|
|
2247 short_u ae_affnr; /* for old affix: new affix number */
|
|
2248 };
|
|
2249
|
|
2250 /*
|
|
2251 * Structure to store a word from a ".dic" file.
|
|
2252 */
|
|
2253 typedef struct dicword_S
|
|
2254 {
|
|
2255 char_u *dw_affnm; /* original affix names */
|
|
2256 char_u dw_word[1]; /* actually longer: the word in 'encoding' */
|
|
2257 } dicword_T;
|
|
2258
|
|
2259 static dicword_T dumdw;
|
|
2260 #define HI2DW(hi) ((dicword_T *)((hi)->hi_key - (dumdw.dw_word - (char_u *)&dumdw)))
|
|
2261
|
|
2262 /*
|
|
2263 * Structure to store a basic word for the spell file.
|
|
2264 * This is used for ":mkspell", not for spell checking.
|
|
2265 */
|
|
2266 typedef struct basicword_S basicword_T;
|
|
2267 struct basicword_S
|
|
2268 {
|
|
2269 basicword_T *bw_next; /* next word with same basic word */
|
|
2270 basicword_T *bw_cnext; /* next word with same caps */
|
|
2271 int bw_flags; /* BWF_ flags */
|
|
2272 garray_T bw_prefix; /* table with prefix numbers */
|
|
2273 garray_T bw_suffix; /* table with suffix numbers */
|
|
2274 int bw_region; /* region bits */
|
|
2275 char_u *bw_caseword; /* keep-case word */
|
|
2276 char_u *bw_leadstring; /* must come before bw_word */
|
|
2277 char_u *bw_addstring; /* must come after bw_word */
|
|
2278 char_u bw_word[1]; /* actually longer: word case folded */
|
|
2279 };
|
|
2280
|
|
2281 static basicword_T dumbw;
|
|
2282 #define KEY2BW(p) ((basicword_T *)((p) - (dumbw.bw_word - (char_u *)&dumbw)))
|
|
2283 #define HI2BW(hi) KEY2BW((hi)->hi_key)
|
|
2284
|
|
2285 /* Store the affix number related with a certain string. */
|
|
2286 typedef struct affhash_S
|
|
2287 {
|
|
2288 short_u as_nr; /* the affix nr */
|
|
2289 char_u as_word[1]; /* actually longer */
|
|
2290 } affhash_T;
|
|
2291
|
|
2292 static affhash_T dumas;
|
|
2293 #define HI2AS(hi) ((affhash_T *)((hi)->hi_key - (dumas.as_word - (char_u *)&dumas)))
|
|
2294
|
|
2295
|
240
|
2296 static afffile_T *spell_read_aff __ARGS((char_u *fname, vimconv_T *conv, int ascii));
|
236
|
2297 static void spell_free_aff __ARGS((afffile_T *aff));
|
240
|
2298 static int has_non_ascii __ARGS((char_u *s));
|
|
2299 static int spell_read_dic __ARGS((hashtab_T *ht, char_u *fname, vimconv_T *conv, int ascii));
|
|
2300 static int get_new_aff __ARGS((hashtab_T *oldaff, garray_T *gap, int prefix));
|
236
|
2301 static void spell_free_dic __ARGS((hashtab_T *dic));
|
|
2302 static int same_affentries __ARGS((affheader_T *ah1, affheader_T *ah2));
|
|
2303 static void add_affhash __ARGS((hashtab_T *ht, char_u *key, int newnr));
|
|
2304 static void clear_affhash __ARGS((hashtab_T *ht));
|
|
2305 static void trans_affixes __ARGS((dicword_T *dw, basicword_T *bw, afffile_T *oldaff, hashtab_T *newwords));
|
|
2306 static int build_wordlist __ARGS((hashtab_T *newwords, hashtab_T *oldwords, afffile_T *oldaff, int regionmask));
|
|
2307 static void combine_regions __ARGS((hashtab_T *newwords));
|
|
2308 static int same_affixes __ARGS((basicword_T *bw, basicword_T *nbw));
|
|
2309 static void expand_affixes __ARGS((hashtab_T *newwords, garray_T *prefgap, garray_T *suffgap));
|
|
2310 static void expand_one_aff __ARGS((basicword_T *bw, garray_T *add_words, affentry_T *pae, affentry_T *sae));
|
|
2311 static void add_to_wordlist __ARGS((hashtab_T *newwords, basicword_T *bw));
|
|
2312 static void put_bytes __ARGS((FILE *fd, long_u nr, int len));
|
|
2313 static void write_affix __ARGS((FILE *fd, affheader_T *ah));
|
|
2314 static void write_affixlist __ARGS((FILE *fd, garray_T *aff, int bytes));
|
|
2315 static void write_vim_spell __ARGS((char_u *fname, garray_T *prefga, garray_T *suffga, hashtab_T *newwords, int regcount, char_u *regchars));
|
|
2316 static void write_bword __ARGS((FILE *fd, basicword_T *bw, int lowcap, basicword_T **prevbw, int regionmask, int prefm, int suffm));
|
|
2317 static void free_wordtable __ARGS((hashtab_T *ht));
|
|
2318 static void free_basicword __ARGS((basicword_T *bw));
|
|
2319 static void free_affixentries __ARGS((affentry_T *first));
|
240
|
2320 static void free_affix_entry __ARGS((affentry_T *ap));
|
236
|
2321
|
|
2322 /*
|
|
2323 * Read an affix ".aff" file.
|
|
2324 * Returns an afffile_T, NULL for failure.
|
|
2325 */
|
|
2326 static afffile_T *
|
240
|
2327 spell_read_aff(fname, conv, ascii)
|
236
|
2328 char_u *fname;
|
|
2329 vimconv_T *conv; /* info for encoding conversion */
|
240
|
2330 int ascii; /* Only accept ASCII characters */
|
236
|
2331 {
|
|
2332 FILE *fd;
|
|
2333 afffile_T *aff;
|
|
2334 char_u rline[MAXLINELEN];
|
|
2335 char_u *line;
|
|
2336 char_u *pc = NULL;
|
|
2337 char_u *(items[6]);
|
|
2338 int itemcnt;
|
|
2339 char_u *p;
|
|
2340 int lnum = 0;
|
|
2341 affheader_T *cur_aff = NULL;
|
|
2342 int aff_todo = 0;
|
|
2343 hashtab_T *tp;
|
|
2344
|
|
2345 fd = fopen((char *)fname, "r");
|
|
2346 if (fd == NULL)
|
|
2347 {
|
|
2348 EMSG2(_(e_notopen), fname);
|
|
2349 return NULL;
|
|
2350 }
|
|
2351
|
|
2352 smsg((char_u *)_("Reading affix file %s..."), fname);
|
|
2353 out_flush();
|
|
2354
|
|
2355 aff = (afffile_T *)alloc_clear((unsigned)sizeof(afffile_T));
|
|
2356 if (aff == NULL)
|
|
2357 return NULL;
|
|
2358 hash_init(&aff->af_pref);
|
|
2359 hash_init(&aff->af_suff);
|
|
2360 ga_init2(&aff->af_rep, (int)sizeof(repentry_T), 20);
|
|
2361
|
|
2362 /*
|
|
2363 * Read all the lines in the file one by one.
|
|
2364 */
|
|
2365 while (!vim_fgets(rline, MAXLINELEN, fd))
|
|
2366 {
|
|
2367 ++lnum;
|
|
2368
|
|
2369 /* Skip comment lines. */
|
|
2370 if (*rline == '#')
|
|
2371 continue;
|
|
2372
|
|
2373 /* Convert from "SET" to 'encoding' when needed. */
|
|
2374 vim_free(pc);
|
|
2375 if (conv->vc_type != CONV_NONE)
|
|
2376 {
|
|
2377 pc = string_convert(conv, rline, NULL);
|
|
2378 line = pc;
|
|
2379 }
|
|
2380 else
|
|
2381 {
|
|
2382 pc = NULL;
|
|
2383 line = rline;
|
|
2384 }
|
|
2385
|
|
2386 /* Split the line up in white separated items. Put a NUL after each
|
|
2387 * item. */
|
|
2388 itemcnt = 0;
|
|
2389 for (p = line; ; )
|
|
2390 {
|
|
2391 while (*p != NUL && *p <= ' ') /* skip white space and CR/NL */
|
|
2392 ++p;
|
|
2393 if (*p == NUL)
|
|
2394 break;
|
|
2395 items[itemcnt++] = p;
|
|
2396 while (*p > ' ') /* skip until white space or CR/NL */
|
|
2397 ++p;
|
|
2398 if (*p == NUL)
|
|
2399 break;
|
|
2400 *p++ = NUL;
|
|
2401 }
|
|
2402
|
|
2403 /* Handle non-empty lines. */
|
|
2404 if (itemcnt > 0)
|
|
2405 {
|
|
2406 if (STRCMP(items[0], "SET") == 0 && itemcnt == 2
|
|
2407 && aff->af_enc == NULL)
|
|
2408 {
|
|
2409 if (aff->af_enc != NULL)
|
|
2410 smsg((char_u *)_("Duplicate SET line ignored in %s line %d: %s"),
|
|
2411 fname, lnum, line);
|
|
2412 else
|
|
2413 {
|
|
2414 /* Setup for conversion from "ENC" to 'encoding'. */
|
|
2415 aff->af_enc = enc_canonize(items[1]);
|
240
|
2416 if (aff->af_enc != NULL && !ascii
|
236
|
2417 && convert_setup(conv, aff->af_enc, p_enc) == FAIL)
|
|
2418 smsg((char_u *)_("Conversion in %s not supported: from %s to %s"),
|
|
2419 fname, aff->af_enc, p_enc);
|
|
2420 }
|
|
2421 }
|
|
2422 else if (STRCMP(items[0], "TRY") == 0 && itemcnt == 2
|
|
2423 && aff->af_try == NULL)
|
|
2424 aff->af_try = vim_strsave(items[1]);
|
|
2425 else if ((STRCMP(items[0], "PFX") == 0
|
|
2426 || STRCMP(items[0], "SFX") == 0)
|
|
2427 && aff_todo == 0
|
|
2428 && itemcnt == 4)
|
|
2429 {
|
|
2430 /* New affix letter. */
|
|
2431 cur_aff = (affheader_T *)alloc((unsigned)sizeof(affheader_T));
|
|
2432 if (cur_aff == NULL)
|
|
2433 break;
|
|
2434 cur_aff->ah_key[0] = *items[1];
|
|
2435 cur_aff->ah_key[1] = NUL;
|
|
2436 if (items[1][1] != NUL)
|
|
2437 smsg((char_u *)_("Affix name too long in %s line %d: %s"),
|
|
2438 fname, lnum, items[1]);
|
|
2439 if (*items[2] == 'Y')
|
|
2440 cur_aff->ah_combine = TRUE;
|
|
2441 else if (*items[2] == 'N')
|
|
2442 cur_aff->ah_combine = FALSE;
|
|
2443 else if (p_verbose > 0)
|
|
2444 smsg((char_u *)_("Expected Y or N in %s line %d: %s"),
|
|
2445 fname, lnum, items[2]);
|
|
2446 cur_aff->ah_first = NULL;
|
|
2447 if (*items[0] == 'P')
|
|
2448 tp = &aff->af_pref;
|
|
2449 else
|
|
2450 tp = &aff->af_suff;
|
|
2451 if (!HASHITEM_EMPTY(hash_find(tp, cur_aff->ah_key)))
|
|
2452 smsg((char_u *)_("Duplicate affix in %s line %d: %s"),
|
|
2453 fname, lnum, items[1]);
|
|
2454 else
|
|
2455 hash_add(tp, cur_aff->ah_key);
|
|
2456
|
|
2457 aff_todo = atoi((char *)items[3]);
|
|
2458 }
|
|
2459 else if ((STRCMP(items[0], "PFX") == 0
|
|
2460 || STRCMP(items[0], "SFX") == 0)
|
|
2461 && aff_todo > 0
|
|
2462 && STRCMP(cur_aff->ah_key, items[1]) == 0
|
|
2463 && itemcnt == 5)
|
|
2464 {
|
|
2465 affentry_T *aff_entry;
|
|
2466
|
|
2467 /* New item for an affix letter. */
|
|
2468 --aff_todo;
|
|
2469 aff_entry = (affentry_T *)alloc_clear(
|
|
2470 (unsigned)sizeof(affentry_T));
|
|
2471 if (aff_entry == NULL)
|
|
2472 break;
|
240
|
2473
|
236
|
2474 if (STRCMP(items[2], "0") != 0)
|
|
2475 aff_entry->ae_chop = vim_strsave(items[2]);
|
|
2476 if (STRCMP(items[3], "0") != 0)
|
|
2477 aff_entry->ae_add = vim_strsave(items[3]);
|
|
2478 if (STRCMP(items[4], ".") != 0)
|
|
2479 {
|
|
2480 char_u buf[MAXLINELEN];
|
|
2481
|
|
2482 aff_entry->ae_cond = vim_strsave(items[4]);
|
|
2483 if (*items[0] == 'P')
|
|
2484 sprintf((char *)buf, "^%s", items[4]);
|
|
2485 else
|
|
2486 sprintf((char *)buf, "%s$", items[4]);
|
|
2487 aff_entry->ae_prog = vim_regcomp(buf, RE_MAGIC + RE_STRING);
|
|
2488 }
|
240
|
2489
|
|
2490 if (ascii && (has_non_ascii(aff_entry->ae_chop)
|
|
2491 || has_non_ascii(aff_entry->ae_add)))
|
|
2492 {
|
|
2493 /* Don't use an affix entry with non-ASCII characters when
|
|
2494 * "ascii" is TRUE. */
|
|
2495 free_affix_entry(aff_entry);
|
|
2496 }
|
|
2497 else
|
|
2498 {
|
|
2499 aff_entry->ae_next = cur_aff->ah_first;
|
|
2500 cur_aff->ah_first = aff_entry;
|
|
2501 }
|
236
|
2502 }
|
|
2503 else if (STRCMP(items[0], "REP") == 0 && itemcnt == 2)
|
|
2504 /* Ignore REP count */;
|
|
2505 else if (STRCMP(items[0], "REP") == 0 && itemcnt == 3)
|
|
2506 {
|
|
2507 repentry_T *rp;
|
|
2508
|
|
2509 /* REP item */
|
|
2510 if (ga_grow(&aff->af_rep, 1) == FAIL)
|
|
2511 break;
|
|
2512 rp = ((repentry_T *)aff->af_rep.ga_data) + aff->af_rep.ga_len;
|
|
2513 rp->re_from = vim_strsave(items[1]);
|
|
2514 rp->re_to = vim_strsave(items[2]);
|
|
2515 ++aff->af_rep.ga_len;
|
|
2516 }
|
|
2517 else if (p_verbose > 0)
|
|
2518 smsg((char_u *)_("Unrecognized item in %s line %d: %s"),
|
|
2519 fname, lnum, items[0]);
|
|
2520 }
|
|
2521
|
|
2522 }
|
|
2523
|
|
2524 vim_free(pc);
|
|
2525 fclose(fd);
|
|
2526 return aff;
|
|
2527 }
|
|
2528
|
|
2529 /*
|
240
|
2530 * Return TRUE if string "s" contains a non-ASCII character (128 or higher).
