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 */
|
249
|
206 #define BWF_ADDS_M 0x1000 /* there are more than 255 additions */
|
236
|
207
|
243
|
208 #define BWF_ADDHASH 0x8000 /* Internal: use hashtab for additions */
|
|
209
|
|
210 #define NOWC_KEY (char_u *)"x" /* hashtab key used for additions without
|
|
211 any word character */
|
|
212
|
236
|
213 /* flags used for addition in the spell file */
|
|
214 #define ADD_REGION 0x02 /* region byte follows */
|
|
215 #define ADD_ONECAP 0x04 /* first letter must be capital */
|
249
|
216 #define ADD_LEADLEN 0x10 /* there is a leadlen byte */
|
|
217 #define ADD_COPYLEN 0x20 /* there is a copylen byte */
|
236
|
218 #define ADD_ALLCAP 0x40 /* all letters must be capital (not used
|
|
219 for single-letter words) */
|
|
220 #define ADD_KEEPCAP 0x80 /* fixed case */
|
|
221
|
|
222 /* Translate ADD_ flags to BWF_ flags.
|
|
223 * (Needed to keep ADD_ flags in one byte.) */
|
|
224 #define ADD2BWF(x) (((x) & 0x0f) | (((x) & 0xf0) << 4))
|
|
225
|
249
|
226 #define VIMSPELLMAGIC "VIMspell03" /* string at start of Vim spell file */
|
236
|
227 #define VIMSPELLMAGICL 10
|
|
228
|
|
229 /*
|
|
230 * Structure to store info for word matching.
|
|
231 */
|
|
232 typedef struct matchinf_S
|
|
233 {
|
|
234 langp_T *mi_lp; /* info for language and region */
|
|
235 slang_T *mi_slang; /* info for the language */
|
243
|
236
|
|
237 /* pointers to original text to be checked */
|
236
|
238 char_u *mi_line; /* start of line containing word */
|
|
239 char_u *mi_word; /* start of word being checked */
|
|
240 char_u *mi_end; /* first non-word char after mi_word */
|
243
|
241 char_u *mi_wend; /* end of matching word (is mi_end
|
236
|
242 * or further) */
|
243
|
243 char_u *mi_fend; /* next char to be added to mi_fword */
|
|
244
|
|
245 /* case-folded text */
|
|
246 char_u mi_fword[MAXWLEN + 1]; /* mi_word case-folded */
|
|
247 int mi_fendlen; /* byte length of first word in
|
|
248 mi_fword */
|
|
249 int mi_faddlen; /* byte length of text in mi_fword
|
|
250 after first word */
|
|
251 char_u *mi_cword; /* word to check, points in mi_fword */
|
|
252 char_u *mi_awend; /* after next word, to check for
|
|
253 addition (NULL when not done yet) */
|
|
254 int mi_did_awend; /* did compute mi_awend */
|
|
255
|
|
256 /* others */
|
236
|
257 int mi_result; /* result so far: SP_BAD, SP_OK, etc. */
|
|
258 int mi_capflags; /* BWF_ONECAP BWF_ALLCAP BWF_KEEPCAP */
|
|
259 } matchinf_T;
|
|
260
|
|
261 static int word_match __ARGS((matchinf_T *mip));
|
|
262 static int check_adds __ARGS((matchinf_T *mip, fword_T *fw, int req_pref, int req_suf));
|
243
|
263 static void fill_awend __ARGS((matchinf_T *mip));
|
|
264 static void fold_addchars __ARGS((matchinf_T *mip, int addlen));
|
|
265 static int supports_affix __ARGS((int cnt, char_u *afflist, int afflistlen, int nr));
|
236
|
266 static int prefix_match __ARGS((matchinf_T *mip));
|
243
|
267 static int noprefix_match __ARGS((matchinf_T *mip, char_u *pword, char_u *cstart, affitem_T *ai));
|
236
|
268 static int suffix_match __ARGS((matchinf_T *mip));
|
|
269 static int match_caps __ARGS((int flags, char_u *caseword, matchinf_T *mip, char_u *cword, char_u *end));
|
|
270 static slang_T *slang_alloc __ARGS((char_u *lang));
|
|
271 static void slang_free __ARGS((slang_T *lp));
|
|
272 static slang_T *spell_load_lang __ARGS((char_u *lang));
|
|
273 static void spell_load_file __ARGS((char_u *fname, void *cookie));
|
|
274 static void *getroom __ARGS((slang_T *lp, int *bl_used, int len));
|
|
275 static int find_region __ARGS((char_u *rp, char_u *region));
|
|
276 static int captype __ARGS((char_u *word, char_u *end));
|
|
277
|
|
278 /*
|
|
279 * Main spell-checking function.
|
|
280 * "ptr" points to the start of a word.
|
|
281 * "*attrp" is set to the attributes for a badly spelled word. For a non-word
|
|
282 * or when it's OK it remains unchanged.
|
|
283 * This must only be called when 'spelllang' is not empty.
|
|
284 * Returns the length of the word in bytes, also when it's OK, so that the
|
|
285 * caller can skip over the word.
|
|
286 */
|
|
287 int
|
|
288 spell_check(wp, line, ptr, attrp)
|
|
289 win_T *wp; /* current window */
|
|
290 char_u *line; /* start of line where "ptr" points into */
|
|
291 char_u *ptr;
|
|
292 int *attrp;
|
|
293 {
|
|
294 matchinf_T mi; /* Most things are put in "mi" so that it can
|
|
295 be passed to functions quickly. */
|
|
296
|
|
297 /* Find the end of the word. We already know that *ptr is a word char. */
|
|
298 mi.mi_word = ptr;
|
|
299 mi.mi_end = ptr;
|
|
300 do
|
|
301 {
|
|
302 mb_ptr_adv(mi.mi_end);
|
|
303 } while (*mi.mi_end != NUL && spell_iswordc(mi.mi_end));
|
|
304
|
|
305 /* A word starting with a number is always OK. */
|
|
306 if (*ptr >= '0' && *ptr <= '9')
|
|
307 return (int)(mi.mi_end - ptr);
|
|
308
|
243
|
309 /* Make case-folded copy of the word. */
|
236
|
310 (void)str_foldcase(ptr, mi.mi_end - ptr, mi.mi_fword, MAXWLEN + 1);
|
|
311 mi.mi_cword = mi.mi_fword;
|
243
|
312 mi.mi_fendlen = STRLEN(mi.mi_fword);
|
|
313 mi.mi_faddlen = 0;
|
|
314 mi.mi_fend = mi.mi_end;
|
|
315
|
|
316 /* Check the caps type of the word. */
|
|
317 mi.mi_capflags = captype(ptr, mi.mi_end);
|
|
318
|
|
319 /* The word is bad unless we recognize it. */
|
236
|
320 mi.mi_result = SP_BAD;
|
|
321 mi.mi_wend = mi.mi_end;
|
243
|
322
|
|
323 mi.mi_awend = NULL;
|
|
324 mi.mi_did_awend = FALSE;
|
236
|
325 mi.mi_line = line;
|
|
326
|
|
327 /*
|
|
328 * Loop over the languages specified in 'spelllang'.
|
|
329 * We check them all, because a matching word may have additions that are
|
|
330 * longer than an already found matching word.
|
|
331 */
|
|
332 for (mi.mi_lp = LANGP_ENTRY(wp->w_buffer->b_langp, 0);
|
|
333 mi.mi_lp->lp_slang != NULL; ++mi.mi_lp)
|
|
334 {
|
|
335 /*
|
|
336 * Check for a matching word.
|
|
337 * If not found or wrong region try removing prefixes (and then
|
|
338 * suffixes).
|
|
339 * If still not found or wrong region try removing suffixes.
|
|
340 */
|
|
341 mi.mi_slang = mi.mi_lp->lp_slang;
|
|
342 if (!word_match(&mi) || mi.mi_result != SP_OK)
|
|
343 if (!prefix_match(&mi) || mi.mi_result != SP_OK)
|
|
344 suffix_match(&mi);
|
|
345 }
|
|
346
|
|
347 if (mi.mi_result != SP_OK)
|
|
348 {
|
|
349 if (mi.mi_result == SP_BAD)
|
|
350 *attrp = highlight_attr[HLF_SPB];
|
|
351 else if (mi.mi_result == SP_RARE)
|
|
352 *attrp = highlight_attr[HLF_SPR];
|
|
353 else
|
|
354 *attrp = highlight_attr[HLF_SPL];
|
|
355 }
|
|
356
|
|
357 return (int)(mi.mi_wend - ptr);
|
|
358 }
|
|
359
|
|
360 /*
|
243
|
361 * Check if the word "mip->mi_word" matches.
|
|
362 * "mip->mi_fword" is the same word case-folded;
|
|
363 *
|
|
364 * This checks the word as a whole and for prefixes that include a word.
|
|
365 *
|
|
366 * Note that when called mi_fword only contains the word up to mip->mi_end,
|
|
367 * but when checking additions it gets longer.
|
236
|
368 */
|
|
369 static int
|
|
370 word_match(mip)
|
|
371 matchinf_T *mip;
|
|
372 {
|
243
|
373 hash_T fhash = hash_hash(mip->mi_fword);
|
236
|
374 hashitem_T *hi;
|
|
375 fword_T *fw;
|
|
376 int valid = FALSE;
|
243
|
377 char_u *p;
|
|
378 char_u pword[MAXWLEN + 1];
|
|
379 int charlen;
|
|
380 int capflags_save;
|
|
381 affitem_T *ai;
|
|
382 char_u *cstart;
|
|
383 int addlen;
|
|
384 int n;
|
|
385 char_u *save_end;
|
|
386 int cc;
|
|
387
|
|
388 hi = hash_lookup(&mip->mi_slang->sl_words, mip->mi_fword, fhash);
|
|
389 if (!HASHITEM_EMPTY(hi))
|
|
390 {
|
|
391 /*
|
|
392 * Find a basic word for which the case of "mi_word" is correct.
|
|
393 * If it is, check additions and use the longest one.
|
|
394 */
|
|
395 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
396 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
397 mip->mi_word, mip->mi_end))
|
|
398 valid |= check_adds(mip, fw, -1, -1);
|
|
399 }
|
236
|
400
|
|
401 /*
|
243
|
402 * Try finding a matching preword for "mip->mi_word". These are
|
|
403 * prefixes that have a non-word character after a word character:
|
|
404 * "d'", "de-", "'s-", "l'de-". But not "'s".
|
|
405 * Also need to do this when a matching word was already found, because we
|
|
406 * might find a longer match this way (French: "qu" and "qu'a-t-elle").
|
236
|
407 */
|
243
|
408 cc = mip->mi_fword[mip->mi_fendlen];
|
|
409 mip->mi_fword[mip->mi_fendlen] = NUL;
|
|
410 hi = hash_lookup(&mip->mi_slang->sl_prewords, mip->mi_fword, fhash);
|
|
411 mip->mi_fword[mip->mi_fendlen] = cc;
|
|
412 if (!HASHITEM_EMPTY(hi))
|
|
413 {
|
|
414 capflags_save = mip->mi_capflags;
|
|
415
|
|
416 /* Go through the list of matching prewords. */
|
|
417 for (ai = HI2AI(hi); ai != NULL; ai = ai->ai_next)
|
|
418 {
|
|
419 /* Check that the lead string matches before the word. */
|
|
420 p = ai->ai_add + ai->ai_addlen + ai->ai_choplen + 2;
|
|
421 if (ai->ai_leadlen > 0)
|
|
422 {
|
|
423 if (mip->mi_word - mip->mi_line < ai->ai_leadlen
|
|
424 || STRNCMP(mip->mi_word - ai->ai_leadlen, p,
|
|
425 ai->ai_leadlen) != 0)
|
|
426 continue;
|
|
427 p += ai->ai_leadlen + 1; /* advance "p" to tail */
|
|
428 }
|
|
429 else
|
|
430 ++p; /* advance "p" to tail */
|
|
431
|
|
432 /* Check that the tail string matches after the word. Need
|
|
433 * to fold case first. */
|
|
434 if (ai->ai_taillen > 0)
|
|
435 {
|
|
436 if (ai->ai_taillen >= mip->mi_faddlen)
|
|
437 {
|
|
438 fold_addchars(mip, ai->ai_taillen);
|
|
439 if (ai->ai_taillen > mip->mi_faddlen)
|
|
440 continue; /* not enough chars, can't match */
|
|
441 }
|
|
442 if (STRNCMP(mip->mi_fword + mip->mi_fendlen,
|
|
443 p, ai->ai_taillen) != 0)
|
|
444 continue;
|
|
445 }
|
|
446
|
|
447 /*
|
|
448 * This preword matches. Remove the preword and check that
|
|
449 * the resulting word exits.