|
|
2531 * When "s" is NULL FALSE is returned.
|
|
2532 */
|
|
2533 static int
|
|
2534 has_non_ascii(s)
|
|
2535 char_u *s;
|
|
2536 {
|
|
2537 char_u *p;
|
|
2538
|
|
2539 if (s != NULL)
|
|
2540 for (p = s; *p != NUL; ++p)
|
|
2541 if (*p >= 128)
|
|
2542 return TRUE;
|
|
2543 return FALSE;
|
|
2544 }
|
|
2545
|
|
2546 /*
|
236
|
2547 * Free the structure filled by spell_read_aff().
|
|
2548 */
|
|
2549 static void
|
|
2550 spell_free_aff(aff)
|
|
2551 afffile_T *aff;
|
|
2552 {
|
|
2553 hashtab_T *ht;
|
|
2554 hashitem_T *hi;
|
|
2555 int todo;
|
|
2556 int i;
|
|
2557 repentry_T *rp;
|
|
2558 affheader_T *ah;
|
|
2559
|
|
2560 vim_free(aff->af_enc);
|
|
2561 vim_free(aff->af_try);
|
|
2562
|
|
2563 for (ht = &aff->af_pref; ; ht = &aff->af_suff)
|
|
2564 {
|
|
2565 todo = ht->ht_used;
|
|
2566 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
2567 {
|
|
2568 if (!HASHITEM_EMPTY(hi))
|
|
2569 {
|
|
2570 --todo;
|
|
2571 ah = HI2AH(hi);
|
|
2572 free_affixentries(ah->ah_first);
|
|
2573 vim_free(ah);
|
|
2574 }
|
|
2575 }
|
|
2576 if (ht == &aff->af_suff)
|
|
2577 break;
|
|
2578 }
|
|
2579 hash_clear(&aff->af_pref);
|
|
2580 hash_clear(&aff->af_suff);
|
|
2581
|
|
2582 for (i = 0; i < aff->af_rep.ga_len; ++i)
|
|
2583 {
|
|
2584 rp = ((repentry_T *)aff->af_rep.ga_data) + i;
|
|
2585 vim_free(rp->re_from);
|
|
2586 vim_free(rp->re_to);
|
|
2587 }
|
|
2588 ga_clear(&aff->af_rep);
|
|
2589
|
|
2590 vim_free(aff);
|
|
2591 }
|
|
2592
|
|
2593 /*
|
|
2594 * Read a dictionary ".dic" file.
|
|
2595 * Returns OK or FAIL;
|
|
2596 * Each entry in the hashtab_T is a dicword_T.
|
|
2597 */
|
|
2598 static int
|
240
|
2599 spell_read_dic(ht, fname, conv, ascii)
|
236
|
2600 hashtab_T *ht;
|
|
2601 char_u *fname;
|
|
2602 vimconv_T *conv; /* info for encoding conversion */
|
240
|
2603 int ascii; /* only accept ASCII words */
|
236
|
2604 {
|
|
2605 char_u line[MAXLINELEN];
|
|
2606 char_u *p;
|
|
2607 dicword_T *dw;
|
|
2608 char_u *pc;
|
|
2609 char_u *w;
|
|
2610 int l;
|
|
2611 hash_T hash;
|
|
2612 hashitem_T *hi;
|
|
2613 FILE *fd;
|
|
2614 int lnum = 1;
|
|
2615
|
|
2616 fd = fopen((char *)fname, "r");
|
|
2617 if (fd == NULL)
|
|
2618 {
|
|
2619 EMSG2(_(e_notopen), fname);
|
|
2620 return FAIL;
|
|
2621 }
|
|
2622
|
|
2623 smsg((char_u *)_("Reading dictionary file %s..."), fname);
|
|
2624 out_flush();
|
|
2625
|
|
2626 /* Read and ignore the first line: word count. */
|
|
2627 (void)vim_fgets(line, MAXLINELEN, fd);
|
|
2628 if (!isdigit(*skipwhite(line)))
|
|
2629 EMSG2(_("E760: No word count in %s"), fname);
|
|
2630
|
|
2631 /*
|
|
2632 * Read all the lines in the file one by one.
|
|
2633 * The words are converted to 'encoding' here, before being added to
|
|
2634 * the hashtable.
|
|
2635 */
|
|
2636 while (!vim_fgets(line, MAXLINELEN, fd))
|
|
2637 {
|
|
2638 ++lnum;
|
|
2639
|
|
2640 /* Remove CR, LF and white space from end. */
|
|
2641 l = STRLEN(line);
|
|
2642 while (l > 0 && line[l - 1] <= ' ')
|
|
2643 --l;
|
|
2644 if (l == 0)
|
|
2645 continue; /* empty line */
|
|
2646 line[l] = NUL;
|
|
2647
|
|
2648 /* Find the optional affix names. */
|
|
2649 p = vim_strchr(line, '/');
|
|
2650 if (p != NULL)
|
|
2651 *p++ = NUL;
|
|
2652
|
240
|
2653 /* Skip non-ASCII words when "ascii" is TRUE. */
|
|
2654 if (ascii && has_non_ascii(line))
|
|
2655 continue;
|
|
2656
|
236
|
2657 /* Convert from "SET" to 'encoding' when needed. */
|
|
2658 if (conv->vc_type != CONV_NONE)
|
|
2659 {
|
|
2660 pc = string_convert(conv, line, NULL);
|
|
2661 w = pc;
|
|
2662 }
|
|
2663 else
|
|
2664 {
|
|
2665 pc = NULL;
|
|
2666 w = line;
|
|
2667 }
|
|
2668
|
|
2669 dw = (dicword_T *)alloc_clear((unsigned)sizeof(dicword_T)
|
|
2670 + STRLEN(w));
|
|
2671 if (dw == NULL)
|
|
2672 break;
|
|
2673 STRCPY(dw->dw_word, w);
|
|
2674 vim_free(pc);
|
|
2675
|
|
2676 hash = hash_hash(dw->dw_word);
|
|
2677 hi = hash_lookup(ht, dw->dw_word, hash);
|
|
2678 if (!HASHITEM_EMPTY(hi))
|
|
2679 smsg((char_u *)_("Duplicate word in %s line %d: %s"),
|
|
2680 fname, lnum, line);
|
|
2681 else
|
|
2682 hash_add_item(ht, hi, dw->dw_word, hash);
|
|
2683
|
|
2684 if (p != NULL)
|
|
2685 dw->dw_affnm = vim_strsave(p);
|
|
2686 }
|
|
2687
|
|
2688 fclose(fd);
|
|
2689 return OK;
|
|
2690 }
|
|
2691
|
|
2692 /*
|
|
2693 * Free the structure filled by spell_read_dic().
|
|
2694 */
|
|
2695 static void
|
|
2696 spell_free_dic(dic)
|
|
2697 hashtab_T *dic;
|
|
2698 {
|
|
2699 int todo;
|
|
2700 dicword_T *dw;
|
|
2701 hashitem_T *hi;
|
|
2702
|
|
2703 todo = dic->ht_used;
|
|
2704 for (hi = dic->ht_array; todo > 0; ++hi)
|
|
2705 {
|
|
2706 if (!HASHITEM_EMPTY(hi))
|
|
2707 {
|
|
2708 --todo;
|
|
2709 dw = HI2DW(hi);
|
|
2710 vim_free(dw->dw_affnm);
|
|
2711 vim_free(dw);
|
|
2712 }
|
|
2713 }
|
|
2714 hash_clear(dic);
|
|
2715 }
|
|
2716
|
|
2717 /*
|
|
2718 * Take the affixes read by spell_read_aff() and add them to the new list.
|
|
2719 * Attempts to re-use the same number for identical affixes (ignoring the
|
|
2720 * condition, since we remove that). That is especially important when using
|
|
2721 * multiple regions.
|
|
2722 * Returns OK or FAIL;
|
|
2723 */
|
|
2724 static int
|
240
|
2725 get_new_aff(oldaff, gap, prefix)
|
236
|
2726 hashtab_T *oldaff; /* hashtable with affheader_T */
|
|
2727 garray_T *gap; /* table with new affixes */
|
240
|
2728 int prefix; /* TRUE when doing prefixes, FALSE for
|
|
2729 suffixes */
|
236
|
2730 {
|
|
2731 int oldtodo;
|
|
2732 affheader_T *oldah, *newah, *gapah;
|
|
2733 affentry_T *oldae, *newae;
|
|
2734 hashitem_T *oldhi;
|
|
2735 hashitem_T *hi;
|
|
2736 hashtab_T condht; /* conditions already found */
|
|
2737 char_u condkey[MAXLINELEN];
|
|
2738 int newnr;
|
|
2739 int gapnr;
|
|
2740 int retval = OK;
|
|
2741 char_u *p;
|
|
2742 garray_T tga;
|
243
|
2743 int preword;
|
236
|
2744
|
|
2745 /*
|
|
2746 * Loop over all the old affix names.
|
|
2747 */
|
|
2748 oldtodo = oldaff->ht_used;
|
|
2749 for (oldhi = oldaff->ht_array; oldtodo > 0 && retval == OK; ++oldhi)
|
|
2750 {
|
|
2751 if (!HASHITEM_EMPTY(oldhi))
|
|
2752 {
|
|
2753 --oldtodo;
|
|
2754 oldah = (affheader_T *)oldhi->hi_key;
|
|
2755
|
|
2756 /* Put entries with the same condition under the same new affix
|
|
2757 * nr in "tga". Use hashtable "condht" to find them. */
|
|
2758 ga_init2(&tga, sizeof(affheader_T), 10);
|
|
2759 hash_init(&condht);
|
|
2760
|
|
2761 /*
|
|
2762 * Loop over all affixes with the same name.
|
|
2763 * The affixes with the same condition will get the same number,
|
|
2764 * since they can be used with the same words.
|
|
2765 * 1. build the lists of new affentry_T, with the headers in "tga".
|
|
2766 * 2. Check if some of the lists already exist in "gap", re-use
|
|
2767 * their number.
|
|
2768 * 3. Assign the new numbers to the old affixes.
|
|
2769 */
|
|
2770
|
|
2771 /* 1. build the lists of new affentry_T. */
|
|
2772 for (oldae = oldah->ah_first; oldae != NULL && retval == OK;
|
|
2773 oldae = oldae->ae_next)
|
|
2774 {
|
243
|
2775 preword = FALSE;
|
236
|
2776 oldae->ae_add_nw = NULL;
|
240
|
2777 oldae->ae_add_pw = NULL;
|
236
|
2778 if (oldae->ae_add != NULL)
|
|
2779 {
|
240
|
2780 /* Check for non-word characters in the affix. If there
|
243
|
2781 * is one a suffix will be turned into an addition, a
|
|
2782 * prefix may be turned into a leadstring.
|
236
|
2783 * This is stored with the old affix, that is where
|
|
2784 * trans_affixes() will check. */
|
|
2785 for (p = oldae->ae_add; *p != NUL; mb_ptr_adv(p))
|
|
2786 if (!spell_iswordc(p))
|
240
|
2787 {
|
|
2788 oldae->ae_add_nw = p;
|
236
|
2789 break;
|
240
|
2790 }
|
|
2791
|
|
2792 if (prefix && oldae->ae_add_nw != NULL)
|
|
2793 {
|
243
|
2794 /* If a prefix has non-word characters special
|
|
2795 * treatment is necessary:
|
|
2796 * - If it has only non-word characters it becomes a
|
|
2797 * leadstring.
|
|
2798 * - If it has a sequence of word characters followed
|
|
2799 * by a non-word char it becomes a "preword": "d'",
|
|
2800 * "de-", "d'ai", etc.
|
|
2801 * - if it has another mix of word and non-word
|
|
2802 * characters the part before the last word char
|
|
2803 * becomes a leadstring: "'d", etc.
|
|
2804 */
|
240
|
2805 for (p = oldae->ae_add; *p != NUL; mb_ptr_adv(p))
|
|
2806 if (spell_iswordc(p))
|
|
2807 {
|
|
2808 oldae->ae_add_pw = p;
|
|
2809 break;
|
|
2810 }
|
|
2811 if (oldae->ae_add_pw != NULL)
|
|
2812 {
|
|
2813 /* Mixed prefix, set ae_add_nw to first non-word
|
|
2814 * char after ae_add_pw (if there is one). */
|
|
2815 oldae->ae_add_nw = NULL;
|
|
2816 for ( ; *p != NUL; mb_ptr_adv(p))
|
|
2817 if (!spell_iswordc(p))
|
|
2818 {
|
|
2819 oldae->ae_add_nw = p;
|
|
2820 break;
|
|
2821 }
|
243
|
2822 if (oldae->ae_add_nw != NULL)
|
|
2823 {
|
|
2824 preword = TRUE;
|
|
2825 oldae->ae_add_pw = NULL;
|
|
2826 oldae->ae_add_nw = NULL;
|
|
2827 }
|
240
|
2828 }
|
|
2829 }
|
236
|
2830 }
|
|
2831
|
|
2832 if (oldae->ae_cond == NULL)
|
|
2833 /* hashtable requires a non-empty key */
|
|
2834 STRCPY(condkey, "---");
|
|
2835 else
|
|
2836 STRCPY(condkey, oldae->ae_cond);
|
|
2837
|
|
2838 /* Look for an existing list with this name and condition. */
|
|
2839 hi = hash_find(&condht, condkey);
|
|
2840 if (!HASHITEM_EMPTY(hi))
|
|
2841 /* Match with existing affix, use that one. */
|
|
2842 newnr = HI2AS(hi)->as_nr;
|
|
2843 else
|
|
2844 {
|
|
2845 /* Add a new affix number. */
|
|
2846 newnr = tga.ga_len;
|
|
2847 if (ga_grow(&tga, 1) == FAIL)
|
|
2848 retval = FAIL;
|
|
2849 else
|
|
2850 {
|
|
2851 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2852 newah->ah_combine = oldah->ah_combine;
|
|
2853 newah->ah_first = NULL;
|
|
2854 ++tga.ga_len;
|
|
2855
|
|
2856 /* Add the new list to the condht hashtable. */
|
|
2857 add_affhash(&condht, condkey, newnr);
|
|
2858 }
|
|
2859 }
|
|
2860
|
|
2861 /* Add the new affentry_T to the list. */
|
|
2862 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2863 newae = (affentry_T *)alloc_clear((unsigned)sizeof(affentry_T));
|
|
2864 if (newae == NULL)
|
|
2865 retval = FAIL;
|
|
2866 else
|
|
2867 {
|
|
2868 newae->ae_next = newah->ah_first;
|
|
2869 newah->ah_first = newae;
|
|
2870 if (oldae->ae_chop == NULL)
|
|
2871 newae->ae_chop = NULL;
|
|
2872 else
|
|
2873 newae->ae_chop = vim_strsave(oldae->ae_chop);
|
|
2874 if (oldae->ae_add == NULL)
|
|
2875 newae->ae_add = NULL;
|
|
2876 else
|
|
2877 newae->ae_add = vim_strsave(oldae->ae_add);
|
243
|
2878 newae->ae_preword = preword;
|
236
|
2879
|
|
2880 /* The condition is not copied, since the new affix is
|
|
2881 * only used for words where the condition matches. */
|
|
2882 }
|
|
2883 }
|
|
2884
|
|
2885 /* 2. Check if some of the lists already exist, re-use their
|
|
2886 * number. Otherwise add the list to "gap". */
|
|
2887 for (newnr = 0; newnr < tga.ga_len; ++newnr)
|
|
2888 {
|
|
2889 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2890 for (gapnr = 0; gapnr < gap->ga_len; ++gapnr)
|
|
2891 {
|
|
2892 gapah = ((affheader_T *)gap->ga_data) + gapnr;
|
243
|
2893 if (newah->ah_combine == gapah->ah_combine
|
|
2894 && same_affentries(newah, gapah))
|
236
|
2895 /* Found an existing affheader_T entry with same
|
|
2896 * affentry_T list, use its number. */
|
|
2897 break;
|
|
2898 }
|
|
2899
|
|
2900 newah->ah_affnr = gapnr;
|
|
2901 if (gapnr == gap->ga_len)
|
|
2902 {
|
|
2903 /* This is a new affentry_T list, add it. */
|
|
2904 if (ga_grow(gap, 1) == FAIL)
|
|
2905 retval = FAIL;
|
|
2906 else
|
|
2907 {
|
|
2908 *(((affheader_T *)gap->ga_data) + gap->ga_len) = *newah;
|
|
2909 ++gap->ga_len;
|
|
2910 }
|
|
2911 }
|
|
2912 else
|
|
2913 {
|
|
2914 /* free unused affentry_T list */
|
|
2915 free_affixentries(newah->ah_first);
|
|
2916 }
|
|
2917 }
|
|
2918
|
|
2919 /* 3. Assign the new affix numbers to the old affixes. */
|
|
2920 for (oldae = oldah->ah_first; oldae != NULL && retval == OK;
|
|
2921 oldae = oldae->ae_next)
|
|
2922 {
|
|
2923 if (oldae->ae_cond == NULL)
|
|
2924 /* hashtable requires a non-empty key */
|
|
2925 STRCPY(condkey, "---");
|
|
2926 else
|
|
2927 STRCPY(condkey, oldae->ae_cond);
|
|
2928
|
|
2929 /* Look for an existing affix with this name and condition. */
|
|
2930 hi = hash_find(&condht, condkey);
|
|
2931 if (!HASHITEM_EMPTY(hi))
|
|
2932 /* Match with existing affix, use that one. */
|
|
2933 newnr = HI2AS(hi)->as_nr;
|
|
2934 else
|
|
2935 {
|
|
2936 EMSG(_(e_internal));
|
|
2937 retval = FAIL;
|
|
2938 }
|
|
2939 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2940 oldae->ae_affnr = newah->ah_affnr;
|
|
2941 }
|
|
2942
|
|
2943 ga_clear(&tga);
|
|
2944 clear_affhash(&condht);
|
|
2945 }
|
|
2946 }
|
|
2947
|
|
2948 return retval;
|
|
2949 }
|
|
2950
|
|
2951 /*
|
|
2952 * Return TRUE if the affentry_T lists for "ah1" and "ah2" contain the same
|
|
2953 * items, ignoring the order.
|
|
2954 * Only compares the chop and add strings, not the condition.
|
|
2955 */
|
|
2956 static int
|
|
2957 same_affentries(ah1, ah2)
|
|
2958 affheader_T *ah1;
|
|
2959 affheader_T *ah2;
|
|
2960 {
|
|
2961 affentry_T *ae1, *ae2;
|
|
2962
|
|
2963 /* Check the length of the lists first. */
|
|
2964 ae2 = ah2->ah_first;
|
|
2965 for (ae1 = ah1->ah_first; ae1 != NULL; ae1 = ae1->ae_next)
|
|
2966 {
|
|
2967 if (ae2 == NULL)
|
|
2968 return FALSE; /* "ah1" list is longer */
|
|
2969 ae2 = ae2->ae_next;
|
|
2970 }
|
|
2971 if (ae2 != NULL)
|
|
2972 return FALSE; /* "ah2" list is longer */
|
|
2973
|
|
2974 /* Check that each entry in "ah1" appears in "ah2". */
|
|
2975 for (ae1 = ah1->ah_first; ae1 != NULL; ae1 = ae1->ae_next)
|
|
2976 {
|
|
2977 for (ae2 = ah2->ah_first; ae2 != NULL; ae2 = ae2->ae_next)
|
|
2978 {
|
|
2979 if ((ae1->ae_chop == NULL) == (ae2->ae_chop == NULL)
|
|
2980 && (ae1->ae_add == NULL) == (ae2->ae_add == NULL)
|
|
2981 && (ae1->ae_chop == NULL
|
|
2982 || STRCMP(ae1->ae_chop, ae2->ae_chop) == 0)
|
|
2983 && (ae1->ae_add == NULL
|
|
2984 || STRCMP(ae1->ae_add, ae2->ae_add) == 0))
|
|
2985 break;
|
|
2986 }
|
|
2987 if (ae2 == NULL)
|
|
2988 return FALSE;
|
|
2989 }
|
|
2990
|
|
2991 return TRUE;
|
|
2992 }
|
|
2993
|
|
2994 /*
|
|
2995 * Add a chop/add or cond hashtable entry.
|
|
2996 */
|
|
2997 static void
|
|
2998 add_affhash(ht, key, newnr)
|
|
2999 hashtab_T *ht;
|
|
3000 char_u *key;
|
|
3001 int newnr;
|
|
3002 {
|
|
3003 affhash_T *as;
|
|
3004
|
|
3005 as = (affhash_T *)alloc((unsigned)sizeof(affhash_T) + STRLEN(key));
|
|
3006 if (as != NULL)
|
|
3007 {
|
|
3008 as->as_nr = newnr;
|
|
3009 STRCPY(as->as_word, key);
|
|
3010 hash_add(ht, as->as_word);
|
|
3011 }
|
|
3012 }
|
|
3013
|
|
3014 /*
|
|
3015 * Clear the chop/add hashtable used to detect identical affixes.
|
|
3016 */
|
|
3017 static void
|
|
3018 clear_affhash(ht)
|
|
3019 hashtab_T *ht;
|
|
3020 {
|
|
3021 int todo;
|
|
3022 hashitem_T *hi;
|
|
3023
|
|
3024 todo = ht->ht_used;
|
|
3025 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
3026 {
|
|
3027 if (!HASHITEM_EMPTY(hi))
|
|
3028 {
|
|
3029 --todo;
|
|
3030 vim_free(HI2AS(hi));
|
|
3031 }
|
|
3032 }
|
|
3033 hash_clear(ht);
|
|
3034 }
|
|
3035
|
|
3036 /*
|
|
3037 * Translate list of affix names for an old word to affix numbers in a new
|
|
3038 * basic word.
|
|
3039 * This checks if the conditions match with the old word. The result is that
|
|
3040 * the new affix does not need to store the condition.
|
|
3041 */
|
|
3042 static void
|
|
3043 trans_affixes(dw, bw, oldaff, newwords)
|
|
3044 dicword_T *dw; /* old word */
|
|
3045 basicword_T *bw; /* basic word */
|
|
3046 afffile_T *oldaff; /* affixes for "oldwords" */
|
|
3047 hashtab_T *newwords; /* table with words */
|
|
3048 {
|
|
3049 char_u key[2];
|
|
3050 char_u *p;
|
|
3051 char_u *affnm;
|
|
3052 garray_T *gap;
|
|
3053 hashitem_T *aff_hi;
|
|
3054 affheader_T *ah;
|
|
3055 affentry_T *ae;
|
|
3056 regmatch_T regmatch;
|
|
3057 int i;
|
|
3058 basicword_T *nbw;
|
|
3059 int alen;
|
|
3060 int wlen;
|
240
|
3061 garray_T suffixga; /* list of words with non-word suffixes */
|
|
3062 garray_T prefixga; /* list of words with non-word prefixes */
|
236
|
3063 char_u nword[MAXWLEN];
|
|
3064 int flags;
|
|
3065 int n;
|
|
3066
|
240
|
3067 ga_init2(&suffixga, (int)sizeof(basicword_T *), 5);
|
|
3068 ga_init2(&prefixga, (int)sizeof(basicword_T *), 5);
|
236
|
3069
|
|
3070 /* Loop over all the affix names of the old word. */
|
|
3071 key[1] = NUL;
|
|
3072 for (affnm = dw->dw_affnm; *affnm != NUL; ++affnm)
|
|
3073 {
|
|
3074 key[0] = *affnm;
|
|
3075 aff_hi = hash_find(&oldaff->af_pref, key);
|
|
3076 if (!HASHITEM_EMPTY(aff_hi))
|
|
3077 gap = &bw->bw_prefix; /* found a prefix */
|
|
3078 else
|
|
3079 {
|
|
3080 gap = &bw->bw_suffix; /* must be a suffix */
|
|
3081 aff_hi = hash_find(&oldaff->af_suff, key);
|
|
3082 if (HASHITEM_EMPTY(aff_hi))
|
|
3083 {
|
|
3084 smsg((char_u *)_("No affix entry '%s' for word %s"),
|
|
3085 key, dw->dw_word);
|
|
3086 continue;
|
|
3087 }
|
|
3088 }
|
|
3089
|
|
3090 /* Loop over all the affix entries for this affix name. */
|
|
3091 ah = HI2AH(aff_hi);
|
|
3092 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
|
3093 {
|
243
|
3094 /* Setup for regexp matching. Note that we don't ignore case.
|
|
3095 * This is weird, because he rules in an .aff file don't care
|
|
3096 * about case, but it's necessary for compatibility with Myspell.
|
|
3097 */
|
236
|
3098 regmatch.regprog = ae->ae_prog;
|
243
|
3099 regmatch.rm_ic = FALSE;
|
236
|
3100 if (ae->ae_prog == NULL
|
|
3101 || vim_regexec(®match, dw->dw_word, (colnr_T)0))
|
|
3102 {
|
240
|
3103 if ((ae->ae_add_nw != NULL || ae->ae_add_pw != NULL)
|
|
3104 && (gap != &bw->bw_suffix || bw->bw_addstring == NULL))
|
236
|
3105 {
|
|
3106 /* Affix has a non-word character and isn't prepended to
|
|
3107 * leader or appended to addition. Need to use another
|
|
3108 * word with an addition. It's a copy of the basicword_T
|
|
3109 * "bw". */
|
|
3110 if (gap == &bw->bw_suffix)
|
|
3111 {
|
|
3112 alen = ae->ae_add_nw - ae->ae_add;
|
|
3113 nbw = (basicword_T *)alloc((unsigned)(
|
|
3114 sizeof(basicword_T) + STRLEN(bw->bw_word)
|
|
3115 + alen + 1));
|
|
3116 if (nbw != NULL)
|
|
3117 {
|
|
3118 *nbw = *bw;
|
|
3119 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3120 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3121
|
|
3122 /* Adding the suffix may change the caps. */
|
|
3123 STRCPY(nword, dw->dw_word);
|
|
3124 if (ae->ae_chop != NULL)
|
|
3125 {
|
|
3126 /* Remove chop string. */
|
|
3127 p = nword + STRLEN(nword);
|
|
3128 for (i = mb_charlen(ae->ae_chop); i > 0; --i)
|
|
3129 mb_ptr_back(nword, p);
|
|
3130 *p = NUL;
|
|
3131 }
|
|
3132 STRCAT(nword, ae->ae_add);
|
|
3133 flags = captype(nword, nword + STRLEN(nword));
|
|
3134 if (flags & BWF_KEEPCAP)
|
|
3135 {
|
240
|
3136 /* "caseword" excludes the addition */
|
236
|
3137 nword[STRLEN(dw->dw_word) + alen] = NUL;
|
|
3138 nbw->bw_caseword = vim_strsave(nword);
|
|
3139 }
|
|
3140 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3141 | BWF_KEEPCAP);
|
|
3142 nbw->bw_flags |= flags;
|
|
3143
|
|
3144 if (bw->bw_leadstring != NULL)
|
|
3145 nbw->bw_leadstring =
|
|
3146 vim_strsave(bw->bw_leadstring);
|
|
3147 nbw->bw_addstring = vim_strsave(ae->ae_add_nw);
|
|
3148
|
|
3149 STRCPY(nbw->bw_word, bw->bw_word);
|
|
3150 if (alen > 0 || ae->ae_chop != NULL)
|
|
3151 {
|
240
|
3152 /* Suffix starts with word character and/or
|
|
3153 * chop off something. Append it to the word.