|
|
450 */
|
|
451
|
|
452 /* Find the place in the original word where the tail ends,
|
|
453 * needed for case checks. */
|
|
454 #ifdef FEAT_MBYTE
|
|
455 charlen = mb_charlen(p);
|
|
456 #else
|
|
457 charlen = ai->ai_taillen;
|
|
458 #endif
|
|
459 cstart = mip->mi_end;
|
|
460 for (n = 0; n < charlen; ++n)
|
|
461 mb_ptr_adv(cstart);
|
|
462
|
|
463 /* The new word starts with the chop. Then add up to the next
|
|
464 * non-word char. */
|
|
465 mch_memmove(pword, ai->ai_add + ai->ai_addlen + 1,
|
|
466 ai->ai_choplen);
|
|
467 p = mip->mi_fword + mip->mi_fendlen + ai->ai_taillen;
|
|
468 addlen = ai->ai_taillen;
|
|
469 while (spell_iswordc(p))
|
|
470 {
|
|
471 ++charlen;
|
|
472 #ifdef FEAT_MBYTE
|
|
473 addlen += (*mb_ptr2len_check)(p);
|
|
474 #else
|
|
475 ++addlen;
|
|
476 #endif
|
|
477 mb_ptr_adv(p);
|
|
478 if (addlen >= mip->mi_faddlen)
|
|
479 {
|
|
480 /* Get more folded characters in mip->mi_fword. */
|
|
481 fold_addchars(mip, addlen);
|
|
482 if (addlen >= mip->mi_faddlen)
|
|
483 break; /* not enough chars, can't match */
|
|
484 }
|
|
485 }
|
|
486 mch_memmove(pword + ai->ai_choplen,
|
|
487 mip->mi_fword + mip->mi_fendlen + ai->ai_taillen,
|
|
488 addlen - ai->ai_taillen);
|
|
489 pword[ai->ai_choplen + addlen - ai->ai_taillen] = NUL;
|
|
490
|
|
491 /* Need to set mi_end to find additions. Also set mi_fendlen
|
|
492 * and mi_faddlen. */
|
|
493 save_end = mip->mi_end;
|
|
494 while (--charlen >= 0)
|
|
495 mb_ptr_adv(mip->mi_end);
|
|
496 mip->mi_fendlen += addlen;
|
|
497 mip->mi_faddlen -= addlen;
|
|
498
|
|
499 /* Find the word "pword", caseword "cstart". */
|
|
500 n = noprefix_match(mip, pword, cstart, ai);
|
|
501 mip->mi_end = save_end;
|
|
502 mip->mi_fendlen -= addlen;
|
|
503 mip->mi_faddlen += addlen;
|
|
504 if (n)
|
|
505 valid = TRUE;
|
|
506
|
|
507 /* If we found a valid word, we still need to try other
|
|
508 * suffixes, because it may have an addition that's longer. */
|
|
509 }
|
|
510
|
|
511 mip->mi_capflags = capflags_save;
|
|
512 }
|
236
|
513
|
|
514 return valid;
|
|
515 }
|
|
516
|
|
517 /*
|
|
518 * Check a matching basic word for additions.
|
|
519 * Return TRUE if we have a valid match.
|
|
520 */
|
|
521 static int
|
|
522 check_adds(mip, fw, req_pref, req_suf)
|
|
523 matchinf_T *mip;
|
|
524 fword_T *fw;
|
|
525 int req_pref; /* required prefix nr, -1 if none */
|
|
526 int req_suf; /* required suffix nr, -1 if none */
|
|
527 {
|
|
528 int valid = FALSE;
|
|
529 addword_T *aw;
|
243
|
530 addword_T *naw = NULL;
|
236
|
531 char_u *p;
|
|
532 int addlen;
|
243
|
533 int cc;
|
|
534 hashitem_T *hi;
|
|
535 char_u *cp = NULL;
|
|
536 int n;
|
|
537
|
|
538 /* Check if required prefixes and suffixes are supported. These are on
|
|
539 * the basic word, not on each addition. */
|
|
540 if (req_pref >= 0 || req_suf >= 0)
|
|
541 {
|
|
542 /* Prefix NRs are stored just after the word in fw_word. */
|
|
543 cp = fw->fw_word + STRLEN(fw->fw_word) + 1;
|
|
544 if (req_pref >= 0 && !supports_affix(mip->mi_slang->sl_prefcnt,
|
|
545 cp, fw->fw_prefixcnt, req_pref))
|
|
546 return FALSE;
|
|
547 if (req_suf >= 0)
|
|
548 {
|
|
549 /* Suffix NRs are stored just after the Prefix NRs. */
|
|
550 if (fw->fw_prefixcnt > 0)
|
|
551 {
|
|
552 if (mip->mi_slang->sl_prefcnt > 256)
|
|
553 cp += fw->fw_prefixcnt * 2;
|
|
554 else
|
|
555 cp += fw->fw_prefixcnt;
|
|
556 }
|
|
557 if (!supports_affix(mip->mi_slang->sl_suffcnt,
|
|
558 cp, fw->fw_suffixcnt, req_suf))
|
|
559 return FALSE;
|
|
560 }
|
|
561 }
|
|
562
|
|
563 /* A word may be valid without an addition. */
|
|
564 if (fw->fw_flags & BWF_VALID)
|
236
|
565 {
|
|
566 valid = TRUE;
|
|
567 if (mip->mi_result != SP_OK)
|
|
568 {
|
|
569 if ((fw->fw_region & mip->mi_lp->lp_region) == 0)
|
|
570 mip->mi_result = SP_LOCAL;
|
|
571 else
|
|
572 mip->mi_result = SP_OK;
|
|
573 }
|
243
|
574 /* Set word end, required when matching a word after a preword. */
|
|
575 if (mip->mi_wend < mip->mi_end)
|
|
576 mip->mi_wend = mip->mi_end;
|
236
|
577 }
|
|
578
|
|
579 /*
|
|
580 * Check additions, both before and after the word.
|
|
581 * This may make the word longer, thus we also need to check
|
|
582 * when we already found a matching word.
|
243
|
583 * When the BWF_ADDHASH flag is present then fw_adds points to a hashtable
|
|
584 * for quick lookup. Otherwise it points to the list of all possible
|
|
585 * additions.
|
236
|
586 */
|
243
|
587 if (fw->fw_flags & BWF_ADDHASH)
|
236
|
588 {
|
243
|
589 /* Locate the text up to the next end-of-word. */
|
|
590 if (!mip->mi_did_awend)
|
|
591 fill_awend(mip);
|
|
592 if (mip->mi_awend == NULL)
|
|
593 return valid; /* there is no next word */
|
|
594
|
|
595 cc = *mip->mi_awend;
|
|
596 *mip->mi_awend = NUL;
|
|
597 hi = hash_find((hashtab_T *)fw->fw_adds,
|
|
598 mip->mi_fword + mip->mi_fendlen);
|
|
599 *mip->mi_awend = cc;
|
|
600 if (HASHITEM_EMPTY(hi))
|
|
601 return valid; /* no matching addition */
|
|
602 aw = HI2ADDWORD(hi);
|
|
603
|
|
604 /* Also check additions without word characters. If they are there,
|
|
605 * skip the first dummy entry. */
|
|
606 hi = hash_find((hashtab_T *)fw->fw_adds, NOWC_KEY);
|
|
607 if (!HASHITEM_EMPTY(hi))
|
|
608 naw = HI2ADDWORD(hi)->aw_next;
|
|
609 }
|
|
610 else
|
|
611 aw = fw->fw_adds;
|
|
612
|
|
613 for ( ; ; aw = aw->aw_next)
|
|
614 {
|
|
615 if (aw == NULL)
|
|
616 {
|
|
617 /* At end of list: may also try additions without word chars. */
|
|
618 if (naw == NULL)
|
|
619 break;
|
|
620 aw = naw;
|
|
621 naw = NULL;
|
|
622 }
|
|
623
|
236
|
624 if (aw->aw_leadlen > 0)
|
|
625 {
|
|
626 /* There is a leader, verify that it matches. */
|
|
627 if (aw->aw_leadlen > mip->mi_word - mip->mi_line
|
|
628 || STRNCMP(mip->mi_word - aw->aw_leadlen,
|
|
629 aw->aw_word, aw->aw_leadlen) != 0)
|
|
630 continue;
|
|
631 if (mip->mi_word - aw->aw_leadlen > mip->mi_line)
|
|
632 {
|
|
633 /* There must not be a word character just before the
|
|
634 * leader. */
|
|
635 p = mip->mi_word - aw->aw_leadlen;
|
|
636 mb_ptr_back(mip->mi_line, p);
|
|
637 if (spell_iswordc(p))
|
|
638 continue;
|
|
639 }
|
|
640 /* Leader matches. Addition is rest of "aw_word". */
|
|
641 p = aw->aw_word + aw->aw_leadlen;
|
|
642 }
|
|
643 else
|
|
644 /* No leader, use whole of "aw_word" for addition. */
|
|
645 p = aw->aw_word;
|
|
646
|
|
647 addlen = aw->aw_wordlen - aw->aw_leadlen;
|
|
648 if (addlen > 0)
|
|
649 {
|
|
650 /* Check for matching addition and no word character after it.
|
|
651 * First make sure we have enough case-folded chars to compare
|
|
652 * with. */
|
243
|
653 if (addlen >= mip->mi_faddlen)
|
|
654 fold_addchars(mip, addlen);
|
|
655
|
|
656 /* Put back the saved char, if needed. */
|
|
657 if (aw->aw_saveb != NUL)
|
236
|
658 {
|
243
|
659 cp = p + STRLEN(p);
|
|
660 *cp = aw->aw_saveb;
|
236
|
661 }
|
243
|
662 n = STRNCMP(mip->mi_fword + mip->mi_fendlen, p, addlen);
|
|
663 if (aw->aw_saveb != NUL)
|
|
664 *cp = NUL;
|
|
665
|
|
666 if (n != 0 || (mip->mi_fword[mip->mi_fendlen + addlen] != NUL
|
|
667 && spell_iswordc(mip->mi_fword + mip->mi_fendlen + addlen)))
|
236
|
668 continue;
|
|
669
|
|
670 /* Compute the length in the original word, before case folding. */
|
|
671 #ifdef FEAT_MBYTE
|
|
672 if (has_mbyte)
|
|
673 {
|
|
674 int l;
|
|
675
|
|
676 p = mip->mi_end;
|
243
|
677 for (l = 0; l < addlen; l += (*mb_ptr2len_check)(mip->mi_fword
|
|
678 + mip->mi_fendlen + l))
|
236
|
679 mb_ptr_adv(p);
|
|
680 addlen = p - mip->mi_end;
|
|
681 }
|
|
682 #endif
|
|
683
|
|
684 /* Check case of the addition. */
|
|
685 if (!match_caps(ADD2BWF(aw->aw_flags),
|
|
686 aw->aw_word + aw->aw_wordlen + 1, mip,
|
|
687 mip->mi_end, mip->mi_end + addlen))
|
|
688 continue;
|
|
689 }
|
|
690
|
|
691 /* Match! Use the new length if it's longer. */
|
|
692 if (mip->mi_wend < mip->mi_end + addlen)
|
|
693 mip->mi_wend = mip->mi_end + addlen;
|
|
694
|
|
695 valid = TRUE;
|
|
696 if (mip->mi_result != SP_OK)
|
|
697 {
|
|
698 if ((aw->aw_region & mip->mi_lp->lp_region) == 0)
|
|
699 mip->mi_result = SP_LOCAL;
|
|
700 else
|
|
701 mip->mi_result = SP_OK;
|
|
702 }
|
|
703 }
|
|
704
|
|
705 return valid;
|
|
706 }
|
|
707
|
|
708 /*
|
243
|
709 * Locate the text up to the next end-of-word after mip->mi_end.
|
|
710 */
|
|
711 static void
|
|
712 fill_awend(mip)
|
|
713 matchinf_T *mip;
|
|
714 {
|
|
715 char_u *p = mip->mi_end;
|
|
716 int addlen = 0;
|
|
717 int find_word = TRUE;
|
|
718
|
|
719 mip->mi_did_awend = TRUE;
|
|
720 if (mip->mi_faddlen == 0)
|
|
721 fold_addchars(mip, 0); /* need to fold first char */
|
|
722
|
|
723 /* 1: find_word == TRUE: skip over non-word characters after mi_end.
|
|
724 * 2: find_word == FALSE: skip over following word characters. */
|
|
725 for (p = mip->mi_fword + mip->mi_fendlen; *p != NUL; mb_ptr_adv(p))
|
|
726 {
|
|
727 if (spell_iswordc(p) == find_word)
|
|
728 {
|
|
729 if (!find_word)
|
|
730 break; /* done */
|
|
731 find_word = !find_word;
|
|
732 }
|
|
733 #ifdef FEAT_MBYTE
|
|
734 addlen += (*mb_ptr2len_check)(p);
|
|
735 #else
|
|
736 ++addlen;
|
|
737 #endif
|
|
738 if (addlen >= mip->mi_faddlen)
|
|
739 fold_addchars(mip, addlen); /* need to fold more chars */
|
|
740 }
|
|
741
|
|
742 /* If there are extra chars store the result. */
|
|
743 if (addlen != 0)
|
|
744 mip->mi_awend = p;
|
|
745 }
|
|
746
|
|
747 /*
|
|
748 * Fold enough characters of the checked text to be able to compare with an
|
|
749 * addition of length "addlen" plus one character (to be able to check the
|
|
750 * next character to be a non-word char).
|
|
751 * When there are not enough characters (end of line) mip->mi_faddlen will be
|
|
752 * smaller than "addlen".
|
|
753 */
|
|
754 static void
|
|
755 fold_addchars(mip, addlen)
|
|
756 matchinf_T *mip;
|
|
757 int addlen;
|
|
758 {
|
|
759 int l;
|
|
760 char_u *p = mip->mi_fword + mip->mi_fendlen;
|
|
761
|
|
762 while (mip->mi_faddlen <= addlen)
|
|
763 {
|
|
764 if (*mip->mi_fend == NUL) /* end of the line */
|
|
765 {
|
|
766 p[mip->mi_faddlen] = NUL;
|
|
767 break;
|
|
768 }
|
|
769 #ifdef FEAT_MBYTE
|
|
770 if (has_mbyte)
|
|
771 l = (*mb_ptr2len_check)(mip->mi_fend);
|
|
772 else
|
|
773 #endif
|
|
774 l = 1;
|
|
775 (void)str_foldcase(mip->mi_fend, l, p + mip->mi_faddlen,
|
|
776 MAXWLEN - mip->mi_fendlen - mip->mi_faddlen);
|
|
777 mip->mi_fend += l;
|
|
778 mip->mi_faddlen += STRLEN(p + mip->mi_faddlen);
|
|
779 }
|
|
780 }
|
|
781
|
|
782 /*
|
|
783 * Return TRUE if affix "nr" appears in affix list "afflist[afflistlen]".