|
|
3154 * Add new word entry. */
|
236
|
3155 wlen = STRLEN(nbw->bw_word);
|
|
3156 if (ae->ae_chop != NULL)
|
|
3157 wlen -= STRLEN(ae->ae_chop);
|
|
3158 mch_memmove(nbw->bw_word + wlen, ae->ae_add,
|
|
3159 alen);
|
|
3160 nbw->bw_word[wlen + alen] = NUL;
|
|
3161 add_to_wordlist(newwords, nbw);
|
|
3162 }
|
|
3163 else
|
|
3164 /* Basic word is the same, link "nbw" after
|
|
3165 * "bw". */
|
|
3166 bw->bw_next = nbw;
|
|
3167
|
|
3168 /* Remember this word, we need to set bw_prefix
|
240
|
3169 * and bw_prefix later. */
|
|
3170 if (ga_grow(&suffixga, 1) == OK)
|
|
3171 ((basicword_T **)suffixga.ga_data)
|
|
3172 [suffixga.ga_len++] = nbw;
|
|
3173 }
|
|
3174 }
|
|
3175 else if (ae->ae_add_nw == NULL)
|
|
3176 {
|
|
3177 /* Prefix that starts with non-word char(s) and may be
|
|
3178 * followed by word chars: Make a leadstring and
|
|
3179 * prepend word chars before the word. */
|
|
3180 alen = STRLEN(ae->ae_add_pw);
|
|
3181 nbw = (basicword_T *)alloc((unsigned)(
|
|
3182 sizeof(basicword_T) + STRLEN(bw->bw_word)
|
|
3183 + alen + 1));
|
|
3184 if (nbw != NULL)
|
|
3185 {
|
|
3186 *nbw = *bw;
|
|
3187 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3188 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3189
|
|
3190 /* Adding the prefix may change the caps. */
|
|
3191 STRCPY(nword, ae->ae_add);
|
|
3192 p = dw->dw_word;
|
|
3193 if (ae->ae_chop != NULL)
|
|
3194 /* Skip chop string. */
|
|
3195 for (i = mb_charlen(ae->ae_chop); i > 0; --i)
|
|
3196 mb_ptr_adv( p);
|
|
3197 STRCAT(nword, p);
|
|
3198
|
|
3199 flags = captype(nword, nword + STRLEN(nword));
|
|
3200 if (flags & BWF_KEEPCAP)
|
|
3201 /* "caseword" excludes the addition */
|
|
3202 nbw->bw_caseword = vim_strsave(nword
|
|
3203 + (ae->ae_add_pw - ae->ae_add));
|
|
3204 else
|
|
3205 nbw->bw_caseword = NULL;
|
|
3206 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3207 | BWF_KEEPCAP);
|
|
3208 nbw->bw_flags |= flags;
|
|
3209
|
|
3210 if (bw->bw_addstring != NULL)
|
|
3211 nbw->bw_addstring =
|
|
3212 vim_strsave(bw->bw_addstring);
|
|
3213 else
|
|
3214 nbw->bw_addstring = NULL;
|
|
3215 nbw->bw_leadstring = vim_strnsave(ae->ae_add,
|
|
3216 ae->ae_add_pw - ae->ae_add);
|
|
3217
|
|
3218 if (alen > 0 || ae->ae_chop != NULL)
|
|
3219 {
|
|
3220 /* Prefix ends in word character and/or chop
|
|
3221 * off something. Prepend it to the word.
|
|
3222 * Add new word entry. */
|
|
3223 STRCPY(nbw->bw_word, ae->ae_add_pw);
|
|
3224 p = bw->bw_word;
|
|
3225 if (ae->ae_chop != NULL)
|
|
3226 p += STRLEN(ae->ae_chop);
|
|
3227 STRCAT(nbw->bw_word, p);
|
|
3228 add_to_wordlist(newwords, nbw);
|
|
3229 }
|
|
3230 else
|
|
3231 {
|
|
3232 /* Basic word is the same, link "nbw" after
|
|
3233 * "bw". */
|
|
3234 STRCPY(nbw->bw_word, bw->bw_word);
|
|
3235 bw->bw_next = nbw;
|
|
3236 }
|
|
3237
|
|
3238 /* Remember this word, we need to set bw_suffix
|
236
|
3239 * and bw_suffix later. */
|
240
|
3240 if (ga_grow(&prefixga, 1) == OK)
|
|
3241 ((basicword_T **)prefixga.ga_data)
|
|
3242 [prefixga.ga_len++] = nbw;
|
236
|
3243 }
|
|
3244 }
|
|
3245 else
|
|
3246 {
|
240
|
3247 /* Prefix with both non-word and word characters: Turn
|
|
3248 * prefix into basic word, original word becomes an
|
|
3249 * addstring. */
|
|
3250
|
|
3251 /* Fold-case the word characters in the prefix into
|
|
3252 * nword[]. */
|
|
3253 alen = 0;
|
|
3254 for (p = ae->ae_add_pw; p < ae->ae_add_nw; p += n)
|
|
3255 {
|
|
3256 #ifdef FEAT_MBYTE
|
|
3257 n = (*mb_ptr2len_check)(p);
|
|
3258 #else
|
|
3259 n = 1;
|
|
3260 #endif
|
|
3261 (void)str_foldcase(p, n, nword + alen,
|
|
3262 MAXWLEN - alen);
|
|
3263 alen += STRLEN(nword + alen);
|
|
3264 }
|
|
3265
|
|
3266 /* Allocate a new word entry. */
|
|
3267 nbw = (basicword_T *)alloc((unsigned)(
|
|
3268 sizeof(basicword_T) + alen + 1));
|
|
3269 if (nbw != NULL)
|
|
3270 {
|
|
3271 *nbw = *bw;
|
|
3272 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3273 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3274
|
|
3275 mch_memmove(nbw->bw_word, nword, alen);
|
|
3276 nbw->bw_word[alen] = NUL;
|
|
3277
|
|
3278 /* Use the cap type of the prefix. */
|
|
3279 alen = ae->ae_add_nw - ae->ae_add_pw;
|
|
3280 mch_memmove(nword, ae->ae_add_pw, alen);
|
|
3281 nword[alen] = NUL;
|
|
3282 flags = captype(nword, nword + STRLEN(nword));
|
|
3283 if (flags & BWF_KEEPCAP)
|
|
3284 nbw->bw_caseword = vim_strsave(nword);
|
|
3285 else
|
|
3286 nbw->bw_caseword = NULL;
|
|
3287 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3288 | BWF_KEEPCAP);
|
|
3289 nbw->bw_flags |= flags;
|
|
3290
|
|
3291 /* The addstring is the prefix after the word
|
|
3292 * characters, the original word excluding "chop",
|
|
3293 * plus any addition. */
|
|
3294 STRCPY(nword, ae->ae_add_nw);
|
|
3295 p = bw->bw_word;
|
|
3296 if (ae->ae_chop != NULL)
|
|
3297 p += STRLEN(ae->ae_chop);
|
|
3298 STRCAT(nword, p);
|
|
3299 if (bw->bw_addstring != NULL)
|
|
3300 STRCAT(nword, bw->bw_addstring);
|
|
3301 nbw->bw_addstring = vim_strsave(nword);
|
|
3302
|
|
3303 if (ae->ae_add_pw > ae->ae_add)
|
|
3304 nbw->bw_leadstring = vim_strnsave(ae->ae_add,
|
|
3305 ae->ae_add_pw - ae->ae_add);
|
|
3306 else
|
|
3307 nbw->bw_leadstring = NULL;
|
|
3308
|
|
3309 add_to_wordlist(newwords, nbw);
|
|
3310
|
|
3311 /* Remember this word, we need to set bw_suffix
|
|
3312 * and bw_suffix later. */
|
|
3313 if (ga_grow(&prefixga, 1) == OK)
|
|
3314 ((basicword_T **)prefixga.ga_data)
|
|
3315 [prefixga.ga_len++] = nbw;
|
|
3316 }
|
236
|
3317 }
|
|
3318 }
|
|
3319 else
|
|
3320 {
|
|
3321 /* Affix applies to this word, add the related affix
|
|
3322 * number. But only if it's not there yet. And keep the
|
|
3323 * list sorted, so that we can compare it later. */
|
|
3324 for (i = 0; i < gap->ga_len; ++i)
|
|
3325 {
|
|
3326 n = ((short_u *)gap->ga_data)[i];
|
|
3327 if (n >= ae->ae_affnr)
|
|
3328 {
|
|
3329 if (n == ae->ae_affnr)
|
|
3330 i = -1;
|
|
3331 break;
|
|
3332 }
|
|
3333 }
|
|
3334 if (i >= 0 && ga_grow(gap, 1) == OK)
|
|
3335 {
|
|
3336 if (i < gap->ga_len)
|
|
3337 mch_memmove(((short_u *)gap->ga_data) + i + 1,
|
|
3338 ((short_u *)gap->ga_data) + i,
|
|
3339 sizeof(short_u) * (gap->ga_len - i));
|
|
3340 ((short_u *)gap->ga_data)[i] = ae->ae_affnr;
|
|
3341 ++gap->ga_len;
|
|
3342 }
|
|
3343 }
|
|
3344 }
|
|
3345 }
|
|
3346 }
|
|
3347
|
|
3348 /*
|
|
3349 * For the words that we added for suffixes with non-word characters: Use
|
|
3350 * the prefix list of the main word.
|
|
3351 */
|
240
|
3352 for (i = 0; i < suffixga.ga_len; ++i)
|
236
|
3353 {
|
240
|
3354 nbw = ((basicword_T **)suffixga.ga_data)[i];
|
236
|
3355 if (ga_grow(&nbw->bw_prefix, bw->bw_prefix.ga_len) == OK)
|
|
3356 {
|
|
3357 mch_memmove(nbw->bw_prefix.ga_data, bw->bw_prefix.ga_data,
|
|
3358 bw->bw_prefix.ga_len * sizeof(short_u));
|
|
3359 nbw->bw_prefix.ga_len = bw->bw_prefix.ga_len;
|
|
3360 }
|
|
3361 }
|
|
3362
|
240
|
3363 /*
|
|
3364 * For the words that we added for prefixes with non-word characters: Use
|
|
3365 * the suffix list of the main word.
|
|
3366 */
|
|
3367 for (i = 0; i < prefixga.ga_len; ++i)
|
|
3368 {
|
|
3369 nbw = ((basicword_T **)prefixga.ga_data)[i];
|
|
3370 if (ga_grow(&nbw->bw_suffix, bw->bw_suffix.ga_len) == OK)
|
|
3371 {
|
|
3372 mch_memmove(nbw->bw_suffix.ga_data, bw->bw_suffix.ga_data,
|
|
3373 bw->bw_suffix.ga_len * sizeof(short_u));
|
|
3374 nbw->bw_suffix.ga_len = bw->bw_suffix.ga_len;
|
|
3375 }
|
|
3376 }
|
|
3377
|
|
3378 ga_clear(&suffixga);
|
|
3379 ga_clear(&prefixga);
|
236
|
3380 }
|
|
3381
|
|
3382 /*
|
|
3383 * Go over all words in "oldwords" and change the old affix names to the new
|
|
3384 * affix numbers, check the conditions, fold case, extract the basic word and
|
|
3385 * additions.
|
|
3386 */
|
|
3387 static int
|
|
3388 build_wordlist(newwords, oldwords, oldaff, regionmask)
|
|
3389 hashtab_T *newwords; /* basicword_T entries */
|
|
3390 hashtab_T *oldwords; /* dicword_T entries */
|
|
3391 afffile_T *oldaff; /* affixes for "oldwords" */
|
|
3392 int regionmask; /* value for bw_region */
|
|
3393 {
|
|
3394 int todo;
|
|
3395 hashitem_T *old_hi;
|
|
3396 dicword_T *dw;
|
|
3397 basicword_T *bw;
|
|
3398 char_u foldword[MAXLINELEN];
|
|
3399 int leadlen;
|
|
3400 char_u leadstring[MAXLINELEN];
|
|
3401 int addlen;
|
|
3402 char_u addstring[MAXLINELEN];
|
|
3403 int dwlen;
|
|
3404 char_u *p;
|
|
3405 int clen;
|
|
3406 int flags;
|
240
|
3407 char_u *cp = NULL;
|
236
|
3408 int l;
|
240
|
3409 char_u message[MAXLINELEN + MAXWLEN];
|
236
|
3410
|
|
3411 todo = oldwords->ht_used;
|
|
3412 for (old_hi = oldwords->ht_array; todo > 0; ++old_hi)
|
|
3413 {
|
|
3414 if (!HASHITEM_EMPTY(old_hi))
|
|
3415 {
|
|
3416 --todo;
|
|
3417 dw = HI2DW(old_hi);
|
|
3418
|
|
3419 /* This takes time, print a message now and then. */
|
240
|
3420 if ((todo & 0x3ff) == 0 || todo == (int)oldwords->ht_used - 1)
|
236
|
3421 {
|
240
|
3422 sprintf((char *)message, _("%6d todo - %s"),
|
|
3423 todo, dw->dw_word);
|
|
3424 msg_start();
|
|
3425 msg_outtrans_attr(message, 0);
|
|
3426 msg_clr_eos();
|
|
3427 msg_didout = FALSE;
|
|
3428 msg_col = 0;
|
236
|
3429 out_flush();
|
|
3430 ui_breakcheck();
|
|
3431 if (got_int)
|
|
3432 break;
|
|
3433 }
|
|
3434
|
|
3435 /* The basic words are always stored with folded case. */
|
|
3436 dwlen = STRLEN(dw->dw_word);
|
|
3437 (void)str_foldcase(dw->dw_word, dwlen, foldword, MAXLINELEN);
|
|
3438 flags = captype(dw->dw_word, dw->dw_word + dwlen);
|
|
3439
|
|
3440 /* Check for non-word characters before the word. */
|
|
3441 clen = 0;
|
|
3442 leadlen = 0;
|
|
3443 if (!spell_iswordc(foldword))
|
|
3444 {
|
|
3445 p = foldword;
|
|
3446 for (;;)
|
|
3447 {
|
|
3448 mb_ptr_adv(p);
|
|
3449 ++clen;
|
|
3450 if (*p == NUL) /* Only non-word chars (bad word!) */
|
|
3451 {
|
|
3452 if (p_verbose > 0)
|
|
3453 smsg((char_u *)_("Warning: word without word characters: \"%s\""),
|
|
3454 foldword);
|
|
3455 break;
|
|
3456 }
|
|
3457 if (spell_iswordc(p))
|
|
3458 {
|
|
3459 /* Move the leader to "leadstring" and remove it from
|
|
3460 * "foldword". */
|
|
3461 leadlen = p - foldword;
|
|
3462 mch_memmove(leadstring, foldword, leadlen);
|
|
3463 leadstring[leadlen] = NUL;
|
|
3464 mch_memmove(foldword, p, STRLEN(p) + 1);
|
|
3465 break;
|
|
3466 }
|
|
3467 }
|
|
3468 }
|
|
3469
|
|
3470 /* Check for non-word characters after word characters. */
|
|
3471 addlen = 0;
|
|
3472 for (p = foldword; spell_iswordc(p); mb_ptr_adv(p))
|
|
3473 {
|
|
3474 if (*p == NUL)
|
|
3475 break;
|
|
3476 ++clen;
|
|
3477 }
|
|
3478 if (*p != NUL)
|
|
3479 {
|
|
3480 /* Move the addition to "addstring" and truncate "foldword". */
|
|
3481 if (flags & BWF_KEEPCAP)
|
|
3482 {
|
|
3483 /* Preserve caps, need to skip the right number of
|
|
3484 * characters in the original word (case folding may
|
|
3485 * change the byte count). */
|
|
3486 l = 0;
|
|
3487 for (cp = dw->dw_word; l < clen; mb_ptr_adv(cp))
|
|
3488 ++l;
|
|
3489 addlen = STRLEN(cp);
|
|
3490 mch_memmove(addstring, cp, addlen + 1);
|
|
3491 }
|
|
3492 else
|
|
3493 {
|
|
3494 addlen = STRLEN(p);
|
|
3495 mch_memmove(addstring, p, addlen + 1);
|
|
3496 }
|
|
3497 *p = NUL;
|
|
3498 }
|
|
3499
|
|
3500 bw = (basicword_T *)alloc_clear((unsigned)sizeof(basicword_T)
|
|
3501 + STRLEN(foldword));
|
|
3502 if (bw == NULL)
|
|
3503 break;
|
|
3504 STRCPY(bw->bw_word, foldword);
|
|
3505 bw->bw_region = regionmask;
|
|
3506
|
|
3507 if (leadlen > 0)
|
|
3508 bw->bw_leadstring = vim_strsave(leadstring);
|
|
3509 else
|
|
3510 bw->bw_leadstring = NULL;
|
|
3511 if (addlen > 0)
|
|
3512 bw->bw_addstring = vim_strsave(addstring);
|
|
3513 else
|
|
3514 bw->bw_addstring = NULL;
|
|
3515
|
|
3516 add_to_wordlist(newwords, bw);
|
|
3517
|
|
3518 if (flags & BWF_KEEPCAP)
|
|
3519 {
|
|
3520 if (addlen == 0)
|
|
3521 /* use the whole word */
|
|
3522 bw->bw_caseword = vim_strsave(dw->dw_word + leadlen);
|
|
3523 else
|
|
3524 /* use only up to the addition */
|
|
3525 bw->bw_caseword = vim_strnsave(dw->dw_word + leadlen,
|
|
3526 cp - dw->dw_word - leadlen);
|
|
3527 if (bw->bw_caseword == NULL) /* out of memory */
|
|
3528 flags &= ~BWF_KEEPCAP;
|
|
3529 }
|
|
3530 bw->bw_flags = flags;
|
|
3531
|
|
3532 /* Deal with any affix names on the old word, translate them
|
|
3533 * into affix numbers. */
|
|
3534 ga_init2(&bw->bw_prefix, sizeof(short_u), 10);
|
|
3535 ga_init2(&bw->bw_suffix, sizeof(short_u), 10);
|
|
3536 if (dw->dw_affnm != NULL)
|
|
3537 trans_affixes(dw, bw, oldaff, newwords);
|
|
3538 }
|
|
3539 }
|
|
3540 if (todo > 0)
|
|
3541 return FAIL;
|
|
3542 return OK;
|
|
3543 }
|
|
3544
|
|
3545 /*
|
|
3546 * Go through the list of words and combine the ones that are identical except
|
|
3547 * for the region.
|
|
3548 */
|
|
3549 static void
|
|
3550 combine_regions(newwords)
|
|
3551 hashtab_T *newwords;
|
|
3552 {
|
|
3553 int todo;
|
|
3554 hashitem_T *hi;
|
|
3555 basicword_T *bw, *nbw, *pbw;
|
|
3556
|
|
3557 /* Loop over all basic words in the words table. */
|
|
3558 todo = newwords->ht_used;
|
|
3559 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
3560 {
|
|
3561 if (!HASHITEM_EMPTY(hi))
|
|
3562 {
|
|
3563 --todo;
|
|
3564
|
|
3565 /* Loop over the list of words for this basic word. Compare with
|
|
3566 * each following word in the same list. */
|
|
3567 for (bw = HI2BW(hi); bw != NULL; bw = bw->bw_next)
|
|
3568 {
|
|
3569 pbw = bw;
|
|
3570 for (nbw = pbw->bw_next; nbw != NULL; nbw = pbw->bw_next)
|
|
3571 {
|
|
3572 if (bw->bw_flags == nbw->bw_flags
|
|
3573 && (bw->bw_leadstring == NULL)
|
|
3574 == (nbw->bw_leadstring == NULL)
|
|
3575 && (bw->bw_addstring == NULL)
|
|
3576 == (nbw->bw_addstring == NULL)
|
|
3577 && ((bw->bw_flags & BWF_KEEPCAP) == 0
|
|
3578 || (STRCMP(bw->bw_caseword,
|
|
3579 nbw->bw_caseword) == 0))
|
|
3580 && (bw->bw_leadstring == NULL
|
|
3581 || (STRCMP(bw->bw_leadstring,
|
|
3582 nbw->bw_leadstring) == 0))
|
|
3583 && (bw->bw_addstring == NULL
|
|
3584 || (STRCMP(bw->bw_addstring,
|
|
3585 nbw->bw_addstring) == 0))
|
|
3586 && same_affixes(bw, nbw)
|
|
3587 )
|
|
3588 {
|
|
3589 /* Match, combine regions and delete "nbw". */
|
|
3590 pbw->bw_next = nbw->bw_next;
|
|
3591 bw->bw_region |= nbw->bw_region;
|
|
3592 free_basicword(nbw);
|
|
3593 }
|
|
3594 else
|
|
3595 /* No match, continue with next one. */
|
|
3596 pbw = nbw;
|
|
3597 }
|
|
3598 }
|
|
3599 }
|
|
3600 }
|
|
3601 }
|
|
3602
|
|
3603 /*
|
|
3604 * Return TRUE when the prefixes and suffixes for "bw" and "nbw" are equal.
|
|
3605 */
|
|
3606 static int
|
|
3607 same_affixes(bw, nbw)
|
|
3608 basicword_T *bw;
|
|
3609 basicword_T *nbw;
|
|
3610 {
|
|
3611 return (bw->bw_prefix.ga_len == nbw->bw_prefix.ga_len
|
|
3612 && bw->bw_suffix.ga_len == nbw->bw_suffix.ga_len
|
|
3613 && (bw->bw_prefix.ga_len == 0
|
|
3614 || vim_memcmp(bw->bw_prefix.ga_data,
|
|
3615 nbw->bw_prefix.ga_data,
|
|
3616 bw->bw_prefix.ga_len * sizeof(short_u)) == 0)
|
|
3617 && (bw->bw_suffix.ga_len == 0
|
|
3618 || vim_memcmp(bw->bw_suffix.ga_data,
|
|
3619 nbw->bw_suffix.ga_data,
|
|
3620 bw->bw_suffix.ga_len * sizeof(short_u)) == 0));
|
|
3621 }
|
|
3622
|
|
3623 /*
|
243
|
3624 * For each basic word with additions turn the suffixes into other additions
|
|
3625 * and/or new basic words. For each basic word with a leadstring turn the
|
|
3626 * prefixes into other leadstrings and/or new basic words.
|
|
3627 * The result is that no affixes apply to the additions or leadstring of a
|
|
3628 * word.
|
|
3629 * This is also needed when a word with an addition has a prefix and the word
|
|
3630 * with prefix also exists. E.g., "blurp's/D" (D is prefix "de") and
|
|
3631 * "deblurp". "deblurp" would match and no prefix would be tried.
|
236
|
3632 */
|
|
3633 static void
|
|
3634 expand_affixes(newwords, prefgap, suffgap)
|
|
3635 hashtab_T *newwords;
|
|
3636 garray_T *prefgap;
|
|
3637 garray_T *suffgap;
|
|
3638 {
|
|
3639 int todo;
|
|
3640 hashitem_T *hi;
|
|
3641 basicword_T *bw;
|
|
3642 int pi, si;
|
|
3643 affentry_T *pae, *sae;
|
|
3644 garray_T add_words;
|
|
3645 int n;
|
240
|
3646 char_u message[MAXLINELEN + MAXWLEN];
|
236
|
3647
|
|
3648 ga_init2(&add_words, sizeof(basicword_T *), 10);
|
|
3649
|
|
3650 todo = newwords->ht_used;
|
|
3651 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
3652 {
|
|
3653 if (!HASHITEM_EMPTY(hi))
|
|
3654 {
|
|
3655 --todo;
|
240
|
3656
|
|
3657 /* This takes time, print a message now and then. */
|
|
3658 if ((todo & 0x3ff) == 0 || todo == (int)newwords->ht_used - 1)
|
|
3659 {
|
|
3660 sprintf((char *)message, _("%6d todo - %s"),
|
|
3661 todo, HI2BW(hi)->bw_word);
|
|
3662 msg_start();
|
|
3663 msg_outtrans_attr(message, 0);
|
|
3664 msg_clr_eos();
|
|
3665 msg_didout = FALSE;
|
|
3666 msg_col = 0;
|
|
3667 out_flush();
|
|
3668 ui_breakcheck();
|
|
3669 if (got_int)
|
|
3670 break;
|
|
3671 }
|
|
3672
|
236
|
3673 for (bw = HI2BW(hi); bw != NULL; bw = bw->bw_next)
|
|
3674 {
|
|
3675 /*
|
|
3676 * Need to fix affixes if there is a leader or addition and
|
|
3677 * there are prefixes or suffixes.
|
|
3678 */
|
|
3679 if ((bw->bw_leadstring != NULL || bw->bw_addstring != NULL)
|
|
3680 && (bw->bw_prefix.ga_len != 0
|
|
3681 || bw->bw_suffix.ga_len != 0))
|
|
3682 {
|
|
3683 /* Loop over all prefix numbers, but first without a
|
|
3684 * prefix. */
|
|
3685 for (pi = -1; pi < bw->bw_prefix.ga_len; ++pi)
|
|
3686 {
|
|
3687 pae = NULL;
|
|
3688 if (pi >= 0)
|
|
3689 {
|
|
3690 n = ((short_u *)bw->bw_prefix.ga_data)[pi];
|
|
3691 pae = ((affheader_T *)prefgap->ga_data + n)
|
|
3692 ->ah_first;
|
|
3693 }
|
|
3694
|
|
3695 /* Loop over all entries for prefix "pi". Do it once
|
|
3696 * when there is no prefix (pi == -1). */
|
|
3697 do
|
|
3698 {
|
243
|
3699 /* Skip prewords, they don't need to be expanded. */
|
|
3700 if (pae == NULL || !pae->ae_preword)
|
|
3701 {
|
236
|
3702 /* Loop over all suffix numbers. Do without a
|
|
3703 * suffix first when there is a prefix. */
|
|
3704 for (si = (pi == -1 ? 0 : -1);
|
|
3705 si < bw->bw_suffix.ga_len; ++si)
|
|
3706 {
|
|
3707 sae = NULL;
|
|
3708 if (si >= 0)
|
|
3709 {
|
|
3710 n = ((short_u *)bw->bw_suffix.ga_data)[si];
|
|
3711 sae = ((affheader_T *)suffgap->ga_data + n)
|
|
3712 ->ah_first;
|
|
3713 }
|
|
3714
|
|
3715 /* Loop over all entries for suffix "si". Do
|
|
3716 * it once when there is no suffix (si == -1).
|
|
3717 */
|
|
3718 do
|
|
3719 {
|
|
3720 /* Expand the word for this combination of
|
|
3721 * prefixes and affixes. */
|
|
3722 expand_one_aff(bw, &add_words, pae, sae);
|
|
3723
|
|
3724 /* Advance to next suffix entry, if there
|
|
3725 * is one. */
|
|
3726 if (sae != NULL)
|
|
3727 sae = sae->ae_next;
|
|
3728 } while (sae != NULL);
|
|
3729 }
|
243
|
3730 }
|
|
3731
|
|
3732 /* Advance to next prefix entry, if there is one. */
|
|
3733 if (pae != NULL)
|
|
3734 pae = pae->ae_next;
|
236
|
3735 } while (pae != NULL);
|
|
3736 }
|
|
3737 }
|
|
3738 }
|
|
3739 }
|
|
3740 }
|
|
3741
|
|
3742 /*
|
|
3743 * Add the new words afterwards, can't change "newwords" while going over
|
|
3744 * all its items.
|
|
3745 */
|
|
3746 for (pi = 0; pi < add_words.ga_len; ++pi)
|
|
3747 add_to_wordlist(newwords, ((basicword_T **)add_words.ga_data)[pi]);
|
|
3748
|
|
3749 ga_clear(&add_words);
|
|
3750 }
|
|
3751
|
|
3752 /*
|
|
3753 * Add one word to "add_words" for basic word "bw" with additions, adding
|
|
3754 * prefix "pae" and suffix "sae". Either "pae" or "sae" can be NULL.
|
243
|
3755 * Don't do this when not necessary:
|
|
3756 * - no leadstring and adding prefix doesn't result in existing word.