|
236
|
784 */
|
|
785 static int
|
243
|
786 supports_affix(cnt, afflist, afflistlen, nr)
|
|
787 int cnt; /* total affix NR count */
|
|
788 char_u *afflist;
|
|
789 int afflistlen; /* affix count in "afflist" */
|
236
|
790 int nr;
|
|
791 {
|
243
|
792 char_u *pc = afflist;
|
236
|
793 int i;
|
243
|
794 int nr_msb, nr_lsb;
|
236
|
795
|
|
796 if (cnt <= 256)
|
|
797 {
|
243
|
798 /* one byte affix numbers */
|
|
799 for (i = afflistlen; --i >= 0; )
|
236
|
800 if (*pc++ == nr)
|
|
801 return TRUE;
|
|
802 }
|
|
803 else
|
|
804 {
|
243
|
805 /* two byte affix numbers, MSB first */
|
|
806 nr_msb = (unsigned)nr >> 8;
|
|
807 nr_lsb = nr & 0xff;
|
|
808 for (i = afflistlen; --i >= 0; )
|
|
809 {
|
|
810 if (*pc++ == nr_msb && *pc == nr_lsb)
|
236
|
811 return TRUE;
|
243
|
812 ++pc;
|
|
813 }
|
236
|
814 }
|
|
815 return FALSE;
|
|
816 }
|
|
817
|
|
818 /*
|
|
819 * Try finding a match for "mip->mi_cword" by removing prefixes.
|
|
820 */
|
|
821 static int
|
|
822 prefix_match(mip)
|
|
823 matchinf_T *mip;
|
|
824 {
|
|
825 int len = 0;
|
|
826 int charlen = 0;
|
|
827 int cc;
|
|
828 affitem_T *ai;
|
|
829 char_u pword[MAXWLEN + 1];
|
|
830 hashtab_T *ht;
|
|
831 hashitem_T *hi;
|
|
832 int found_valid = FALSE;
|
|
833 int cstart_charlen = 0;
|
|
834 char_u *cstart = mip->mi_word;
|
|
835 int capflags_save = mip->mi_capflags;
|
|
836
|
|
837 /*
|
|
838 * Check for prefixes with different character lengths.
|
|
839 * Start with zero length (only chop off).
|
|
840 */
|
|
841 for (charlen = 0; charlen <= mip->mi_slang->sl_preftab.ga_len; ++charlen)
|
|
842 {
|
|
843 if (charlen > 0)
|
|
844 {
|
|
845 #ifdef FEAT_MBYTE
|
|
846 if (has_mbyte)
|
243
|
847 len += (*mb_ptr2len_check)(mip->mi_cword + len);
|
236
|
848 else
|
|
849 #endif
|
|
850 len += 1;
|
|
851 }
|
|
852 if (mip->mi_cword[len] == NUL) /* end of word, no prefix possible */
|
|
853 break;
|
|
854
|
|
855 if (charlen == 0)
|
|
856 ai = mip->mi_slang->sl_prefzero;
|
|
857 else
|
|
858 {
|
|
859 /* Get pointer to hashtab for prefix of this many chars. */
|
|
860 ht = ((hashtab_T *)mip->mi_slang->sl_preftab.ga_data) + charlen - 1;
|
|
861 if (ht->ht_used == 0)
|
|
862 continue;
|
|
863
|
|
864 cc = mip->mi_cword[len];
|
|
865 mip->mi_cword[len] = NUL;
|
|
866 hi = hash_find(ht, mip->mi_cword);
|
|
867 mip->mi_cword[len] = cc;
|
|
868
|
|
869 if (HASHITEM_EMPTY(hi))
|
|
870 ai = NULL;
|
|
871 else
|
|
872 ai = HI2AI(hi);
|
|
873 }
|
|
874
|
|
875 /* Loop over all matching prefixes. */
|
|
876 for ( ; ai != NULL; ai = ai->ai_next)
|
|
877 {
|
243
|
878 /* Create the basic word from the chop string and the word after
|
|
879 * the matching add string. */
|
|
880 mch_memmove(pword, ai->ai_add + ai->ai_addlen + 1, ai->ai_choplen);
|
|
881 mch_memmove(pword + ai->ai_choplen, mip->mi_cword + ai->ai_addlen,
|
|
882 mip->mi_fendlen - ai->ai_addlen);
|
|
883 pword[mip->mi_fendlen - ai->ai_addlen] = NUL;
|
236
|
884
|
|
885 /* Adjust the word start for case checks, we only check the
|
|
886 * part after the prefix. */
|
|
887 while (cstart_charlen < charlen)
|
|
888 {
|
|
889 mb_ptr_adv(cstart);
|
|
890 ++cstart_charlen;
|
|
891 }
|
|
892
|
243
|
893 /* Find the word "pword", caseword "cstart". */
|
|
894 found_valid |= noprefix_match(mip, pword, cstart, ai);
|
|
895
|
|
896 if (found_valid && mip->mi_result == SP_OK)
|
236
|
897 {
|
243
|
898 /* Found a valid word, no need to try other suffixes. */
|
|
899 mip->mi_capflags = capflags_save;
|
|
900 return TRUE;
|
236
|
901 }
|
|
902 }
|
|
903 }
|
|
904
|
|
905 mip->mi_capflags = capflags_save;
|
|
906 return FALSE;
|
|
907 }
|
|
908
|
|
909 /*
|
243
|
910 * Check for matching word after removing a prefix.
|
|
911 * Return TRUE if found.
|
|
912 */
|
|
913 static int
|
|
914 noprefix_match(mip, pword, cstart, ai)
|
|
915 matchinf_T *mip;
|
|
916 char_u *pword; /* case-folded word */
|
|
917 char_u *cstart; /* original word after removed prefix */
|
|
918 affitem_T *ai; /* the prefix item */
|
|
919 {
|
|
920 hashitem_T *hi;
|
|
921 fword_T *fw;
|
|
922 int found_valid = FALSE;
|
|
923 char_u *word;
|
|
924 int i;
|
|
925 int fendlen;
|
|
926
|
|
927 /* Removing the prefix may change the caps, e.g. for
|
|
928 * "deAlf" removing "de" makes it ONECAP. */
|
|
929 mip->mi_capflags = captype(cstart, mip->mi_end);
|
|
930
|
|
931 /* Find the basic word. */
|
|
932 hi = hash_find(&mip->mi_slang->sl_words, pword);
|
|
933 if (!HASHITEM_EMPTY(hi))
|
|
934 {
|
|
935 /* Check if the word supports this prefix. */
|
|
936 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
937 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
938 cstart, mip->mi_end))
|
|
939 found_valid |= check_adds(mip, fw, ai->ai_nr, -1);
|
|
940
|
|
941 if (found_valid && mip->mi_result == SP_OK)
|
|
942 /* Found a valid word, no need to try other suffixes. */
|
|
943 return TRUE;
|
|
944 }
|
|
945
|
|
946 /* No matching basic word without prefix. When combining is
|
|
947 * allowed try with suffixes. */
|
|
948 if (ai->ai_flags & AFF_COMBINE)
|
|
949 {
|
|
950 /* Pass the word with prefix removed to suffix_match(). */
|
|
951 mip->mi_cword = pword;
|
|
952 word = mip->mi_word;
|
|
953 mip->mi_word = cstart;
|
|
954 fendlen = mip->mi_fendlen;
|
|
955 mip->mi_fendlen = STRLEN(pword);
|
|
956 i = suffix_match(mip);
|
|
957 mip->mi_cword = mip->mi_fword;
|
|
958 mip->mi_word = word;
|
|
959 mip->mi_fendlen = fendlen;
|
|
960 if (i)
|
|
961 return TRUE;
|
|
962 }
|
|
963
|
|
964 return FALSE;
|
|
965 }
|
|
966
|
|
967 /*
|
236
|
968 * Try finding a match for "mip->mi_cword" by removing suffixes.
|
|
969 */
|
|
970 static int
|
|
971 suffix_match(mip)
|
|
972 matchinf_T *mip;
|
|
973 {
|
|
974 char_u *sufp;
|
243
|
975 char_u *endw = mip->mi_cword + mip->mi_fendlen;
|
|
976 int endw_c = *endw;
|
236
|
977 int charlen;
|
|
978 affitem_T *ai;
|
|
979 char_u pword[MAXWLEN + 1];
|
|
980 fword_T *fw;
|
|
981 hashtab_T *ht;
|
|
982 hashitem_T *hi;
|
|
983 int tlen;
|
|
984 int cend_charlen = 0;
|
|
985 char_u *cend = mip->mi_end;
|
|
986 int found_valid = FALSE;
|
|
987 int capflags_save = mip->mi_capflags;
|
|
988
|
|
989 /*
|
|
990 * Try suffixes of different length, starting with an empty suffix (chop
|
|
991 * only, thus adds something).
|
|
992 * Stop checking if there are no suffixes with so many characters.
|
|
993 */
|
243
|
994 sufp = endw;
|
236
|
995 for (charlen = 0; charlen <= mip->mi_slang->sl_sufftab.ga_len; ++charlen)
|
|
996 {
|
|
997 /* Move the pointer to the possible suffix back one character, unless
|
|
998 * doing the first round (empty suffix). */
|
|
999 if (charlen > 0)
|
|
1000 {
|
|
1001 mb_ptr_back(mip->mi_cword, sufp);
|
|
1002 if (sufp <= mip->mi_cword) /* start of word, no suffix possible */
|
|
1003 break;
|
|
1004 }
|
|
1005
|
|
1006 if (charlen == 0)
|
|
1007 ai = mip->mi_slang->sl_suffzero;
|
|
1008 else
|
|
1009 {
|
|
1010 /* Get pointer to hashtab for suffix of this many chars. */
|
|
1011 ht = ((hashtab_T *)mip->mi_slang->sl_sufftab.ga_data) + charlen - 1;
|
|
1012 if (ht->ht_used == 0)
|
|
1013 continue;
|
|
1014
|
243
|
1015 *endw = NUL; /* truncate after possible suffix */
|
236
|
1016 hi = hash_find(ht, sufp);
|
|
1017 if (HASHITEM_EMPTY(hi))
|
|
1018 ai = NULL;
|
|
1019 else
|
|
1020 ai = HI2AI(hi);
|
243
|
1021 *endw = endw_c;
|
236
|
1022 }
|
|
1023
|
|
1024 if (ai != NULL)
|
|
1025 {
|
|
1026 /* Found a list of matching suffixes. Now check that there is one
|
|
1027 * we can use. */
|
|
1028 tlen = sufp - mip->mi_cword; /* length of word without suffix */
|
|
1029 mch_memmove(pword, mip->mi_cword, tlen);
|
|
1030
|
|
1031 for ( ; ai != NULL; ai = ai->ai_next)
|
|
1032 {
|
|
1033 /* Found a matching suffix. Create the basic word by removing
|
|
1034 * the suffix and adding the chop string. */
|
|
1035 if (ai->ai_choplen == 0)
|
|
1036 pword[tlen] = NUL;
|
|
1037 else
|
243
|
1038 mch_memmove(pword + tlen, ai->ai_add + ai->ai_addlen + 1,
|
|
1039 ai->ai_choplen + 1);
|
236
|
1040
|
|
1041 /* Find the basic word. */
|
|
1042 hi = hash_find(&mip->mi_slang->sl_words, pword);
|
|
1043 if (!HASHITEM_EMPTY(hi))
|
|
1044 {
|
|
1045 /* Adjust the end for case checks, we only check the part
|
|
1046 * before the suffix. */
|
|
1047 while (cend_charlen < charlen)
|
|
1048 {
|
|
1049 mb_ptr_back(mip->mi_word, cend);
|
|
1050 ++cend_charlen;
|
|
1051 }
|
|
1052
|
|
1053 /* Removing the suffix may change the caps, e.g. for
|
|
1054 * "UFOs" removing 's' makes it ALLCAP. */
|
|
1055 mip->mi_capflags = captype(mip->mi_word, cend);
|
|
1056
|
|
1057 /* Check if the word supports this suffix. */
|
|
1058 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
1059 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
1060 mip->mi_word, cend))
|
|
1061 found_valid |= check_adds(mip, fw, -1, ai->ai_nr);
|
|
1062
|
|
1063 if (found_valid && mip->mi_result == SP_OK)
|
|
1064 {
|
|
1065 /* Found a valid word, no need to try other suffixes. */
|
|
1066 mip->mi_capflags = capflags_save;
|
|
1067 return TRUE;
|
|
1068 }
|
|
1069 }
|
|
1070 }
|
|
1071 }
|
|
1072 }
|
|
1073
|
|
1074 mip->mi_capflags = capflags_save;
|
|
1075 return FALSE;
|
|
1076 }
|
|
1077
|
|
1078 /*
|
|
1079 * Return TRUE if case of "cword" meets the requirements of case flags
|
|
1080 * "flags".