|
236
|
3757 */
|
|
3758 static void
|
|
3759 expand_one_aff(bw, add_words, pae, sae)
|
|
3760 basicword_T *bw;
|
|
3761 garray_T *add_words;
|
|
3762 affentry_T *pae;
|
|
3763 affentry_T *sae;
|
|
3764 {
|
|
3765 char_u word[MAXWLEN + 1];
|
|
3766 char_u caseword[MAXWLEN + 1];
|
|
3767 int l = 0;
|
|
3768 int choplen = 0;
|
|
3769 int ll;
|
|
3770 basicword_T *nbw;
|
|
3771
|
|
3772 /* Prepend prefix to the basic word if there is a prefix and there is no
|
|
3773 * leadstring. */
|
|
3774 if (pae != NULL && bw->bw_leadstring == NULL)
|
|
3775 {
|
|
3776 if (pae->ae_add != NULL)
|
|
3777 {
|
|
3778 l = STRLEN(pae->ae_add);
|
|
3779 mch_memmove(word, pae->ae_add, l);
|
|
3780 }
|
|
3781 if (pae->ae_chop != NULL)
|
|
3782 choplen = STRLEN(pae->ae_chop);
|
|
3783 }
|
|
3784
|
|
3785 /* Copy the body of the word. */
|
|
3786 STRCPY(word + l, bw->bw_word + choplen);
|
|
3787
|
|
3788 /* Do the same for bw_caseword, if it's there. */
|
|
3789 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3790 {
|
|
3791 if (l > 0)
|
|
3792 mch_memmove(caseword, pae->ae_add, l);
|
|
3793 STRCPY(caseword + l, bw->bw_caseword + choplen);
|
|
3794 }
|
|
3795
|
|
3796 /* Append suffix to the basic word if there is a suffix and there is no
|
|
3797 * addstring. */
|
|
3798 if (sae != 0 && bw->bw_addstring == NULL)
|
|
3799 {
|
|
3800 l = STRLEN(word);
|
|
3801 if (sae->ae_chop != NULL)
|
|
3802 l -= STRLEN(sae->ae_chop);
|
|
3803 if (sae->ae_add == NULL)
|
|
3804 word[l] = NUL;
|
|
3805 else
|
|
3806 STRCPY(word + l, sae->ae_add);
|
|
3807
|
|
3808 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3809 {
|
|
3810 /* Do the same for the caseword. */
|
|
3811 l = STRLEN(caseword);
|
|
3812 if (sae->ae_chop != NULL)
|
|
3813 l -= STRLEN(sae->ae_chop);
|
|
3814 if (sae->ae_add == NULL)
|
|
3815 caseword[l] = NUL;
|
|
3816 else
|
|
3817 STRCPY(caseword + l, sae->ae_add);
|
|
3818 }
|
|
3819 }
|
|
3820
|
|
3821 nbw = (basicword_T *)alloc_clear((unsigned)
|
|
3822 sizeof(basicword_T) + STRLEN(word));
|
|
3823 if (nbw != NULL)
|
|
3824 {
|
|
3825 /* Add the new word to the list of words to be added later. */
|
|
3826 if (ga_grow(add_words, 1) == FAIL)
|
|
3827 {
|
|
3828 vim_free(nbw);
|
|
3829 return;
|
|
3830 }
|
|
3831 ((basicword_T **)add_words->ga_data)[add_words->ga_len++] = nbw;
|
|
3832
|
|
3833 /* Copy the (modified) basic word, flags and region. */
|
|
3834 STRCPY(nbw->bw_word, word);
|
|
3835 nbw->bw_flags = bw->bw_flags;
|
|
3836 nbw->bw_region = bw->bw_region;
|
|
3837
|
|
3838 /* Set the (modified) caseword. */
|
|
3839 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3840 if ((nbw->bw_caseword = vim_strsave(caseword)) == NULL)
|
|
3841 nbw->bw_flags &= ~BWF_KEEPCAP;
|
|
3842
|
|
3843 if (bw->bw_leadstring != NULL)
|
|
3844 {
|
|
3845 if (pae != NULL)
|
|
3846 {
|
|
3847 /* Prepend prefix to leadstring. */
|
|
3848 ll = STRLEN(bw->bw_leadstring);
|
|
3849 l = choplen = 0;
|
|
3850 if (pae->ae_add != NULL)
|
|
3851 l = STRLEN(pae->ae_add);
|
|
3852 if (pae->ae_chop != NULL)
|
|
3853 {
|
|
3854 choplen = STRLEN(pae->ae_chop);
|
|
3855 if (choplen > ll) /* TODO: error? */
|
|
3856 choplen = ll;
|
|
3857 }
|
|
3858 nbw->bw_leadstring = alloc((unsigned)(ll + l - choplen + 1));
|
|
3859 if (nbw->bw_leadstring != NULL)
|
|
3860 {
|
|
3861 if (l > 0)
|
|
3862 mch_memmove(nbw->bw_leadstring, pae->ae_add, l);
|
|
3863 STRCPY(nbw->bw_leadstring + l, bw->bw_leadstring + choplen);
|
|
3864 }
|
|
3865 }
|
|
3866 else
|
|
3867 nbw->bw_leadstring = vim_strsave(bw->bw_leadstring);
|
|
3868 }
|
243
|
3869 else if (bw->bw_prefix.ga_len > 0)
|
|
3870 {
|
|
3871 /* There is no leadstring, copy the list of possible prefixes. */
|
|
3872 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3873 if (ga_grow(&nbw->bw_prefix, bw->bw_prefix.ga_len) == OK)
|
|
3874 {
|
|
3875 mch_memmove(nbw->bw_prefix.ga_data, bw->bw_prefix.ga_data,
|
|
3876 bw->bw_prefix.ga_len * sizeof(short_u));
|
|
3877 nbw->bw_prefix.ga_len = bw->bw_prefix.ga_len;
|
|
3878 }
|
|
3879 }
|
236
|
3880
|
|
3881 if (bw->bw_addstring != NULL)
|
|
3882 {
|
|
3883 if (sae != NULL)
|
|
3884 {
|
|
3885 /* Append suffix to addstring. */
|
|
3886 l = STRLEN(bw->bw_addstring);
|
|
3887 if (sae->ae_chop != NULL)
|
|
3888 {
|
|
3889 l -= STRLEN(sae->ae_chop);
|
|
3890 if (l < 0) /* TODO: error? */
|
|
3891 l = 0;
|
|
3892 }
|
|
3893 if (sae->ae_add == NULL)
|
|
3894 ll = 0;
|
|
3895 else
|
|
3896 ll = STRLEN(sae->ae_add);
|
|
3897 nbw->bw_addstring = alloc((unsigned)(ll + l - choplen + 1));
|
|
3898 if (nbw->bw_addstring != NULL)
|
|
3899 {
|
|
3900 STRCPY(nbw->bw_addstring, bw->bw_addstring);
|
|
3901 if (sae->ae_add == NULL)
|
|
3902 nbw->bw_addstring[l] = NUL;
|
|
3903 else
|
|
3904 STRCPY(nbw->bw_addstring + l, sae->ae_add);
|
|
3905 }
|
|
3906 }
|
|
3907 else
|
|
3908 nbw->bw_addstring = vim_strsave(bw->bw_addstring);
|
|
3909 }
|
|
3910 }
|
|
3911 }
|
|
3912
|
|
3913 /*
|
|
3914 * Add basicword_T "*bw" to wordlist "newwords".
|
|
3915 */
|
|
3916 static void
|
|
3917 add_to_wordlist(newwords, bw)
|
|
3918 hashtab_T *newwords;
|
|
3919 basicword_T *bw;
|
|
3920 {
|
|
3921 hashitem_T *hi;
|
|
3922 basicword_T *bw2;
|
|
3923
|
|
3924 hi = hash_find(newwords, bw->bw_word);
|
|
3925 if (HASHITEM_EMPTY(hi))
|
|
3926 {
|
|
3927 /* New entry, add to hashlist. */
|
|
3928 hash_add(newwords, bw->bw_word);
|
|
3929 bw->bw_next = NULL;
|
|
3930 }
|
|
3931 else
|
|
3932 {
|
|
3933 /* Existing entry, append to list of basic words. */
|
|
3934 bw2 = HI2BW(hi);
|
|
3935 bw->bw_next = bw2->bw_next;
|
|
3936 bw2->bw_next = bw;
|
|
3937 }
|
|
3938 }
|
|
3939
|
|
3940 /*
|
|
3941 * Write a number to file "fd", MSB first, in "len" bytes.
|
|
3942 */
|
|
3943 static void
|
|
3944 put_bytes(fd, nr, len)
|
|
3945 FILE *fd;
|
|
3946 long_u nr;
|
|
3947 int len;
|
|
3948 {
|
|
3949 int i;
|
|
3950
|
|
3951 for (i = len - 1; i >= 0; --i)
|
|
3952 putc((int)(nr >> (i * 8)), fd);
|
|
3953 }
|
|
3954
|
|
3955 /*
|
243
|
3956 * Write affix info. <affitemcnt> <affitem> ...
|
236
|
3957 */
|
|
3958 static void
|
|
3959 write_affix(fd, ah)
|
|
3960 FILE *fd;
|
|
3961 affheader_T *ah;
|
|
3962 {
|
|
3963 int i = 0;
|
|
3964 affentry_T *ae;
|
|
3965 char_u *p;
|
|
3966 int round;
|
243
|
3967 int flags;
|
236
|
3968
|
|
3969 /* Count the number of entries. */
|
|
3970 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
|
3971 ++i;
|
|
3972 put_bytes(fd, (long_u)i, 2); /* <affitemcnt> */
|
|
3973
|
243
|
3974 /* <affitem>: <affflags> <affchoplen> <affchop> <affaddlen> <affadd> */
|
236
|
3975 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
243
|
3976 {
|
|
3977 flags = ah->ah_combine ? AFF_COMBINE : 0;
|
|
3978 if (ae->ae_preword)
|
|
3979 flags |= AFF_PREWORD;
|
|
3980 fputc(flags, fd); /* <affflags> */
|
|
3981
|
236
|
3982 for (round = 1; round <= 2; ++round)
|
|
3983 {
|
|
3984 p = round == 1 ? ae->ae_chop : ae->ae_add;
|
|
3985 if (p == NULL)
|
|
3986 putc(0, fd); /* <affchoplen> / <affaddlen> */
|
|
3987 else
|
|
3988 {
|
|
3989 putc(STRLEN(p), fd); /* <affchoplen> / <affaddlen> */
|
|
3990 /* <affchop> / <affadd> */
|
|
3991 fwrite(p, STRLEN(p), (size_t)1, fd);
|
|
3992 }
|
|
3993 }
|
243
|
3994 }
|
236
|
3995 }
|
|
3996
|
|
3997 /*
|
|
3998 * Write list of affix NRs: <affixcnt> <affixNR> ...
|
|
3999 */
|
|
4000 static void
|
|
4001 write_affixlist(fd, aff, bytes)
|
|
4002 FILE *fd;
|
|
4003 garray_T *aff;
|
|
4004 int bytes;
|
|
4005 {
|
|
4006 int i;
|
|
4007
|
|
4008 if (aff->ga_len > 0)
|
|
4009 {
|
|
4010 putc(aff->ga_len, fd); /* <affixcnt> */
|
|
4011 for (i = 0; i < aff->ga_len; ++i)
|
|
4012 put_bytes(fd, (long_u )((short_u *)aff->ga_data)[i], bytes);
|
|
4013 }
|
|
4014 }
|
|
4015
|
|
4016 /*
|
|
4017 * Vim spell file format: <HEADER> <PREFIXLIST> <SUFFIXLIST>
|
|
4018 * <SUGGEST> <WORDLIST>
|
|
4019 *
|
|
4020 * <HEADER>: <fileID> <regioncnt> <regionname> ...
|
|
4021 *
|
243
|
4022 * <fileID> 10 bytes "VIMspell02"
|
236
|
4023 * <regioncnt> 1 byte number of regions following (8 supported)
|
|
4024 * <regionname> 2 bytes Region name: ca, au, etc.
|
|
4025 * First <regionname> is region 1.
|
|
4026 *
|
|
4027 *
|
|
4028 * <PREFIXLIST>: <affcount> <afftotcnt> <affix> ...
|
|
4029 * <SUFFIXLIST>: <affcount> <afftotcnt> <affix> ...
|
|
4030 * list of possible affixes: prefixes and suffixes.
|
|
4031 *
|
|
4032 * <affcount> 2 bytes Number of affixes (MSB comes first).
|
|
4033 * When more than 256 an affixNR is 2 bytes.
|
|
4034 * This is separate for prefixes and suffixes!
|
|
4035 * First affixNR is 0.
|
|
4036 * <afftotcnt> 2 bytes Total number of affix items (MSB comes first).
|
|
4037 *
|
243
|
4038 * <affix>: <affitemcnt> <affitem> ...
|
236
|
4039 *
|
|
4040 * <affitemcnt> 2 bytes Number of affixes with this affixNR (MSB first).
|
|
4041 *
|
243
|
4042 * <affitem>: <affflags> <affchoplen> <affchop> <affaddlen> <affadd>
|
236
|
4043 *
|
243
|
4044 * <affflags> 1 byte 0x01: prefix combines with suffix, AFF_COMBINE
|
|
4045 * 0x02: prefix includes word, AFF_PREWORD
|
|
4046 * 0x04-0x80: unset
|
236
|
4047 * <affchoplen> 1 byte Length of <affchop> in bytes.
|
|
4048 * <affchop> N bytes To be removed from basic word.
|
|
4049 * <affaddlen> 1 byte Length of <affadd> in bytes.
|
|
4050 * <affadd> N bytes To be added to basic word.
|
|
4051 *
|
|
4052 *
|
|
4053 * <SUGGEST> : <suggestlen> <more> ...
|
|
4054 *
|
|
4055 * <suggestlen> 4 bytes Length of <SUGGEST> in bytes, excluding
|
|
4056 * <suggestlen>. MSB first.
|
|
4057 * <more> To be defined.
|
|
4058 *
|
|
4059 *
|
|
4060 * <WORDLIST>: <wordcount> <worditem> ...
|
|
4061 *
|
|
4062 * <wordcount> 4 bytes Number of <worditem> following. MSB first.
|
|
4063 *
|
|
4064 * <worditem>: <nr> <string> <flags> [<flags2>]
|
|
4065 * [<caselen> <caseword>]
|
|
4066 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
4067 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
4068 * [<region>]
|
|
4069 * [<addcnt> <add> ...]
|
|
4070 *
|
|
4071 * <nr> i 1 byte Number of bytes copied from previous word.
|
|
4072 * <string> N bytes Additional bytes for word, up to byte smaller than
|
|
4073 * 0x20 (space).
|
|
4074 * Must only contain case-folded word characters.
|
243
|
4075 * <flags> 1 byte 0x01: word is valid without addition, BWF_VALID
|
|
4076 * 0x02: has region byte, BWF_REGION
|
|
4077 * 0x04: first letter must be upper-case, BWF_ONECAP
|
236
|
4078 * 0x08: has suffixes, <affixcnt> and <affixNR> follow
|
243
|
4079 * BWF_SUFFIX
|
|
4080 * 0x10: more flags, <flags2> follows next, BWF_SECOND
|
236
|
4081 * 0x20-0x80: can't be used, unset
|
243
|
4082 * <flags2> 1 byte 0x01: has additions, <addcnt> and <add> follow,
|
|
4083 * BWF_ADDS
|
236
|
4084 * 0x02: has prefixes, <affixcnt> and <affixNR> follow
|
243
|
4085 * BWF_PREFIX
|
|
4086 * 0x04: all letters must be upper-case, BWF_ALLCAP
|
|
4087 * 0x08: case must match, BWF_KEEPCAP
|
236
|
4088 * 0x10-0x80: unset
|
|
4089 * <caselen> 1 byte Length of <caseword>.
|
|
4090 * <caseword> N bytes Word with matching case.
|
|
4091 * <affixcnt> 1 byte Number of affix NRs following.
|
|
4092 * <affixNR> 1 or 2 byte Number of possible affix for this word.
|
|
4093 * When using 2 bytes MSB comes first.