|
|
1081 */
|
|
1082 static int
|
|
1083 match_caps(flags, caseword, mip, cword, end)
|
|
1084 int flags; /* flags required by basic word or addition */
|
|
1085 char_u *caseword; /* word with case as required */
|
|
1086 matchinf_T *mip;
|
|
1087 char_u *cword; /* word to compare against "caseword" */
|
|
1088 char_u *end; /* end of "cword" */
|
|
1089 {
|
|
1090 char_u *p;
|
|
1091 int c;
|
|
1092 int len;
|
|
1093 int capflags = mip->mi_capflags; /* flags of checked word */
|
|
1094 int past_second;
|
|
1095
|
|
1096 if ((capflags & BWF_KEEPCAP) == 0 && end > mip->mi_end)
|
|
1097 {
|
243
|
1098 /* If "end" is past "mip->mi_end" we need to adjust the caps type for
|
|
1099 * characters after the basic word. */
|
236
|
1100 #ifdef FEAT_MBYTE
|
|
1101 past_second = (mip->mi_word + (*mb_ptr2len_check)(mip->mi_word)
|
|
1102 < mip->mi_end);
|
|
1103 #else
|
|
1104 past_second = mip->mi_word + 1 < mip->mi_end;
|
|
1105 #endif
|
|
1106 for (p = mip->mi_end; p < end; )
|
|
1107 {
|
|
1108 if (!spell_iswordc(p))
|
|
1109 mb_ptr_adv(p);
|
|
1110 else
|
|
1111 {
|
|
1112 #ifdef FEAT_MBYTE
|
|
1113 if (has_mbyte)
|
|
1114 c = mb_ptr2char_adv(&p);
|
|
1115 else
|
|
1116 #endif
|
|
1117 c = *p++;
|
|
1118 if (MB_ISUPPER(c))
|
|
1119 {
|
|
1120 if (capflags == 0 || (capflags & BWF_ONECAP))
|
|
1121 {
|
|
1122 capflags = BWF_KEEPCAP; /* lU or UlU */
|
|
1123 break;
|
|
1124 }
|
|
1125 }
|
|
1126 else
|
|
1127 {
|
|
1128 if (capflags & BWF_ALLCAP)
|
|
1129 {
|
|
1130 if (past_second)
|
|
1131 {
|
|
1132 capflags = BWF_KEEPCAP; /* UUl */
|
|
1133 break;
|
|
1134 }
|
|
1135 capflags = BWF_ONECAP; /* Uu */
|
|
1136 }
|
|
1137 }
|
|
1138 past_second = TRUE;
|
|
1139 }
|
|
1140 }
|
|
1141 }
|
|
1142
|
|
1143 if (capflags == BWF_ALLCAP)
|
|
1144 return TRUE; /* All caps is always OK. */
|
|
1145
|
|
1146 if (flags & BWF_KEEPCAP)
|
|
1147 {
|
|
1148 len = STRLEN(caseword);
|
|
1149 return (len == end - cword && STRNCMP(caseword, cword, len) == 0);
|
|
1150 }
|
|
1151
|
|
1152 if (flags & BWF_ALLCAP)
|
|
1153 return FALSE; /* need ALLCAP, already checked above */
|
|
1154
|
|
1155 if (flags & BWF_ONECAP)
|
|
1156 return capflags == BWF_ONECAP;
|
|
1157
|
|
1158 return capflags != BWF_KEEPCAP; /* no case check, only KEEPCAP is bad */
|
|
1159 }
|
|
1160
|
|
1161 /*
|
|
1162 * Move to next spell error.
|
|
1163 * Return OK if found, FAIL otherwise.
|
|
1164 */
|
|
1165 int
|
|
1166 spell_move_to(dir, allwords)
|
|
1167 int dir; /* FORWARD or BACKWARD */
|
|
1168 int allwords; /* TRUE for "[s" and "]s" */
|
|
1169 {
|
249
|
1170 linenr_T lnum;
|
|
1171 pos_T found_pos;
|
236
|
1172 char_u *line;
|
|
1173 char_u *p;
|
|
1174 int wc;
|
|
1175 int nwc;
|
|
1176 int attr = 0;
|
|
1177 int len;
|
249
|
1178 int has_syntax = syntax_present(curbuf);
|
|
1179 int col;
|
|
1180 int can_spell;
|
236
|
1181
|
|
1182 if (!curwin->w_p_spell || *curwin->w_buffer->b_p_spl == NUL)
|
|
1183 {
|
|
1184 EMSG(_("E756: Spell checking not enabled"));
|
|
1185 return FAIL;
|
|
1186 }
|
|
1187
|
249
|
1188 /*
|
|
1189 * Start looking for bad word at the start of the line, because we can't
|
|
1190 * start halfway a word, we don't know where it starts or ends.
|
|
1191 *
|
|
1192 * When searching backwards, we continue in the line to find the last
|
|
1193 * bad word (in the cursor line: before the cursor).
|
|
1194 */
|
|
1195 lnum = curwin->w_cursor.lnum;
|
|
1196 found_pos.lnum = 0;
|
236
|
1197
|
|
1198 while (!got_int)
|
|
1199 {
|
249
|
1200 line = ml_get(lnum);
|
|
1201 p = line;
|
|
1202 wc = FALSE;
|
|
1203
|
236
|
1204 while (*p != NUL)
|
|
1205 {
|
|
1206 nwc = spell_iswordc(p);
|
|
1207 if (!wc && nwc)
|
|
1208 {
|
249
|
1209 /* When searching backward don't search after the cursor. */
|
|
1210 if (dir == BACKWARD
|
|
1211 && lnum == curwin->w_cursor.lnum
|
|
1212 && (colnr_T)(p - line) >= curwin->w_cursor.col)
|
|
1213 break;
|
|
1214
|
236
|
1215 /* start of word */
|
|
1216 len = spell_check(curwin, line, p, &attr);
|
249
|
1217
|
236
|
1218 if (attr != 0)
|
|
1219 {
|
249
|
1220 /* We found a bad word. Check the attribute. */
|
|
1221 /* TODO: check for syntax @Spell cluster. */
|
|
1222 if (allwords || attr == highlight_attr[HLF_SPB])
|
236
|
1223 {
|
249
|
1224 /* When searching forward only accept a bad word after
|
|
1225 * the cursor. */
|
|
1226 if (dir == BACKWARD
|
|
1227 || lnum > curwin->w_cursor.lnum
|
|
1228 || (lnum == curwin->w_cursor.lnum
|
|
1229 && (colnr_T)(p - line)
|
|
1230 > curwin->w_cursor.col))
|
|
1231 {
|
|
1232 if (has_syntax)
|
|
1233 {
|
|
1234 col = p - line;
|
|
1235 (void)syn_get_id(lnum, (colnr_T)col,
|
|
1236 FALSE, &can_spell);
|
|
1237
|
|
1238 /* have to get the line again, a multi-line
|
|
1239 * regexp may make it invalid */
|
|
1240 line = ml_get(lnum);
|
|
1241 p = line + col;
|
|
1242 }
|
|
1243 else
|
|
1244 can_spell = TRUE;
|
|
1245
|
|
1246 if (can_spell)
|
|
1247 {
|
|
1248 found_pos.lnum = lnum;
|
|
1249 found_pos.col = p - line;
|
|
1250 #ifdef FEAT_VIRTUALEDIT
|
|
1251 found_pos.coladd = 0;
|
|
1252 #endif
|
|
1253 if (dir == FORWARD)
|
|
1254 {
|
|
1255 /* No need to search further. */
|
|
1256 curwin->w_cursor = found_pos;
|
|
1257 return OK;
|
|
1258 }
|
|
1259 }
|
|
1260 }
|
236
|
1261 }
|
249
|
1262 attr = 0;
|
236
|
1263 }
|
|
1264 p += len;
|
|
1265 if (*p == NUL)
|
|
1266 break;
|
|
1267 nwc = FALSE;
|
|
1268 }
|
|
1269
|
|
1270 /* advance to next character */
|
|
1271 mb_ptr_adv(p);
|
|
1272 wc = nwc;
|
|
1273 }
|
|
1274
|
|
1275 /* Advance to next line. */
|
249
|
1276 if (dir == BACKWARD)
|
|
1277 {
|
|
1278 if (found_pos.lnum != 0)
|
|
1279 {
|
|
1280 /* Use the last match in the line. */
|
|
1281 curwin->w_cursor = found_pos;
|
|
1282 return OK;
|
|
1283 }
|
|
1284 if (lnum == 1)
|
|
1285 return FAIL;
|
|
1286 --lnum;
|
|
1287 }
|
|
1288 else
|
|
1289 {
|
|
1290 if (lnum == curbuf->b_ml.ml_line_count)
|
|
1291 return FAIL;
|
|
1292 ++lnum;
|
|
1293 }
|
236
|
1294
|
|
1295 line_breakcheck();
|
|
1296 }
|
|
1297
|
|
1298 return FAIL; /* interrupted */
|
|
1299 }
|
|
1300
|
|
1301 /*
|
|
1302 * Load word list for "lang" from a Vim spell file.
|
|
1303 * "lang" must be the language without the region: "en" or "en-rare".
|
|
1304 */
|
|
1305 static slang_T *
|
|
1306 spell_load_lang(lang)
|
|
1307 char_u *lang;
|
|
1308 {
|
|
1309 slang_T *lp;
|
|
1310 char_u fname_enc[80];
|
|
1311 char_u *p;
|
|
1312 int r;
|
|
1313
|
|
1314 lp = slang_alloc(lang);
|
|
1315 if (lp != NULL)
|
|
1316 {
|
|
1317 /* Find all spell files for "lang" in 'runtimepath' and load them.
|
|
1318 * Use 'encoding', except that we use "latin1" for "latin9". */
|
|
1319 #ifdef FEAT_MBYTE
|
|
1320 if (STRLEN(p_enc) < 60 && STRCMP(p_enc, "iso-8859-15") != 0)
|
|
1321 p = p_enc;
|
|
1322 else
|
|
1323 #endif
|
|
1324 p = (char_u *)"latin1";
|
|
1325 sprintf((char *)fname_enc, "spell/%s.%s.spl", lang, p);
|
|
1326
|
|
1327 r = do_in_runtimepath(fname_enc, TRUE, spell_load_file, lp);
|
240
|
1328 if (r == FAIL && !lp->sl_error)
|
|
1329 {
|
|
1330 /* Try loading the ASCII version. */
|
|
1331 sprintf((char *)fname_enc, "spell/%s.ascii.spl", lang);
|
|
1332
|
|
1333 r = do_in_runtimepath(fname_enc, TRUE, spell_load_file, lp);
|
|
1334 }
|
236
|
1335 if (r == FAIL || lp->sl_error)
|
|
1336 {
|
|
1337 slang_free(lp);
|
|
1338 lp = NULL;
|
|
1339 if (r == FAIL)
|
|
1340 smsg((char_u *)_("Warning: Cannot find word list \"%s\""),
|
|
1341 fname_enc + 6);
|
|
1342 }
|
|
1343 else
|
|
1344 {
|
|
1345 lp->sl_next = first_lang;
|
|
1346 first_lang = lp;
|
|
1347 }
|
|
1348 }
|
|
1349
|
|
1350 return lp;
|
|
1351 }
|
|
1352
|
|
1353 /*
|
|
1354 * Allocate a new slang_T.
|
|
1355 * Caller must fill "sl_next".
|
|
1356 */
|
|
1357 static slang_T *
|
|
1358 slang_alloc(lang)
|
|
1359 char_u *lang;
|
|
1360 {
|
|
1361 slang_T *lp;
|
|
1362
|
|
1363 lp = (slang_T *)alloc(sizeof(slang_T));
|
|
1364 if (lp != NULL)
|
|
1365 {
|
|
1366 lp->sl_name = vim_strsave(lang);
|
|
1367 hash_init(&lp->sl_words);
|
|
1368 ga_init2(&lp->sl_preftab, sizeof(hashtab_T), 4);
|
243
|
1369 hash_init(&lp->sl_prewords);
|
236
|
1370 ga_init2(&lp->sl_sufftab, sizeof(hashtab_T), 4);
|
|
1371 lp->sl_prefzero = NULL;
|
|
1372 lp->sl_suffzero = NULL;
|
|
1373 lp->sl_try = NULL;
|
|
1374 ga_init2(&lp->sl_rep, sizeof(repentry_T), 4);
|
|
1375 lp->sl_regions[0] = NUL;
|
|
1376 lp->sl_block = NULL;
|
|
1377 lp->sl_error = FALSE;
|
|
1378 }
|
|
1379 return lp;
|
|
1380 }
|
|
1381
|
|
1382 /*
|
|
1383 * Free the contents of an slang_T and the structure itself.
|
|
1384 */
|
|
1385 static void
|
|
1386 slang_free(lp)
|
|
1387 slang_T *lp;
|
|
1388 {
|
|
1389 sblock_T *sp;
|
|
1390 int i;
|
243
|
1391 fword_T *fw;
|
|
1392 int todo;
|
|
1393 hashitem_T *hi;
|
236
|
1394
|
|
1395 vim_free(lp->sl_name);
|
243
|
1396
|
|
1397 /* The words themselves are in memory blocks referenced by "sl_block".
|
|
1398 * Only the hashtables for additions need to be cleared. */
|
|
1399 todo = lp->sl_words.ht_used;
|
|
1400 for (hi = lp->sl_words.ht_array; todo > 0; ++hi)
|
|
1401 {
|
|
1402 if (!HASHITEM_EMPTY(hi))
|
|
1403 {
|
|
1404 --todo;
|
|
1405 fw = HI2FWORD(hi);
|
|
1406 if (fw->fw_flags & BWF_ADDHASH)
|
|
1407 hash_clear((hashtab_T *)fw->fw_adds);
|
|
1408 }
|
|
1409 }
|
236
|
1410 hash_clear(&lp->sl_words);
|
243
|
1411
|
236
|
1412 for (i = 0; i < lp->sl_preftab.ga_len; ++i)
|
|
1413 hash_clear(((hashtab_T *)lp->sl_preftab.ga_data) + i);
|
|
1414 ga_clear(&lp->sl_preftab);
|
243
|
1415 hash_clear(&lp->sl_prewords);
|
236
|
1416 for (i = 0; i < lp->sl_sufftab.ga_len; ++i)
|
|
1417 hash_clear(((hashtab_T *)lp->sl_sufftab.ga_data) + i);
|
|
1418 ga_clear(&lp->sl_sufftab);
|
|
1419 ga_clear(&lp->sl_rep);
|
|
1420 vim_free(lp->sl_try);
|
|
1421 while (lp->sl_block != NULL)
|
|
1422 {
|
|
1423 sp = lp->sl_block;
|
|
1424 lp->sl_block = sp->sb_next;
|
|
1425 vim_free(sp);
|
|
1426 }
|
|
1427 vim_free(lp);
|
|
1428 }
|
|
1429
|
|
1430 /*
|
|
1431 * Load one spell file into an slang_T.
|
|
1432 * Invoked through do_in_runtimepath().
|
|
1433 */
|
|
1434 static void
|
|
1435 spell_load_file(fname, cookie)
|
|
1436 char_u *fname;
|
|
1437 void *cookie; /* points to the slang_T to be filled */
|
|
1438 {
|
|
1439 slang_T *lp = cookie;
|
|
1440 FILE *fd;
|
|
1441 char_u buf[MAXWLEN + 1];
|
|
1442 char_u cbuf[MAXWLEN + 1];
|
|
1443 char_u fbuf[MAXWLEN + 1];
|
243
|
1444 char_u affixbuf[256 * 2 * 2]; /* max 2 * 256 affix nrs of 2 bytes */
|
236
|
1445 char_u *p;
|
|
1446 int itm;
|
|
1447 int i;
|
|
1448 int affcount;
|
|
1449 int affnr;
|
|
1450 int affflags;
|
|
1451 int affitemcnt;
|
243
|
1452 int prefixcnt, suffixcnt;
|
236
|
1453 int bl_used = SBLOCKSIZE;
|
|
1454 int widx;
|
240
|
1455 int prefm = 0; /* 1 if <= 256 prefixes, sizeof(short_u) otherw. */
|
|
1456 int suffm = 0; /* 1 if <= 256 suffixes, sizeof(short_u) otherw. */
|
236
|
1457 int wlen;
|
|
1458 int flags;
|
|
1459 affitem_T *ai, *ai2, **aip;
|
|
1460 int round;
|
|
1461 char_u *save_sourcing_name = sourcing_name;
|
|
1462 linenr_T save_sourcing_lnum = sourcing_lnum;
|
|
1463 int cnt;
|
|
1464 int choplen;
|
|
1465 int addlen;
|
|
1466 int leadlen;
|
|
1467 int wordcount;
|
|
1468 fword_T *fw, *fw2;
|
|
1469 garray_T *gap;
|
|
1470 hashtab_T *ht;
|
|
1471 hashitem_T *hi;
|
|
1472 hash_T hash;
|
|
1473 int adds;
|
243
|
1474 addword_T *aw, *naw;
|
236
|
1475 int flen;
|
243
|
1476 int xlen;
|
236
|
1477
|
|
1478 fd = fopen((char *)fname, "r");
|
|
1479 if (fd == NULL)
|
|
1480 {
|
|
1481 EMSG2(_(e_notopen), fname);
|
|
1482 goto errorend;
|
|
1483 }
|
|
1484
|
|
1485 /* Set sourcing_name, so that error messages mention the file name. */
|
|
1486 sourcing_name = fname;
|
|
1487 sourcing_lnum = 0;
|
|
1488
|
|
1489 /* <HEADER>: <fileID> <regioncnt> <regionname> ... */
|
|
1490 for (i = 0; i < VIMSPELLMAGICL; ++i)
|
|
1491 buf[i] = getc(fd); /* <fileID> */
|
|
1492 if (STRNCMP(buf, VIMSPELLMAGIC, VIMSPELLMAGICL) != 0)
|
|
1493 {
|
|
1494 EMSG(_("E757: Wrong file ID in spell file"));
|
|
1495 goto errorend;
|
|
1496 }
|
|
1497
|
|
1498 cnt = getc(fd); /* <regioncnt> */
|
|
1499 if (cnt == EOF)
|
|
1500 {
|
|
1501 truncerr:
|
|
1502 EMSG(_("E758: Truncated spell file"));
|
|
1503 goto errorend;
|
|
1504 }
|
|
1505 if (cnt > 8)
|
|
1506 {
|
|
1507 formerr:
|
|
1508 EMSG(_("E759: Format error in spell file"));
|
|
1509 goto errorend;
|
|
1510 }
|
|
1511 for (i = 0; i < cnt; ++i)
|
|
1512 {
|
|
1513 lp->sl_regions[i * 2] = getc(fd); /* <regionname> */
|
|
1514 lp->sl_regions[i * 2 + 1] = getc(fd);
|
|
1515 }
|
|
1516 lp->sl_regions[cnt * 2] = NUL;
|
|
1517
|
|
1518 /* round 1: <PREFIXLIST>: <affcount> <afftotcnt> <affix> ...
|
|
1519 * round 2: <SUFFIXLIST>: <affcount> <afftotcnt> <affix> ... */
|
|
1520 for (round = 1; round <= 2; ++round)
|
|
1521 {
|
|
1522 affcount = (getc(fd) << 8) + getc(fd); /* <affcount> */
|
|
1523 if (affcount < 0)
|
|
1524 goto truncerr;
|
|
1525 if (round == 1)
|
|
1526 {
|
|
1527 gap = &lp->sl_preftab;
|
|
1528 aip = &lp->sl_prefzero;
|
|
1529 lp->sl_prefcnt = affcount;
|
243
|
1530 prefm = affcount > 256 ? 2 : 1;
|
236
|
1531 }
|
|
1532 else
|
|
1533 {
|
|
1534 gap = &lp->sl_sufftab;
|
|
1535 aip = &lp->sl_suffzero;
|
|
1536 lp->sl_suffcnt = affcount;
|
243
|
1537 suffm = affcount > 256 ? 2 : 1;
|
236
|
1538 }
|
|
1539
|
|
1540 i = (getc(fd) << 8) + getc(fd); /* <afftotcnt> */
|
|
1541 /* afftotcnt is not used */
|
|
1542
|
|
1543 /*
|
|
1544 * For each affix NR there can be several affixes.
|
|
1545 */
|
|
1546 for (affnr = 0; affnr < affcount; ++affnr)
|
|
1547 {
|
243
|
1548 /* <affix>: <affitemcnt> <affitem> ... */
|
236
|
1549 affitemcnt = (getc(fd) << 8) + getc(fd); /* <affitemcnt> */
|
|
1550 if (affitemcnt < 0)
|
|
1551 goto truncerr;
|
|
1552 for (itm = 0; itm < affitemcnt; ++itm)
|
|
1553 {
|
243
|
1554 /* <affitem>: <affflags> <affchoplen> <affchop>
|
|
1555 * <affaddlen> <affadd> */
|
|
1556 affflags = getc(fd); /* <affflags> */
|
236
|
1557 choplen = getc(fd); /* <affchoplen> */
|
|
1558 if (choplen == EOF)
|
|
1559 goto truncerr;
|
|
1560 if (choplen >= MAXWLEN)
|
|
1561 goto formerr;
|
|
1562 for (i = 0; i < choplen; ++i) /* <affchop> */
|
|
1563 buf[i] = getc(fd);
|
|
1564 buf[i] = NUL;
|
|
1565 addlen = getc(fd); /* <affaddlen> */
|
|
1566 if (addlen == EOF)
|
|
1567 goto truncerr;
|
243
|
1568 if (affflags & AFF_PREWORD)
|
|
1569 xlen = addlen + 2; /* space for lead and trail string */
|
|
1570 else
|
|
1571 xlen = 0;
|
|
1572
|
236
|
1573 /* Get room to store the affitem_T, chop and add strings. */
|
|
1574 p = (char_u *)getroom(lp, &bl_used,
|
243
|
1575 sizeof(affitem_T) + addlen + choplen + 1 + xlen);
|
236
|
1576 if (p == NULL)
|
|
1577 goto errorend;
|
|
1578
|
|
1579 ai = (affitem_T *)p;
|
|
1580 ai->ai_nr = affnr;
|
243
|
1581 ai->ai_flags = affflags;
|
236
|
1582 ai->ai_choplen = choplen;
|
|
1583 ai->ai_addlen = addlen;
|
|
1584
|
243
|
1585 /* Chop string is at ai_add[ai_addlen + 1]. */
|
|
1586 p = ai->ai_add + addlen + 1;
|
236
|
1587 STRCPY(p, buf);
|
|
1588
|
|
1589 p = ai->ai_add;
|
|
1590 for (i = 0; i < addlen; ++i) /* <affadd> */
|
|
1591 p[i] = getc(fd);
|
|
1592 p[i] = NUL;
|
|
1593
|
243
|
1594 if (affflags & AFF_PREWORD)
|
236
|
1595 {
|
243
|
1596 int l, leadoff, trailoff;
|
|
1597
|
|
1598 /*
|
|
1599 * Separate lead and trail string, put word at ai_add, so
|
|
1600 * that it can be used as hashtable key.
|
|
1601 */
|
|
1602 /* lead string: up to first word char */
|
|
1603 while (*p != NUL && !spell_iswordc(p))
|
|
1604 mb_ptr_adv(p);
|
|
1605 ai->ai_leadlen = p - ai->ai_add;
|
|
1606 leadoff = addlen + choplen + 2;
|
|
1607 mch_memmove(ai->ai_add + leadoff, ai->ai_add,
|
|
1608 ai->ai_leadlen);
|
|
1609 ai->ai_add[leadoff + ai->ai_leadlen] = NUL;
|
|
1610
|
|
1611 /* trail string: after last word char */
|
|
1612 while (*p != NUL && spell_iswordc(p))
|
|
1613 mb_ptr_adv(p);
|
|
1614 trailoff = leadoff + ai->ai_leadlen + 1;
|
|
1615 STRCPY(ai->ai_add + trailoff, p);
|
|
1616 ai->ai_taillen = STRLEN(p);
|
|
1617
|
|
1618 /* word itself */
|
|
1619 l = (p - ai->ai_add) - ai->ai_leadlen;
|
|
1620 mch_memmove(ai->ai_add, ai->ai_add + ai->ai_leadlen, l);
|
|
1621 ai->ai_add[l] = NUL;
|
|
1622 hash = hash_hash(ai->ai_add);
|
|
1623 hi = hash_lookup(&lp->sl_prewords, ai->ai_add, hash);
|
236
|
1624 if (HASHITEM_EMPTY(hi))
|
|
1625 {
|
243
|
1626 /* First affix with this word, add to hashtable. */
|
|
1627 hash_add_item(&lp->sl_prewords, hi, ai->ai_add, hash);
|
236
|
1628 ai->ai_next = NULL;
|
|
1629 }
|
|
1630 else
|
|
1631 {
|
243
|
1632 /* There already is an affix with this word, link in
|
|
1633 * the list. */
|
236
|
1634 ai2 = HI2AI(hi);
|
|
1635 ai->ai_next = ai2->ai_next;
|
|
1636 ai2->ai_next = ai;
|
|
1637 }
|
|
1638 }
|
243
|
1639 else
|
|
1640 {
|
|
1641 /*
|
|
1642 * Add the affix to a hashtable. Which one depends on the
|
|
1643 * length of the added string in characters.
|
|
1644 */
|
|
1645 #ifdef FEAT_MBYTE
|
|
1646 /* Change "addlen" from length in bytes to length in
|
|
1647 * chars. */
|
|
1648 if (has_mbyte)
|
|
1649 addlen = mb_charlen(p);
|
|
1650 #endif
|
|
1651 if (addlen == 0)
|
|
1652 {
|
|
1653 /* Link in list of zero length affixes. */
|
|
1654 ai->ai_next = *aip;
|
|
1655 *aip = ai;
|
|
1656 }
|
|
1657 else
|
|
1658 {
|
|
1659 if (gap->ga_len < addlen)
|
|
1660 {
|
|
1661 /* Longer affix, need more hashtables. */
|
|
1662 if (ga_grow(gap, addlen - gap->ga_len) == FAIL)
|
|
1663 goto errorend;
|
|
1664
|
|
1665 /* Re-allocating ga_data means that an ht_array
|
|
1666 * pointing to ht_smallarray becomes invalid. We
|
|
1667 * can recognize this: ht_mask is at its init
|
|
1668 * value. */
|
|
1669 for (i = 0; i < gap->ga_len; ++i)
|
|
1670 {
|
|
1671 ht = ((hashtab_T *)gap->ga_data) + i;
|
|
1672 if (ht->ht_mask == HT_INIT_SIZE - 1)
|
|
1673 ht->ht_array = ht->ht_smallarray;
|
|
1674 }
|
|
1675
|
|
1676 /* Init the newly used hashtable(s). */
|
|
1677 while (gap->ga_len < addlen)
|
|
1678 {
|
|
1679 hash_init(((hashtab_T *)gap->ga_data)
|
|
1680 + gap->ga_len);
|
|
1681 ++gap->ga_len;
|
|
1682 }
|
|
1683 }
|
|
1684 ht = ((hashtab_T *)gap->ga_data) + addlen - 1;
|
|
1685 hash = hash_hash(p);
|
|
1686 hi = hash_lookup(ht, p, hash);
|
|
1687 if (HASHITEM_EMPTY(hi))
|
|
1688 {
|
|
1689 /* First affix with this "ai_add", add to
|
|
1690 * hashtable. */
|
|
1691 hash_add_item(ht, hi, p, hash);
|
|
1692 ai->ai_next = NULL;
|
|
1693 }
|
|
1694 else
|
|
1695 {
|
|
1696 /* There already is an affix with this "ai_add",
|
|
1697 * link in the list. */
|
|
1698 ai2 = HI2AI(hi);
|
|
1699 ai->ai_next = ai2->ai_next;
|
|
1700 ai2->ai_next = ai;
|
|
1701 }
|
|
1702 }
|
|
1703 }
|
236
|
1704 }
|
|
1705 }
|
|
1706 }
|
|
1707
|
|
1708 /* <SUGGEST> : <suggestlen> <more> ... */
|
|
1709 /* TODO, just skip this for now */
|
|
1710 i = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8) + getc(fd);
|
|
1711 while (i-- > 0)
|
|
1712 if (getc(fd) == EOF) /* <suggestlen> */
|
|
1713 goto truncerr;
|
|
1714
|
|
1715 /* <WORDLIST>: <wordcount> <worditem> ... */ /* <wordcount> */
|
|
1716 wordcount = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8)
|
|
1717 + getc(fd);
|
|
1718 if (wordcount < 0)
|
|
1719 goto truncerr;
|
|
1720
|
|
1721 /* Init hashtable for this number of words, so that it doesn't need to
|
|
1722 * reallocate the table halfway. */
|
|
1723 hash_lock_size(&lp->sl_words, wordcount);
|
|
1724
|
|
1725 for (widx = 0; ; ++widx)
|
|
1726 {
|
|
1727 /* <worditem>: <nr> <string> <flags> [<flags2>]
|
|
1728 * [<caselen> <caseword>]
|
|
1729 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
1730 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
1731 * [<region>]
|
|
1732 * [<addcnt> <add> ...]