|
|
4094 * <region> 1 byte Bitmask for regions in which word is valid. When
|
|
4095 * omitted it's valid in all regions.
|
|
4096 * Lowest bit is for region 1.
|
|
4097 * <addcnt> 2 bytes Number of <add> items following.
|
|
4098 *
|
|
4099 * <add>: <addflags> <addlen> [<leadlen> <addstring>] [<region>]
|
|
4100 *
|
243
|
4101 * <addflags> 1 byte 0x01: unset
|
|
4102 * 0x02: has region byte, ADD_REGION
|
|
4103 * 0x04: first letter must be upper-case, ADD_ONECAP
|
|
4104 * 0x08-0x20: unset
|
|
4105 * 0x40: all letters must be upper-case, ADD_ALLCAP
|
|
4106 * 0x80: fixed case, <addstring> is the whole word
|
|
4107 * with matching case, ADD_KEEPCAP.
|
236
|
4108 * <addlen> 1 byte Length of <addstring> in bytes.
|
|
4109 * <leadlen> 1 byte Number of bytes at start of <addstring> that must
|
|
4110 * come before the start of the basic word.
|
|
4111 * <addstring> N bytes Word characters, before/in/after the word.
|
|
4112 *
|
|
4113 * All text characters are in 'encoding': <affchop>, <affadd>, <string>,
|
|
4114 * <caseword>> and <addstring>.
|
|
4115 * All other fields are ASCII: <regionname>
|
|
4116 * <string> is always case-folded.
|
|
4117 */
|
|
4118
|
|
4119 /*
|
|
4120 * Write the Vim spell file "fname".
|
|
4121 */
|
|
4122 static void
|
|
4123 write_vim_spell(fname, prefga, suffga, newwords, regcount, regchars)
|
|
4124 char_u *fname;
|
|
4125 garray_T *prefga; /* prefixes, affheader_T entries */
|
|
4126 garray_T *suffga; /* suffixes, affheader_T entries */
|
|
4127 hashtab_T *newwords; /* basic words, basicword_T entries */
|
|
4128 int regcount; /* number of regions */
|
|
4129 char_u *regchars; /* region names */
|
|
4130 {
|
|
4131 FILE *fd;
|
|
4132 garray_T *gap;
|
|
4133 hashitem_T *hi;
|
|
4134 char_u **wtab;
|
|
4135 int todo;
|
|
4136 int flags, aflags;
|
240
|
4137 basicword_T *bw, *bwf, *bw2 = NULL, *prevbw = NULL;
|
236
|
4138 int regionmask; /* mask for all relevant region bits */
|
|
4139 int i;
|
|
4140 int cnt;
|
|
4141 affentry_T *ae;
|
|
4142 int round;
|
|
4143 int prefm, suffm;
|
|
4144 garray_T bwga;
|
|
4145
|
|
4146 fd = fopen((char *)fname, "w");
|
|
4147 if (fd == NULL)
|
|
4148 {
|
|
4149 EMSG2(_(e_notopen), fname);
|
|
4150 return;
|
|
4151 }
|
|
4152
|
|
4153 fwrite(VIMSPELLMAGIC, VIMSPELLMAGICL, (size_t)1, fd);
|
|
4154
|
|
4155 /* write the region names if there is more than one */
|
|
4156 if (regcount > 1)
|
|
4157 {
|
|
4158 putc(regcount, fd);
|
|
4159 fwrite(regchars, (size_t)(regcount * 2), (size_t)1, fd);
|
|
4160 regionmask = (1 << regcount) - 1;
|
|
4161 }
|
|
4162 else
|
|
4163 {
|
|
4164 putc(0, fd);
|
|
4165 regionmask = 0;
|
|
4166 }
|
|
4167
|
|
4168 /* Write the prefix and suffix lists. */
|
|
4169 for (round = 1; round <= 2; ++round)
|
|
4170 {
|
|
4171 gap = round == 1 ? prefga : suffga;
|
|
4172 put_bytes(fd, (long_u)gap->ga_len, 2); /* <affcount> */
|
|
4173
|
|
4174 /* Count the total number of affix items. */
|
|
4175 cnt = 0;
|
|
4176 for (i = 0; i < gap->ga_len; ++i)
|
|
4177 for (ae = ((affheader_T *)gap->ga_data + i)->ah_first;
|
|
4178 ae != NULL; ae = ae->ae_next)
|
|
4179 ++cnt;
|
|
4180 put_bytes(fd, (long_u)cnt, 2); /* <afftotcnt> */
|
|
4181
|
|
4182 for (i = 0; i < gap->ga_len; ++i)
|
|
4183 write_affix(fd, (affheader_T *)gap->ga_data + i);
|
|
4184 }
|
|
4185
|
|
4186 /* Number of bytes used for affix NR depends on affix count. */
|
|
4187 prefm = (prefga->ga_len > 256) ? 2 : 1;
|
|
4188 suffm = (suffga->ga_len > 256) ? 2 : 1;
|
|
4189
|
|
4190 /* Write the suggest info. TODO */
|
|
4191 put_bytes(fd, 0L, 4);
|
|
4192
|
|
4193 /*
|
|
4194 * Write the word list. <wordcount> <worditem> ...
|
|
4195 */
|
|
4196 /* number of basic words in 4 bytes */
|
|
4197 put_bytes(fd, newwords->ht_used, 4); /* <wordcount> */
|
|
4198
|
|
4199 /*
|
|
4200 * Sort the word list, so that we can reuse as many bytes as possible.
|
|
4201 */
|
|
4202 wtab = (char_u **)alloc((unsigned)(sizeof(char_u *) * newwords->ht_used));
|
|
4203 if (wtab != NULL)
|
|
4204 {
|
|
4205 /* Make a table with pointers to each word. */
|
|
4206 todo = newwords->ht_used;
|
|
4207 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
4208 if (!HASHITEM_EMPTY(hi))
|
|
4209 wtab[--todo] = hi->hi_key;
|
|
4210
|
|
4211 /* Sort. */
|
|
4212 sort_strings(wtab, (int)newwords->ht_used);
|
|
4213
|
|
4214 /* Now write each basic word to the spell file. */
|
|
4215 ga_init2(&bwga, sizeof(basicword_T *), 10);
|
240
|
4216 for (todo = 0; (long_u)todo < newwords->ht_used; ++todo)
|
236
|
4217 {
|
|
4218 bwf = KEY2BW(wtab[todo]);
|
|
4219
|
|
4220 /*
|
|
4221 * Reorder the list of basicword_T words: make a list for words
|
|
4222 * with the same case-folded word. Put them together for same
|
|
4223 * caps (ONECAP, ALLCAP and various KEEPCAP words) and same
|
|
4224 * affixes. Each list will then be put as a basic word with
|
|
4225 * additions.
|
|
4226 * This won't take much space, since the basic word is the same
|
|
4227 * every time, only its length is written.
|
|
4228 */
|
|
4229 bwga.ga_len = 0;
|
|
4230 for (bw = bwf; bw != NULL; bw = bw->bw_next)
|
|
4231 {
|
|
4232 flags = bw->bw_flags & (BWF_ONECAP | BWF_KEEPCAP | BWF_ALLCAP);
|
|
4233
|
|
4234 /* Go through the lists we found so far. Break when the case
|
|
4235 * matches. */
|
|
4236 for (i = 0; i < bwga.ga_len; ++i)
|
|
4237 {
|
|
4238 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4239 aflags = bw2->bw_flags & (BWF_ONECAP | BWF_KEEPCAP
|
|
4240 | BWF_ALLCAP);
|
|
4241 if (flags == aflags
|
|
4242 && ((flags & BWF_KEEPCAP) == 0
|
|
4243 || (STRCMP(bw->bw_caseword,
|
|
4244 bw2->bw_caseword) == 0))
|
|
4245 && same_affixes(bw, bw2))
|
|
4246 break;
|
|
4247 }
|
|
4248 if (i == bwga.ga_len)
|
|
4249 {
|
|
4250 /* No word with similar caps, make a new list. */
|
|
4251 if (ga_grow(&bwga, 1) == FAIL)
|
|
4252 break;
|
|
4253 ((basicword_T **)bwga.ga_data)[i] = bw;
|
|
4254 bw->bw_cnext = NULL;
|
|
4255 ++bwga.ga_len;
|
|
4256 }
|
|
4257 else
|
|
4258 {
|
|
4259 /* Add to list of words with similar caps. */
|
|
4260 bw->bw_cnext = bw2->bw_cnext;
|
|
4261 bw2->bw_cnext = bw;
|
|
4262 }
|
|
4263 }
|
|
4264
|
|
4265 /* Prefer the word with no caps to use as the first basic word.
|
|
4266 * At least one without KEEPCAP. */
|
|
4267 bw = NULL;
|
|
4268 for (i = 0; i < bwga.ga_len; ++i)
|
|
4269 {
|
|
4270 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4271 if (bw == NULL
|
|
4272 || (bw2->bw_flags & (BWF_ONECAP | BWF_KEEPCAP
|
|
4273 | BWF_ALLCAP)) == 0
|
|
4274 || (bw->bw_flags & BWF_KEEPCAP))
|
|
4275 bw = bw2;
|
|
4276 }
|
|
4277
|
|
4278 /* Write first basic word. If it's KEEPCAP then we need a word
|
|
4279 * without VALID flag first (makes it easier to read the list back
|
|
4280 * in). */
|
|
4281 if (bw->bw_flags & BWF_KEEPCAP)
|
|
4282 write_bword(fd, bw, TRUE, &prevbw, regionmask, prefm, suffm);
|
|
4283 write_bword(fd, bw, FALSE, &prevbw, regionmask, prefm, suffm);
|
|
4284
|
|
4285 /* Write other basic words, with different caps. */
|
|
4286 for (i = 0; i < bwga.ga_len; ++i)
|
|
4287 {
|
|
4288 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4289 if (bw2 != bw)
|
|
4290 write_bword(fd, bw2, FALSE, &prevbw, regionmask,
|
|
4291 prefm, suffm);
|
|
4292 }
|
|
4293 }
|
|
4294
|
|
4295 ga_clear(&bwga);
|
|
4296 }
|
|
4297
|
|
4298 fclose(fd);
|
|
4299 }
|
|
4300
|
|
4301 /*
|
|
4302 * Write basic word, followed by any additions.
|
|
4303 *
|
|
4304 * <worditem>: <nr> <string> <flags> [<flags2>]
|
|
4305 * [<caselen> <caseword>]
|
|
4306 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
4307 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
4308 * [<region>]
|
|
4309 * [<addcnt> <add> ...]
|
|
4310 */
|
|
4311 static void
|
|
4312 write_bword(fd, bwf, lowcap, prevbw, regionmask, prefm, suffm)
|
|
4313 FILE *fd;
|
|
4314 basicword_T *bwf;
|
|
4315 int lowcap; /* write KEEPKAP word as not-valid */
|
|
4316 basicword_T **prevbw; /* last written basic word */
|
|
4317 int regionmask; /* mask that includes all possible regions */
|
|
4318 int prefm;
|
|
4319 int suffm;
|
|
4320 {
|
|
4321 int flags;
|
|
4322 int aflags;
|
|
4323 int len;
|
|
4324 int leadlen, addlen;
|
|
4325 int clen;
|
|
4326 int adds = 0;
|
|
4327 int i;
|
|
4328 basicword_T *bw, *bw2;
|
|
4329
|
|
4330 /* Check how many bytes can be copied from the previous word. */
|
|
4331 len = STRLEN(bwf->bw_word);
|
|
4332 if (*prevbw == NULL)
|
|
4333 clen = 0;
|
|
4334 else
|
|
4335 for (clen = 0; clen < len
|
|
4336 && (*prevbw)->bw_word[clen] == bwf->bw_word[clen]; ++clen)
|
|
4337 ;
|
|
4338 putc(clen, fd); /* <nr> */
|
|
4339 *prevbw = bwf;
|
|
4340 /* <string> */
|
|
4341 if (len > clen)
|
|
4342 fwrite(bwf->bw_word + clen, (size_t)(len - clen), (size_t)1, fd);
|
|
4343
|
|
4344 /* Try to find a word without additions to use first. */
|
|
4345 bw = bwf;
|
|
4346 for (bw2 = bwf; bw2 != NULL; bw2 = bw2->bw_cnext)
|
|
4347 {
|
|
4348 if (bw2->bw_addstring != NULL || bw2->bw_leadstring != NULL)
|
|
4349 ++adds;
|
|
4350 else
|
|
4351 bw = bw2;
|
|
4352 }
|
|
4353
|
|
4354 /* Flags: If there is no leadstring and no addstring the basic word is
|
|
4355 * valid, may have prefixes, suffixes and region. */
|
|
4356 flags = bw->bw_flags;
|
|
4357 if (bw->bw_addstring == NULL && bw->bw_leadstring == NULL)
|
|
4358 {
|
|
4359 flags |= BWF_VALID;
|
|
4360
|
|
4361 /* Flags: add the region byte if the word isn't valid in all
|
|
4362 * regions. */
|
|
4363 if (regionmask != 0 && (bw->bw_region & regionmask) != regionmask)
|
|
4364 flags |= BWF_REGION;
|
|
4365 }
|
243
|
4366 /* Add the prefix/suffix list if there are prefixes/suffixes. */
|
|
4367 if (bw->bw_leadstring == NULL && bw->bw_prefix.ga_len > 0)
|
|
4368 flags |= BWF_PREFIX;
|
|
4369 if (bw->bw_addstring == NULL && bw->bw_suffix.ga_len > 0)
|
|
4370 flags |= BWF_SUFFIX;
|
236
|
4371
|
|
4372 /* Flags: may have additions. */
|
|
4373 if (adds > 0)
|
|
4374 flags |= BWF_ADDS;
|
|
4375
|
|
4376 /* The dummy word before a KEEPCAP word doesn't have any flags, they are
|
|
4377 * in the actual word that follows. */
|
|
4378 if (lowcap)
|
|
4379 flags = 0;
|
|
4380
|
|
4381 /* Flags: when the upper byte is not used we only write one flags
|
|
4382 * byte, if it's used then set an extra flag in the first byte and
|
|
4383 * also write the second byte. */
|
|
4384 if ((flags & 0xff00) == 0)
|
|
4385 putc(flags, fd); /* <flags> */
|
|
4386 else
|
|
4387 {
|
|
4388 putc(flags | BWF_SECOND, fd); /* <flags> */
|
|
4389 putc((int)((unsigned)flags >> 8), fd); /* <flags2> */
|
|
4390 }
|
|
4391
|
|
4392 /* First dummy word doesn't need anything but flags. */
|
|
4393 if (lowcap)
|
|
4394 return;
|
|
4395
|
|
4396 if (flags & BWF_KEEPCAP)
|
|
4397 {
|
|
4398 len = STRLEN(bw->bw_caseword);
|
|
4399 putc(len, fd); /* <caselen> */
|
|
4400 for (i = 0; i < len; ++i)
|
|
4401 putc(bw->bw_caseword[i], fd); /* <caseword> */
|
|
4402 }
|
|
4403
|
|
4404 /* write prefix and suffix lists: <affixcnt> <affixNR> ... */
|
|
4405 if (flags & BWF_PREFIX)
|
|
4406 write_affixlist(fd, &bw->bw_prefix, prefm);
|
|
4407 if (flags & BWF_SUFFIX)
|
|
4408 write_affixlist(fd, &bw->bw_suffix, suffm);
|
|
4409
|
|
4410 if (flags & BWF_REGION)
|
|
4411 putc(bw->bw_region, fd); /* <region> */
|
|
4412
|
|
4413 /*
|
|
4414 * Additions.