|
|
1733 */
|
|
1734 /* Use <nr> bytes from the previous word. */
|
|
1735 wlen = getc(fd); /* <nr> */
|
|
1736 if (wlen == EOF)
|
|
1737 {
|
|
1738 if (widx >= wordcount) /* normal way to end the file */
|
|
1739 break;
|
|
1740 goto truncerr;
|
|
1741 }
|
|
1742
|
|
1743 /* Read further word bytes until one below 0x20, that must be the
|
|
1744 * flags. Keep this fast! */
|
|
1745 for (;;)
|
|
1746 {
|
|
1747 if ((buf[wlen] = getc(fd)) < 0x20) /* <string> */
|
|
1748 break;
|
|
1749 if (++wlen == MAXWLEN)
|
|
1750 goto formerr;
|
|
1751 }
|
|
1752 flags = buf[wlen]; /* <flags> */
|
|
1753 buf[wlen] = NUL;
|
|
1754
|
|
1755 /* Get more flags if they're there. */
|
|
1756 if (flags & BWF_SECOND)
|
|
1757 flags += getc(fd) << 8; /* <flags2> */
|
|
1758
|
|
1759 if (flags & BWF_KEEPCAP)
|
|
1760 {
|
|
1761 /* Read <caselen> and <caseword> first, its length may differ from
|
|
1762 * the case-folded word. Note: this should only happen after the
|
|
1763 * basic word! */
|
|
1764 wlen = getc(fd);
|
243
|
1765 if (wlen < 0)
|
236
|
1766 goto truncerr;
|
|
1767 for (i = 0; i < wlen; ++i)
|
|
1768 cbuf[i] = getc(fd);
|
|
1769 cbuf[i] = NUL;
|
|
1770 }
|
|
1771
|
243
|
1772 /* Optional prefixes */
|
|
1773 p = affixbuf;
|
|
1774 if (flags & BWF_PREFIX)
|
|
1775 {
|
|
1776 cnt = getc(fd); /* <affixcnt> */
|
|
1777 if (cnt < 0)
|
|
1778 goto truncerr;
|
|
1779 prefixcnt = cnt;
|
|
1780 for (i = cnt * prefm; --i >= 0; ) /* <affixNR> */
|
|
1781 *p++ = getc(fd);
|
|
1782 }
|
|
1783 else
|
|
1784 prefixcnt = 0;
|
|
1785
|
|
1786 /* Optional suffixes */
|
|
1787 if (flags & BWF_SUFFIX)
|
|
1788 {
|
|
1789 cnt = getc(fd); /* <affixcnt> */
|
|
1790 if (cnt < 0)
|
|
1791 goto truncerr;
|
|
1792 suffixcnt = cnt;
|
|
1793 for (i = cnt * suffm; --i >= 0; ) /* <affixNR> */
|
|
1794 *p++ = getc(fd);
|
|
1795 }
|
|
1796 else
|
|
1797 suffixcnt = 0;
|
|
1798
|
|
1799 /* Find room to store the word in an fword_T. */
|
|
1800 fw = (fword_T *)getroom(lp, &bl_used, (int)sizeof(fword_T) + wlen
|
|
1801 + (p - affixbuf));
|
236
|
1802 if (fw == NULL)
|
|
1803 goto errorend;
|
|
1804 mch_memmove(fw->fw_word, (flags & BWF_KEEPCAP) ? cbuf : buf, wlen + 1);
|
243
|
1805
|
|
1806 /* Put the affix NRs just after the word, if any. */
|
|
1807 if (p > affixbuf)
|
|
1808 mch_memmove(fw->fw_word + wlen + 1, affixbuf, p - affixbuf);
|
|
1809
|
236
|
1810 fw->fw_flags = flags;
|
243
|
1811 fw->fw_prefixcnt = prefixcnt;
|
|
1812 fw->fw_suffixcnt = suffixcnt;
|
236
|
1813
|
|
1814 hash = hash_hash(buf);
|
|
1815 hi = hash_lookup(&lp->sl_words, buf, hash);
|
|
1816 if (HASHITEM_EMPTY(hi))
|
|
1817 {
|
|
1818 if (hash_add_item(&lp->sl_words, hi, fw->fw_word, hash) == FAIL)
|
|
1819 goto errorend;
|
|
1820 fw->fw_next = NULL;
|
|
1821 }
|
|
1822 else
|
|
1823 {
|
|
1824 /* Already have this basic word in the hashtable, this one will
|
243
|
1825 * have different case flags and/or affixes. */
|
236
|
1826 fw2 = HI2FWORD(hi);
|
|
1827 fw->fw_next = fw2->fw_next;
|
|
1828 fw2->fw_next = fw;
|
|
1829 --widx; /* don't count this one */
|
|
1830 }
|
|
1831
|
|
1832 if (flags & BWF_REGION)
|
|
1833 fw->fw_region = getc(fd); /* <region> */
|
|
1834 else
|
|
1835 fw->fw_region = REGION_ALL;
|
|
1836
|
|
1837 fw->fw_adds = NULL;
|
|
1838 if (flags & BWF_ADDS)
|
|
1839 {
|
249
|
1840 if (flags & BWF_ADDS_M)
|
|
1841 adds = (getc(fd) << 8) + getc(fd); /* <addcnt> */
|
|
1842 else
|
|
1843 adds = getc(fd); /* <addcnt> */
|
236
|
1844
|
243
|
1845 if (adds > 30)
|
|
1846 {
|
|
1847 /* Use a hashtable to loopup the part until the next word end.
|
|
1848 * This uses more memory and involves some overhead, thus only
|
|
1849 * do it when there are many additions (e.g., for French). */
|
|
1850 ht = (hashtab_T *)getroom(lp, &bl_used, sizeof(hashtab_T));
|
|
1851 if (ht == NULL)
|
|
1852 goto errorend;
|
|
1853 hash_init(ht);
|
|
1854 fw->fw_adds = (addword_T *)ht;
|
|
1855 fw->fw_flags |= BWF_ADDHASH;
|
|
1856
|
|
1857 /* Preset the size of the hashtable. It's never unlocked. */
|
|
1858 hash_lock_size(ht, adds + 1);
|
|
1859 }
|
|
1860 else
|
|
1861 ht = NULL;
|
|
1862
|
236
|
1863 while (--adds >= 0)
|
|
1864 {
|
249
|
1865 /* <add>: <addflags> <addlen> [<leadlen>] [<copylen>]
|
|
1866 * [<addstring>] [<region>] */
|
236
|
1867 flags = getc(fd); /* <addflags> */
|
|
1868 addlen = getc(fd); /* <addlen> */
|
|
1869 if (addlen == EOF)
|
|
1870 goto truncerr;
|
|
1871 if (addlen >= MAXWLEN)
|
|
1872 goto formerr;
|
|
1873
|
249
|
1874 if (flags & ADD_LEADLEN)
|
|
1875 leadlen = getc(fd); /* <leadlen> */
|
|
1876 else
|
|
1877 leadlen = 0;
|
|
1878
|
236
|
1879 if (addlen > 0)
|
|
1880 {
|
249
|
1881 if (flags & ADD_COPYLEN)
|
|
1882 i = getc(fd); /* <copylen> */
|
|
1883 else
|
|
1884 i = 0;
|
|
1885 for ( ; i < addlen; ++i) /* <addstring> */
|
236
|
1886 cbuf[i] = getc(fd);
|
|
1887 cbuf[i] = NUL;
|
|
1888 }
|
|
1889
|
|
1890 if (flags & ADD_KEEPCAP)
|
|
1891 {
|
|
1892 /* <addstring> is in original case, need to get
|
|
1893 * case-folded word too. */
|
|
1894 (void)str_foldcase(cbuf, addlen, fbuf, MAXWLEN);
|
|
1895 flen = addlen - leadlen + 1;
|
|
1896 addlen = STRLEN(fbuf);
|
|
1897 }
|
|
1898 else
|
|
1899 flen = 0;
|
|
1900
|
|
1901 aw = (addword_T *)getroom(lp, &bl_used,
|
|
1902 sizeof(addword_T) + addlen + flen);
|
|
1903 if (aw == NULL)
|
|
1904 goto errorend;
|
|
1905
|
|
1906 if (flags & ADD_KEEPCAP)
|
|
1907 {
|
|
1908 /* Put the addition in original case after the case-folded
|
|
1909 * string. */
|
|
1910 STRCPY(aw->aw_word, fbuf);
|
|
1911 STRCPY(aw->aw_word + addlen + 1, cbuf + leadlen);
|
|
1912 }
|
|
1913 else
|
|
1914 STRCPY(aw->aw_word, cbuf);
|
|
1915
|
|
1916 aw->aw_flags = flags;
|
|
1917 aw->aw_wordlen = addlen;
|
243
|
1918 aw->aw_leadlen = leadlen;
|
236
|
1919
|
|
1920 if (flags & ADD_REGION)
|
|
1921 aw->aw_region = getc(fd); /* <region> */
|
|
1922 else
|
|
1923 aw->aw_region = REGION_ALL;
|
243
|
1924
|
|
1925 if (ht == NULL)
|
|
1926 {
|
|
1927 /* Using simple linked list, put it in front. */
|
|
1928 aw->aw_next = fw->fw_adds;
|
|
1929 fw->fw_adds = aw;
|
|
1930 aw->aw_saveb = NUL;
|
|
1931 }
|
|
1932 else
|
|
1933 {
|
|
1934 /* Put addition in hashtable. For key we use the part up
|
|
1935 * to the next end-of-word. */
|
|
1936 if (leadlen == 0)
|
|
1937 {
|
|
1938 p = aw->aw_word;
|
|
1939 while (*p != NUL && !spell_iswordc(p))
|
|
1940 mb_ptr_adv(p);
|
|
1941 }
|
|
1942
|
|
1943 if (leadlen != 0 || *p == NUL)
|
|
1944 {
|
|
1945 /* Only non-word characters in addition, add it to the
|
|
1946 * list with the special key NOWC_KEY. Also do this
|
|
1947 * when there is a leadstring, it would get too
|
|
1948 * complicated. */
|
|
1949 hash = hash_hash(NOWC_KEY);
|
|
1950 hi = hash_lookup(ht, NOWC_KEY, hash);
|
|
1951 if (HASHITEM_EMPTY(hi))
|
|
1952 {
|
|
1953 /* we use a dummy item as the list header */
|
|
1954 naw = (addword_T *)getroom(lp, &bl_used,
|
|
1955 sizeof(addword_T) + STRLEN(NOWC_KEY));
|
|
1956 if (naw == NULL)
|
|
1957 goto errorend;
|
|
1958 STRCPY(naw->aw_word, NOWC_KEY);
|
|
1959 hash_add_item(ht, hi, naw->aw_word, hash);
|
|
1960 naw->aw_next = aw;
|
|
1961 aw->aw_next = NULL;
|
|
1962 }
|
|
1963 else
|
|
1964 {
|
|
1965 naw = HI2ADDWORD(hi);
|
|
1966 aw->aw_next = naw->aw_next;
|
|
1967 naw->aw_next = aw;
|
|
1968 }
|
|
1969 aw->aw_saveb = NUL;
|
|
1970 }
|
|
1971 else
|
|
1972 {
|
|
1973 /* Truncate at next non-word character, store that
|
|
1974 * byte in "aw_saveb". */
|
|
1975 while (*p != NUL && spell_iswordc(p))
|
|
1976 mb_ptr_adv(p);
|
|
1977 aw->aw_saveb = *p;
|
|
1978 *p = NUL;
|
|
1979 hash = hash_hash(aw->aw_word);
|
|
1980 hi = hash_lookup(ht, aw->aw_word, hash);
|
|
1981 if (HASHITEM_EMPTY(hi))
|
|
1982 {
|
|
1983 hash_add_item(ht, hi, aw->aw_word, hash);
|
|
1984 aw->aw_next = NULL;
|
|
1985 }
|
|
1986 else
|
|
1987 {
|
|
1988 naw = HI2ADDWORD(hi);
|
|
1989 aw->aw_next = naw->aw_next;
|
|
1990 naw->aw_next = aw;
|
|
1991 }
|
|
1992 }
|
|
1993 }
|
236
|
1994 }
|
|
1995 }
|
|
1996 }
|
|
1997 goto end_OK;
|
|
1998
|
|
1999 errorend:
|
|
2000 lp->sl_error = TRUE;
|
|
2001 end_OK:
|
|
2002 if (fd != NULL)
|
|
2003 fclose(fd);
|
|
2004 hash_unlock(&lp->sl_words);
|
|
2005 sourcing_name = save_sourcing_name;
|
|
2006 sourcing_lnum = save_sourcing_lnum;
|
|
2007 }
|
|
2008
|
|
2009 /*
|
|
2010 * Get part of an sblock_T, at least "len" bytes long.
|
|
2011 * Returns NULL when out of memory.
|
|
2012 */
|
|
2013 static void *
|
|
2014 getroom(lp, bl_used, len)
|
|
2015 slang_T *lp; /* lp->sl_block is current block or NULL */
|
|
2016 int *bl_used; /* used up from current block */
|
|
2017 int len; /* length needed */
|
|
2018 {
|
|
2019 char_u *p;
|
|
2020 sblock_T *bl = lp->sl_block;
|
|
2021
|
|
2022 if (bl == NULL || *bl_used + len > SBLOCKSIZE)
|
|
2023 {
|
|
2024 /* Allocate a block of memory. This is not freed until spell_reload()
|
|
2025 * is called. */
|
|
2026 bl = (sblock_T *)alloc((unsigned)(sizeof(sblock_T) + SBLOCKSIZE));
|
|
2027 if (bl == NULL)
|
|
2028 return NULL;
|
|
2029 bl->sb_next = lp->sl_block;
|
|
2030 lp->sl_block = bl;
|
|
2031 *bl_used = 0;
|
|
2032 }
|
|
2033
|
|
2034 p = bl->sb_data + *bl_used;
|
|
2035 *bl_used += len;
|
|
2036
|
|
2037 return p;
|
|
2038 }
|
|
2039
|
|
2040 /*
|
|
2041 * Parse 'spelllang' and set buf->b_langp accordingly.