|
|
4415 */
|
|
4416 if (adds > 0)
|
|
4417 {
|
|
4418 put_bytes(fd, (long_u)adds, 2); /* <addcnt> */
|
|
4419
|
|
4420 for (bw = bwf; bw != NULL; bw = bw->bw_cnext)
|
|
4421 if (bw->bw_leadstring != NULL || bw->bw_addstring != NULL)
|
|
4422 {
|
|
4423 /* <add>: <addflags> <addlen> [<leadlen> <addstring>]
|
|
4424 * [<region>] */
|
|
4425 aflags = 0;
|
|
4426 if (bw->bw_flags & BWF_ONECAP)
|
|
4427 aflags |= ADD_ONECAP;
|
|
4428 if (bw->bw_flags & BWF_ALLCAP)
|
|
4429 aflags |= ADD_ALLCAP;
|
|
4430 if (bw->bw_flags & BWF_KEEPCAP)
|
|
4431 aflags |= ADD_KEEPCAP;
|
|
4432 if (regionmask != 0
|
|
4433 && (bw->bw_region & regionmask) != regionmask)
|
|
4434 aflags |= ADD_REGION;
|
|
4435 putc(aflags, fd); /* <addflags> */
|
|
4436
|
|
4437 if (bw->bw_leadstring == NULL)
|
|
4438 leadlen = 0;
|
|
4439 else
|
|
4440 leadlen = STRLEN(bw->bw_leadstring);
|
|
4441 if (bw->bw_addstring == NULL)
|
|
4442 addlen = 0;
|
|
4443 else
|
|
4444 addlen = STRLEN(bw->bw_addstring);
|
|
4445 putc(leadlen + addlen, fd); /* <addlen> */
|
|
4446 putc(leadlen, fd); /* <leadlen> */
|
|
4447 /* <addstring> */
|
|
4448 if (bw->bw_leadstring != NULL)
|
|
4449 fwrite(bw->bw_leadstring, (size_t)leadlen, (size_t)1, fd);
|
|
4450 if (bw->bw_addstring != NULL)
|
|
4451 fwrite(bw->bw_addstring, (size_t)addlen, (size_t)1, fd);
|
|
4452
|
|
4453 if (aflags & ADD_REGION)
|
|
4454 putc(bw->bw_region, fd); /* <region> */
|
|
4455 }
|
|
4456 }
|
|
4457 }
|
|
4458
|
|
4459
|
|
4460 /*
|
|
4461 * ":mkspell outfile infile ..."
|
|
4462 */
|
|
4463 void
|
|
4464 ex_mkspell(eap)
|
|
4465 exarg_T *eap;
|
|
4466 {
|
|
4467 int fcount;
|
|
4468 char_u **fnames;
|
|
4469 char_u fname[MAXPATHL];
|
|
4470 char_u wfname[MAXPATHL];
|
|
4471 afffile_T *(afile[8]);
|
|
4472 hashtab_T dfile[8];
|
|
4473 int i;
|
|
4474 int len;
|
|
4475 char_u region_name[16];
|
|
4476 struct stat st;
|
|
4477 int round;
|
|
4478 vimconv_T conv;
|
240
|
4479 int ascii = FALSE;
|
|
4480 char_u *arg = eap->arg;
|
|
4481
|
|
4482 if (STRNCMP(arg, "-ascii", 6) == 0)
|
|
4483 {
|
|
4484 ascii = TRUE;
|
|
4485 arg = skipwhite(arg + 6);
|
|
4486 }
|
|
4487
|
|
4488 /* Expand all the remaining arguments (e.g., $VIMRUNTIME). */
|
|
4489 if (get_arglist_exp(arg, &fcount, &fnames) == FAIL)
|
236
|
4490 return;
|
|
4491 if (fcount < 2)
|
|
4492 EMSG(_(e_invarg)); /* need at least output and input names */
|
|
4493 else if (fcount > 9)
|
|
4494 EMSG(_("E754: Only up to 8 regions supported"));
|
|
4495 else
|
|
4496 {
|
|
4497 /* Check for overwriting before doing things that may take a lot of
|
|
4498 * time. */
|
240
|
4499 sprintf((char *)wfname, "%s.%s.spl", fnames[0],
|
|
4500 ascii ? (char_u *)"ascii" : p_enc);
|
236
|
4501 if (!eap->forceit && mch_stat((char *)wfname, &st) >= 0)
|
|
4502 {
|
|
4503 EMSG(_(e_exists));
|
|
4504 goto theend;
|
|
4505 }
|
|
4506 if (mch_isdir(fnames[0]))
|
|
4507 {
|
|
4508 EMSG2(_(e_isadir2), fnames[0]);
|
|
4509 goto theend;
|
|
4510 }
|
|
4511
|
|
4512 /*
|
|
4513 * Init the aff and dic pointers.
|
|
4514 * Get the region names if there are more than 2 arguments.
|
|
4515 */
|
|
4516 for (i = 1; i < fcount; ++i)
|
|
4517 {
|
|
4518 afile[i - 1] = NULL;
|
|
4519 hash_init(&dfile[i - 1]);
|
|
4520 if (fcount > 2)
|
|
4521 {
|
|
4522 len = STRLEN(fnames[i]);
|
|
4523 if (STRLEN(gettail(fnames[i])) < 5 || fnames[i][len - 3] != '_')
|
|
4524 {
|
|
4525 EMSG2(_("E755: Invalid region in %s"), fnames[i]);
|
|
4526 goto theend;
|
|
4527 }
|
|
4528 else
|
|
4529 {
|
|
4530 region_name[(i - 1) * 2] = TOLOWER_ASC(fnames[i][len - 2]);
|
|
4531 region_name[(i - 1) * 2 + 1] =
|
|
4532 TOLOWER_ASC(fnames[i][len - 1]);
|
|
4533 }
|
|
4534 }
|
|
4535 }
|
|
4536
|
|
4537 /*
|
|
4538 * Read all the .aff and .dic files.
|
|
4539 * Text is converted to 'encoding'.
|
|
4540 */
|
|
4541 for (i = 1; i < fcount; ++i)
|
|
4542 {
|
|
4543 /* Read the .aff file. Will init "conv" based on the "SET" line. */
|
|
4544 conv.vc_type = CONV_NONE;
|
|
4545 sprintf((char *)fname, "%s.aff", fnames[i]);
|
240
|
4546 if ((afile[i - 1] = spell_read_aff(fname, &conv, ascii)) == NULL)
|
236
|
4547 break;
|
|
4548
|
|
4549 /* Read the .dic file. */
|
|
4550 sprintf((char *)fname, "%s.dic", fnames[i]);
|
240
|
4551 if (spell_read_dic(&dfile[i - 1], fname, &conv, ascii) == FAIL)
|
236
|
4552 break;
|
|
4553
|
|
4554 /* Free any conversion stuff. */
|
|
4555 convert_setup(&conv, NULL, NULL);
|
|
4556 }
|
|
4557
|
|
4558 /* Process the data when all the files could be read. */
|
|
4559 if (i == fcount)
|
|
4560 {
|
|
4561 garray_T prefga;
|
|
4562 garray_T suffga;
|
|
4563 garray_T *gap;
|
|
4564 hashtab_T newwords;
|
|
4565
|
|
4566 /*
|
|
4567 * Combine all the affixes into one new affix list. This is done
|
|
4568 * for prefixes and suffixes separately.
|
|
4569 * We need to do this for each region, try to re-use the same
|
|
4570 * affixes.
|
|
4571 * Since we number the new affix entries, a growarray will do. In
|
|
4572 * the affheader_T the ah_key is unused.
|
|
4573 */
|
|
4574 MSG(_("Combining affixes..."));
|
|
4575 out_flush();
|
|
4576 for (round = 1; round <= 2; ++round)
|
|
4577 {
|
|
4578 gap = round == 1 ? &prefga : &suffga;
|
|
4579 ga_init2(gap, sizeof(affheader_T), 50);
|
|
4580 for (i = 1; i < fcount; ++i)
|
|
4581 get_new_aff(round == 1 ? &afile[i - 1]->af_pref
|
240
|
4582 : &afile[i - 1]->af_suff,
|
|
4583 gap, round == 1);
|
236
|
4584 }
|
|
4585
|
|
4586 /*
|
|
4587 * Go over all words and:
|
|
4588 * - change the old affix names to the new affix numbers
|
|
4589 * - check the conditions
|
|
4590 * - fold case
|
|
4591 * - extract the basic word and additions.
|
|
4592 * Do this for each region.
|
|
4593 */
|
|
4594 MSG(_("Building word list..."));
|
|
4595 out_flush();
|
|
4596 hash_init(&newwords);
|
|
4597
|
|
4598 for (i = 1; i < fcount; ++i)
|
|
4599 build_wordlist(&newwords, &dfile[i - 1], afile[i - 1],
|
|
4600 1 << (i - 1));
|
|
4601
|
|
4602 if (fcount > 2)
|
|
4603 {
|
|
4604 /* Combine words for the different regions into one. */
|
|
4605 MSG(_("Combining regions..."));
|
|
4606 out_flush();
|
|
4607 combine_regions(&newwords);
|
|
4608 }
|
|
4609
|
|
4610 /*
|
|
4611 * Affixes on a word with additions are clumsy, would require
|
|
4612 * inefficient searching. Turn the affixes into additions and/or
|
|
4613 * the expanded word.
|
|
4614 */
|
|
4615 MSG(_("Processing words..."));
|
|
4616 out_flush();
|
|
4617 expand_affixes(&newwords, &prefga, &suffga);
|
|
4618
|
|
4619 /* Write the info in the spell file. */
|
|
4620 smsg((char_u *)_("Writing spell file %s..."), wfname);
|
|
4621 out_flush();
|
|
4622 write_vim_spell(wfname, &prefga, &suffga, &newwords,
|
|
4623 fcount - 1, region_name);
|
|
4624 MSG(_("Done!"));
|
|
4625 out_flush();
|
|
4626
|
|
4627 /* Free the allocated stuff. */
|
|
4628 free_wordtable(&newwords);
|
|
4629 for (round = 1; round <= 2; ++round)
|
|
4630 {
|
|
4631 gap = round == 1 ? &prefga: &suffga;
|
|
4632 for (i = 0; i < gap->ga_len; ++i)
|
|
4633 free_affixentries(((affheader_T *)gap->ga_data + i)
|
|
4634 ->ah_first);
|
|
4635 ga_clear(gap);
|
|
4636 }
|
|
4637 }
|
|
4638
|
|
4639 /* Free the .aff and .dic file structures. */
|
|
4640 for (i = 1; i < fcount; ++i)
|
|
4641 {
|
|
4642 if (afile[i - 1] != NULL)
|
|
4643 spell_free_aff(afile[i - 1]);
|
|
4644 spell_free_dic(&dfile[i - 1]);
|
|
4645 }
|
|
4646 }
|
|
4647
|
|
4648 theend:
|
|
4649 FreeWild(fcount, fnames);
|
|
4650 }
|
|
4651
|
|
4652 static void
|
|
4653 free_wordtable(ht)
|
|
4654 hashtab_T *ht;
|
|
4655 {
|
|
4656 int todo;
|
|
4657 basicword_T *bw, *nbw;
|
|
4658 hashitem_T *hi;
|
|
4659
|
|
4660 todo = ht->ht_used;
|
|
4661 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
4662 {
|
|
4663 if (!HASHITEM_EMPTY(hi))
|
|
4664 {
|
|
4665 --todo;
|
|
4666 for (bw = HI2BW(hi); bw != NULL; bw = nbw)
|
|
4667 {
|
|
4668 nbw = bw->bw_next;
|
|
4669 free_basicword(bw);
|
|
4670 }
|
|
4671 }
|
|
4672 }
|
|
4673 }
|
|
4674
|
|
4675 /*
|
|
4676 * Free a basicword_T and what it contains.
|
|
4677 */
|
|
4678 static void
|
|
4679 free_basicword(bw)
|
|
4680 basicword_T *bw;
|
|
4681 {
|
|
4682 ga_clear(&bw->bw_prefix);
|
|
4683 ga_clear(&bw->bw_suffix);
|
|
4684 vim_free(bw->bw_caseword);
|
|
4685 vim_free(bw->bw_leadstring);
|
|
4686 vim_free(bw->bw_addstring);
|
|
4687 vim_free(bw);
|
|
4688 }
|
|
4689
|
|
4690 /*
|
240
|
4691 * Free a list of affentry_T and what they contain.
|
236
|
4692 */
|
|
4693 static void
|
|
4694 free_affixentries(first)
|
|
4695 affentry_T *first;
|
|
4696 {
|
|
4697 affentry_T *ap, *an;
|
|
4698
|
|
4699 for (ap = first; ap != NULL; ap = an)
|
|
4700 {
|
|
4701 an = ap->ae_next;
|
240
|
4702 free_affix_entry(ap);
|
236
|
4703 }
|
|
4704 }
|
|
4705
|
240
|
4706 /*
|
|
4707 * Free one affentry_T and what it contains.
|
|
4708 */
|
|
4709 static void
|
|
4710 free_affix_entry(ap)
|
|
4711 affentry_T *ap;
|
|
4712 {
|
|
4713 vim_free(ap->ae_chop);
|
|
4714 vim_free(ap->ae_add);
|
|
4715 vim_free(ap->ae_cond);
|
|
4716 vim_free(ap->ae_prog);
|
|
4717 vim_free(ap);
|
|
4718 }
|
|
4719
|
236
|
4720 #endif /* FEAT_MBYTE */
|
|
4721
|
|
4722 #endif /* FEAT_SYN_HL */
|