|
|
2042 * Returns an error message or NULL.
|
|
2043 */
|
|
2044 char_u *
|
|
2045 did_set_spelllang(buf)
|
|
2046 buf_T *buf;
|
|
2047 {
|
|
2048 garray_T ga;
|
|
2049 char_u *lang;
|
|
2050 char_u *e;
|
|
2051 char_u *region;
|
|
2052 int region_mask;
|
|
2053 slang_T *lp;
|
|
2054 int c;
|
|
2055 char_u lbuf[MAXWLEN + 1];
|
|
2056
|
|
2057 ga_init2(&ga, sizeof(langp_T), 2);
|
|
2058
|
|
2059 /* loop over comma separated languages. */
|
|
2060 for (lang = buf->b_p_spl; *lang != NUL; lang = e)
|
|
2061 {
|
|
2062 e = vim_strchr(lang, ',');
|
|
2063 if (e == NULL)
|
|
2064 e = lang + STRLEN(lang);
|
240
|
2065 region = NULL;
|
236
|
2066 if (e > lang + 2)
|
|
2067 {
|
|
2068 if (e - lang >= MAXWLEN)
|
|
2069 {
|
|
2070 ga_clear(&ga);
|
|
2071 return e_invarg;
|
|
2072 }
|
|
2073 if (lang[2] == '_')
|
|
2074 region = lang + 3;
|
|
2075 }
|
|
2076
|
|
2077 for (lp = first_lang; lp != NULL; lp = lp->sl_next)
|
|
2078 if (STRNICMP(lp->sl_name, lang, 2) == 0)
|
|
2079 break;
|
|
2080
|
|
2081 if (lp == NULL)
|
|
2082 {
|
|
2083 /* Not found, load the language. */
|
|
2084 STRNCPY(lbuf, lang, e - lang);
|
|
2085 lbuf[e - lang] = NUL;
|
|
2086 if (region != NULL)
|
|
2087 mch_memmove(lbuf + 2, lbuf + 5, e - lang - 4);
|
|
2088 lp = spell_load_lang(lbuf);
|
|
2089 }
|
|
2090
|
|
2091 if (lp != NULL)
|
|
2092 {
|
|
2093 if (region == NULL)
|
|
2094 region_mask = REGION_ALL;
|
|
2095 else
|
|
2096 {
|
|
2097 /* find region in sl_regions */
|
|
2098 c = find_region(lp->sl_regions, region);
|
|
2099 if (c == REGION_ALL)
|
|
2100 {
|
|
2101 c = *e;
|
|
2102 *e = NUL;
|
|
2103 smsg((char_u *)_("Warning: region %s not supported"), lang);
|
|
2104 *e = c;
|
|
2105 region_mask = REGION_ALL;
|
|
2106 }
|
|
2107 else
|
|
2108 region_mask = 1 << c;
|
|
2109 }
|
|
2110
|
|
2111 if (ga_grow(&ga, 1) == FAIL)
|
|
2112 {
|
|
2113 ga_clear(&ga);
|
|
2114 return e_outofmem;
|
|
2115 }
|
|
2116 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = lp;
|
|
2117 LANGP_ENTRY(ga, ga.ga_len)->lp_region = region_mask;
|
|
2118 ++ga.ga_len;
|
|
2119 }
|
|
2120
|
|
2121 if (*e == ',')
|
|
2122 ++e;
|
|
2123 }
|
|
2124
|
|
2125 /* Add a NULL entry to mark the end of the list. */
|
|
2126 if (ga_grow(&ga, 1) == FAIL)
|
|
2127 {
|
|
2128 ga_clear(&ga);
|
|
2129 return e_outofmem;
|
|
2130 }
|
|
2131 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = NULL;
|
|
2132 ++ga.ga_len;
|
|
2133
|
|
2134 /* Everything is fine, store the new b_langp value. */
|
|
2135 ga_clear(&buf->b_langp);
|
|
2136 buf->b_langp = ga;
|
|
2137
|
|
2138 return NULL;
|
|
2139 }
|
|
2140
|
|
2141 /*
|
|
2142 * Find the region "region[2]" in "rp" (points to "sl_regions").
|
|
2143 * Each region is simply stored as the two characters of it's name.
|
|
2144 * Returns the index if found, REGION_ALL if not found.
|
|
2145 */
|
|
2146 static int
|
|
2147 find_region(rp, region)
|
|
2148 char_u *rp;
|
|
2149 char_u *region;
|
|
2150 {
|
|
2151 int i;
|
|
2152
|
|
2153 for (i = 0; ; i += 2)
|
|
2154 {
|
|
2155 if (rp[i] == NUL)
|
|
2156 return REGION_ALL;
|
|
2157 if (rp[i] == region[0] && rp[i + 1] == region[1])
|
|
2158 break;
|
|
2159 }
|
|
2160 return i / 2;
|
|
2161 }
|
|
2162
|
|
2163 /*
|
|
2164 * Return type of word:
|
|
2165 * w word 0
|
|
2166 * Word BWF_ONECAP
|
|
2167 * W WORD BWF_ALLCAP
|
|
2168 * WoRd wOrd BWF_KEEPCAP
|
|
2169 */
|
|
2170 static int
|
|
2171 captype(word, end)
|
|
2172 char_u *word;
|
|
2173 char_u *end;
|
|
2174 {
|
|
2175 char_u *p;
|
|
2176 int c;
|
|
2177 int firstcap;
|
|
2178 int allcap;
|
|
2179 int past_second = FALSE; /* past second word char */
|
|
2180
|
|
2181 /* find first letter */
|
|
2182 for (p = word; !spell_iswordc(p); mb_ptr_adv(p))
|
|
2183 if (p >= end)
|
|
2184 return 0; /* only non-word characters, illegal word */
|
|
2185 #ifdef FEAT_MBYTE
|
|
2186 c = mb_ptr2char_adv(&p);
|
|
2187 #else
|
|
2188 c = *p++;
|
|
2189 #endif
|
|
2190 firstcap = allcap = MB_ISUPPER(c);
|
|
2191
|
|
2192 /*
|
|
2193 * Need to check all letters to find a word with mixed upper/lower.
|
|
2194 * But a word with an upper char only at start is a ONECAP.
|
|
2195 */
|
|
2196 for ( ; p < end; mb_ptr_adv(p))
|
|
2197 if (spell_iswordc(p))
|
|
2198 {
|
|
2199 #ifdef FEAT_MBYTE
|
|
2200 c = mb_ptr2char(p);
|
|
2201 #else
|
|
2202 c = *p;
|
|
2203 #endif
|
|
2204 if (!MB_ISUPPER(c))
|
|
2205 {
|
|
2206 /* UUl -> KEEPCAP */
|
|
2207 if (past_second && allcap)
|
|
2208 return BWF_KEEPCAP;
|
|
2209 allcap = FALSE;
|
|
2210 }
|
|
2211 else if (!allcap)
|
|
2212 /* UlU -> KEEPCAP */
|
|
2213 return BWF_KEEPCAP;
|
|
2214 past_second = TRUE;
|
|
2215 }
|
|
2216
|
|
2217 if (allcap)
|
|
2218 return BWF_ALLCAP;
|
|
2219 if (firstcap)
|
|
2220 return BWF_ONECAP;
|
|
2221 return 0;
|
|
2222 }
|
|
2223
|
|
2224 # if defined(FEAT_MBYTE) || defined(PROTO)
|
|
2225 /*
|
|
2226 * Clear all spelling tables and reload them.
|
|
2227 * Used after 'encoding' is set.
|
|
2228 */
|
|
2229 void
|
|
2230 spell_reload()
|
|
2231 {
|
|
2232 buf_T *buf;
|
|
2233 slang_T *lp;
|
|
2234
|
|
2235 /* Initialize the table for spell_iswordc(). */
|
|
2236 init_spell_chartab();
|
|
2237
|
|
2238 /* Unload all allocated memory. */
|
|
2239 while (first_lang != NULL)
|
|
2240 {
|
|
2241 lp = first_lang;
|
|
2242 first_lang = lp->sl_next;
|
|
2243 slang_free(lp);
|
|
2244 }
|
|
2245
|
|
2246 /* Go through all buffers and handle 'spelllang'. */
|
|
2247 for (buf = firstbuf; buf != NULL; buf = buf->b_next)
|
|
2248 {
|
|
2249 ga_clear(&buf->b_langp);
|
|
2250 if (*buf->b_p_spl != NUL)
|
|
2251 did_set_spelllang(buf);
|
|
2252 }
|
|
2253 }
|
|
2254 # endif
|
|
2255
|
|
2256 /*
|
|
2257 * Recognizing words uses a two-step mechanism:
|
|
2258 * 1. Locate a basic word, made out of word characters only and separated by
|
|
2259 * non-word characters.
|
|
2260 * 2. When a basic word is found, check if (possibly required) additions
|
|
2261 * before and after the word are present.
|
|
2262 *
|
|
2263 * Both mechanisms use affixes (prefixes and suffixes) to reduce the number of
|
|
2264 * words. When no matching word was found in the hashtable the start of the
|
|
2265 * word is checked for matching prefixes and the end of the word for matching
|
|
2266 * suffixes. All matching affixes are removed and then the resulting word is
|
|
2267 * searched for. If found it is checked if it supports the used affix.
|
|
2268 */
|
|
2269
|
|
2270
|
|
2271 #if defined(FEAT_MBYTE) || defined(PROTO)
|
|
2272 /*
|
|
2273 * Functions for ":mkspell".
|
|
2274 * Only possible with the multi-byte feature.
|
|
2275 */
|
|
2276
|
|
2277 #define MAXLINELEN 300 /* Maximum length in bytes of a line in a .aff
|
|
2278 and .dic file. */
|
|
2279 /*
|
|
2280 * Main structure to store the contents of a ".aff" file.
|
|
2281 */
|
|
2282 typedef struct afffile_S
|
|
2283 {
|
|
2284 char_u *af_enc; /* "SET", normalized, alloc'ed string or NULL */
|
|
2285 char_u *af_try; /* "TRY" line in "af_enc" encoding */
|
|
2286 hashtab_T af_pref; /* hashtable for prefixes, affheader_T */
|
|
2287 hashtab_T af_suff; /* hashtable for suffixes, affheader_T */
|
|
2288 garray_T af_rep; /* list of repentry_T entries from REP lines */
|
|
2289 } afffile_T;
|
|
2290
|
|
2291 typedef struct affentry_S affentry_T;
|
|
2292
|
|
2293 /* Affix header from ".aff" file. Used for af_pref and af_suff. */
|
|
2294 typedef struct affheader_S
|
|
2295 {
|
|
2296 char_u ah_key[2]; /* key for hashtable == name of affix entry */
|
|
2297 int ah_combine;
|
|
2298 affentry_T *ah_first; /* first affix entry */
|
|
2299 short_u ah_affnr; /* used in get_new_aff() */
|
|
2300 } affheader_T;
|
|
2301
|
|
2302 #define HI2AH(hi) ((affheader_T *)(hi)->hi_key)
|
|
2303
|
|
2304 /* Affix entry from ".aff" file. Used for prefixes and suffixes. */
|
|
2305 struct affentry_S
|
|
2306 {
|
|
2307 affentry_T *ae_next; /* next affix with same name/number */
|
|
2308 char_u *ae_chop; /* text to chop off basic word (can be NULL) */
|
|
2309 char_u *ae_add; /* text to add to basic word (can be NULL) */
|
240
|
2310 char_u *ae_add_nw; /* For a suffix: first non-word char in
|
|
2311 * "ae_add"; for a prefix with only non-word
|
|
2312 * chars: equal to "ae_add", for a prefix with
|
|
2313 * word and non-word chars: first non-word
|
|
2314 * char after word char. NULL otherwise. */
|
|
2315 char_u *ae_add_pw; /* For a prefix with both word and non-word
|
|
2316 * chars: first word char. NULL otherwise. */
|
243
|
2317 char_u ae_preword; /* TRUE for a prefix with one word */
|
236
|
2318 char_u *ae_cond; /* condition (NULL for ".") */
|
|
2319 regprog_T *ae_prog; /* regexp program for ae_cond or NULL */
|
|
2320 short_u ae_affnr; /* for old affix: new affix number */
|
|
2321 };
|
|
2322
|
|
2323 /*
|
|
2324 * Structure to store a word from a ".dic" file.
|
|
2325 */
|
|
2326 typedef struct dicword_S
|
|
2327 {
|
|
2328 char_u *dw_affnm; /* original affix names */
|
|
2329 char_u dw_word[1]; /* actually longer: the word in 'encoding' */
|
|
2330 } dicword_T;
|
|
2331
|
|
2332 static dicword_T dumdw;
|
|
2333 #define HI2DW(hi) ((dicword_T *)((hi)->hi_key - (dumdw.dw_word - (char_u *)&dumdw)))
|
|
2334
|
|
2335 /*
|
|
2336 * Structure to store a basic word for the spell file.
|
|
2337 * This is used for ":mkspell", not for spell checking.
|
|
2338 */
|
|
2339 typedef struct basicword_S basicword_T;
|
|
2340 struct basicword_S
|
|
2341 {
|
|
2342 basicword_T *bw_next; /* next word with same basic word */
|
|
2343 basicword_T *bw_cnext; /* next word with same caps */
|
|
2344 int bw_flags; /* BWF_ flags */
|
|
2345 garray_T bw_prefix; /* table with prefix numbers */
|
|
2346 garray_T bw_suffix; /* table with suffix numbers */
|
|
2347 int bw_region; /* region bits */
|
|
2348 char_u *bw_caseword; /* keep-case word */
|
|
2349 char_u *bw_leadstring; /* must come before bw_word */
|
|
2350 char_u *bw_addstring; /* must come after bw_word */
|
|
2351 char_u bw_word[1]; /* actually longer: word case folded */
|
|
2352 };
|
|
2353
|
|
2354 static basicword_T dumbw;
|
|
2355 #define KEY2BW(p) ((basicword_T *)((p) - (dumbw.bw_word - (char_u *)&dumbw)))
|
|
2356 #define HI2BW(hi) KEY2BW((hi)->hi_key)
|
|
2357
|
|
2358 /* Store the affix number related with a certain string. */
|
|
2359 typedef struct affhash_S
|
|
2360 {
|
|
2361 short_u as_nr; /* the affix nr */
|
|
2362 char_u as_word[1]; /* actually longer */
|
|
2363 } affhash_T;
|
|
2364
|
|
2365 static affhash_T dumas;
|
|
2366 #define HI2AS(hi) ((affhash_T *)((hi)->hi_key - (dumas.as_word - (char_u *)&dumas)))
|
|
2367
|
249
|
2368 /* info for writing the spell file */
|
|
2369 typedef struct winfo_S
|
|
2370 {
|
|
2371 FILE *wif_fd;
|
|
2372 basicword_T *wif_prevbw; /* last written basic word */
|
|
2373 int wif_regionmask; /* regions supported */
|
|
2374 int wif_prefm; /* 1 or 2 bytes used for prefix NR */
|
|
2375 int wif_suffm; /* 1 or 2 bytes used for suffix NR */
|
|
2376 long wif_wcount; /* written word count */
|
|
2377 long wif_acount; /* written addition count */
|
|
2378 long wif_addmax; /* max number of additions on one word */
|
|
2379 char_u *wif_addmaxw; /* word with max additions */
|
|
2380 } winfo_T;
|
|
2381
|
236
|
2382
|
240
|
2383 static afffile_T *spell_read_aff __ARGS((char_u *fname, vimconv_T *conv, int ascii));
|
236
|
2384 static void spell_free_aff __ARGS((afffile_T *aff));
|
240
|
2385 static int has_non_ascii __ARGS((char_u *s));
|
|
2386 static int spell_read_dic __ARGS((hashtab_T *ht, char_u *fname, vimconv_T *conv, int ascii));
|
|
2387 static int get_new_aff __ARGS((hashtab_T *oldaff, garray_T *gap, int prefix));
|
236
|
2388 static void spell_free_dic __ARGS((hashtab_T *dic));
|
|
2389 static int same_affentries __ARGS((affheader_T *ah1, affheader_T *ah2));
|
|
2390 static void add_affhash __ARGS((hashtab_T *ht, char_u *key, int newnr));
|
|
2391 static void clear_affhash __ARGS((hashtab_T *ht));
|
|
2392 static void trans_affixes __ARGS((dicword_T *dw, basicword_T *bw, afffile_T *oldaff, hashtab_T *newwords));
|
|
2393 static int build_wordlist __ARGS((hashtab_T *newwords, hashtab_T *oldwords, afffile_T *oldaff, int regionmask));
|
|
2394 static void combine_regions __ARGS((hashtab_T *newwords));
|
|
2395 static int same_affixes __ARGS((basicword_T *bw, basicword_T *nbw));
|
|
2396 static void expand_affixes __ARGS((hashtab_T *newwords, garray_T *prefgap, garray_T *suffgap));
|
|
2397 static void expand_one_aff __ARGS((basicword_T *bw, garray_T *add_words, affentry_T *pae, affentry_T *sae));
|
|
2398 static void add_to_wordlist __ARGS((hashtab_T *newwords, basicword_T *bw));
|
|
2399 static void put_bytes __ARGS((FILE *fd, long_u nr, int len));
|
|
2400 static void write_affix __ARGS((FILE *fd, affheader_T *ah));
|
|
2401 static void write_affixlist __ARGS((FILE *fd, garray_T *aff, int bytes));
|
|
2402 static void write_vim_spell __ARGS((char_u *fname, garray_T *prefga, garray_T *suffga, hashtab_T *newwords, int regcount, char_u *regchars));
|
249
|
2403 static void write_bword __ARGS((winfo_T *wif, basicword_T *bw, int lowcap));
|
236
|
2404 static void free_wordtable __ARGS((hashtab_T *ht));
|
|
2405 static void free_basicword __ARGS((basicword_T *bw));
|
|
2406 static void free_affixentries __ARGS((affentry_T *first));
|
240
|
2407 static void free_affix_entry __ARGS((affentry_T *ap));
|
236
|
2408
|
|
2409 /*
|
|
2410 * Read an affix ".aff" file.
|
|
2411 * Returns an afffile_T, NULL for failure.
|
|
2412 */
|
|
2413 static afffile_T *
|
240
|
2414 spell_read_aff(fname, conv, ascii)
|
236
|
2415 char_u *fname;
|
|
2416 vimconv_T *conv; /* info for encoding conversion */
|
240
|
2417 int ascii; /* Only accept ASCII characters */
|
236
|
2418 {
|
|
2419 FILE *fd;
|
|
2420 afffile_T *aff;
|
|
2421 char_u rline[MAXLINELEN];
|
|
2422 char_u *line;
|
|
2423 char_u *pc = NULL;
|
|
2424 char_u *(items[6]);
|
|
2425 int itemcnt;
|
|
2426 char_u *p;
|
|
2427 int lnum = 0;
|
|
2428 affheader_T *cur_aff = NULL;
|
|
2429 int aff_todo = 0;
|
|
2430 hashtab_T *tp;
|
|
2431
|
|
2432 fd = fopen((char *)fname, "r");
|
|
2433 if (fd == NULL)
|
|
2434 {
|
|
2435 EMSG2(_(e_notopen), fname);
|
|
2436 return NULL;
|
|
2437 }
|
|
2438
|
|
2439 smsg((char_u *)_("Reading affix file %s..."), fname);
|
|
2440 out_flush();
|
|
2441
|
|
2442 aff = (afffile_T *)alloc_clear((unsigned)sizeof(afffile_T));
|
|
2443 if (aff == NULL)
|
|
2444 return NULL;
|
|
2445 hash_init(&aff->af_pref);
|
|
2446 hash_init(&aff->af_suff);
|
|
2447 ga_init2(&aff->af_rep, (int)sizeof(repentry_T), 20);
|
|
2448
|
|
2449 /*
|
|
2450 * Read all the lines in the file one by one.
|
|
2451 */
|
|
2452 while (!vim_fgets(rline, MAXLINELEN, fd))
|
|
2453 {
|
|
2454 ++lnum;
|
|
2455
|
|
2456 /* Skip comment lines. */
|
|
2457 if (*rline == '#')
|
|
2458 continue;
|
|
2459
|
|
2460 /* Convert from "SET" to 'encoding' when needed. */
|
|
2461 vim_free(pc);
|
|
2462 if (conv->vc_type != CONV_NONE)
|
|
2463 {
|
|
2464 pc = string_convert(conv, rline, NULL);
|
|
2465 line = pc;
|
|
2466 }
|
|
2467 else
|
|
2468 {
|
|
2469 pc = NULL;
|
|
2470 line = rline;
|
|
2471 }
|
|
2472
|
|
2473 /* Split the line up in white separated items. Put a NUL after each
|
|
2474 * item. */
|
|
2475 itemcnt = 0;
|
|
2476 for (p = line; ; )
|
|
2477 {
|
|
2478 while (*p != NUL && *p <= ' ') /* skip white space and CR/NL */
|
|
2479 ++p;
|
|
2480 if (*p == NUL)
|
|
2481 break;
|
|
2482 items[itemcnt++] = p;
|
|
2483 while (*p > ' ') /* skip until white space or CR/NL */
|
|
2484 ++p;
|
|
2485 if (*p == NUL)
|
|
2486 break;
|
|
2487 *p++ = NUL;
|
|
2488 }
|
|
2489
|
|
2490 /* Handle non-empty lines. */
|
|
2491 if (itemcnt > 0)
|
|
2492 {
|
|
2493 if (STRCMP(items[0], "SET") == 0 && itemcnt == 2
|
|
2494 && aff->af_enc == NULL)
|
|
2495 {
|
|
2496 if (aff->af_enc != NULL)
|
|
2497 smsg((char_u *)_("Duplicate SET line ignored in %s line %d: %s"),
|
|
2498 fname, lnum, line);
|
|
2499 else
|
|
2500 {
|
|
2501 /* Setup for conversion from "ENC" to 'encoding'. */
|
|
2502 aff->af_enc = enc_canonize(items[1]);
|
240
|
2503 if (aff->af_enc != NULL && !ascii
|
236
|
2504 && convert_setup(conv, aff->af_enc, p_enc) == FAIL)
|
|
2505 smsg((char_u *)_("Conversion in %s not supported: from %s to %s"),
|
|
2506 fname, aff->af_enc, p_enc);
|
|
2507 }
|
|
2508 }
|
|
2509 else if (STRCMP(items[0], "TRY") == 0 && itemcnt == 2
|
|
2510 && aff->af_try == NULL)
|
|
2511 aff->af_try = vim_strsave(items[1]);
|
|
2512 else if ((STRCMP(items[0], "PFX") == 0
|
|
2513 || STRCMP(items[0], "SFX") == 0)
|
|
2514 && aff_todo == 0
|
|
2515 && itemcnt == 4)
|
|
2516 {
|
|
2517 /* New affix letter. */
|
|
2518 cur_aff = (affheader_T *)alloc((unsigned)sizeof(affheader_T));
|
|
2519 if (cur_aff == NULL)
|
|
2520 break;
|
|
2521 cur_aff->ah_key[0] = *items[1];
|
|
2522 cur_aff->ah_key[1] = NUL;
|
|
2523 if (items[1][1] != NUL)
|
|
2524 smsg((char_u *)_("Affix name too long in %s line %d: %s"),
|
|
2525 fname, lnum, items[1]);
|
|
2526 if (*items[2] == 'Y')
|
|
2527 cur_aff->ah_combine = TRUE;
|
|
2528 else if (*items[2] == 'N')
|
|
2529 cur_aff->ah_combine = FALSE;
|
|
2530 else if (p_verbose > 0)
|
|
2531 smsg((char_u *)_("Expected Y or N in %s line %d: %s"),
|
|
2532 fname, lnum, items[2]);
|
|
2533 cur_aff->ah_first = NULL;
|
|
2534 if (*items[0] == 'P')
|
|
2535 tp = &aff->af_pref;
|
|
2536 else
|
|
2537 tp = &aff->af_suff;
|
|
2538 if (!HASHITEM_EMPTY(hash_find(tp, cur_aff->ah_key)))
|
|
2539 smsg((char_u *)_("Duplicate affix in %s line %d: %s"),
|
|
2540 fname, lnum, items[1]);
|
|
2541 else
|
|
2542 hash_add(tp, cur_aff->ah_key);
|
|
2543
|
|
2544 aff_todo = atoi((char *)items[3]);
|
|
2545 }
|
|
2546 else if ((STRCMP(items[0], "PFX") == 0
|
|
2547 || STRCMP(items[0], "SFX") == 0)
|
|
2548 && aff_todo > 0
|
|
2549 && STRCMP(cur_aff->ah_key, items[1]) == 0
|
|
2550 && itemcnt == 5)
|
|
2551 {
|
|
2552 affentry_T *aff_entry;
|
|
2553
|
|
2554 /* New item for an affix letter. */
|
|
2555 --aff_todo;
|
|
2556 aff_entry = (affentry_T *)alloc_clear(
|
|
2557 (unsigned)sizeof(affentry_T));
|
|
2558 if (aff_entry == NULL)
|
|
2559 break;
|
240
|
2560
|
236
|
2561 if (STRCMP(items[2], "0") != 0)
|
|
2562 aff_entry->ae_chop = vim_strsave(items[2]);
|
|
2563 if (STRCMP(items[3], "0") != 0)
|
|
2564 aff_entry->ae_add = vim_strsave(items[3]);
|
|
2565 if (STRCMP(items[4], ".") != 0)
|
|
2566 {
|
|
2567 char_u buf[MAXLINELEN];
|
|
2568
|
|
2569 aff_entry->ae_cond = vim_strsave(items[4]);
|
|
2570 if (*items[0] == 'P')
|
|
2571 sprintf((char *)buf, "^%s", items[4]);
|
|
2572 else
|
|
2573 sprintf((char *)buf, "%s$", items[4]);
|
|
2574 aff_entry->ae_prog = vim_regcomp(buf, RE_MAGIC + RE_STRING);
|
|
2575 }
|
240
|
2576
|
|
2577 if (ascii && (has_non_ascii(aff_entry->ae_chop)
|
|
2578 || has_non_ascii(aff_entry->ae_add)))
|
|
2579 {
|
|
2580 /* Don't use an affix entry with non-ASCII characters when
|
|
2581 * "ascii" is TRUE. */
|
|
2582 free_affix_entry(aff_entry);
|
|
2583 }
|
|
2584 else
|
|
2585 {
|
|
2586 aff_entry->ae_next = cur_aff->ah_first;
|
|
2587 cur_aff->ah_first = aff_entry;
|
|
2588 }
|
236
|
2589 }
|
|
2590 else if (STRCMP(items[0], "REP") == 0 && itemcnt == 2)
|
|
2591 /* Ignore REP count */;
|
|
2592 else if (STRCMP(items[0], "REP") == 0 && itemcnt == 3)
|
|
2593 {
|
|
2594 repentry_T *rp;
|
|
2595
|
|
2596 /* REP item */
|
|
2597 if (ga_grow(&aff->af_rep, 1) == FAIL)
|
|
2598 break;
|
|
2599 rp = ((repentry_T *)aff->af_rep.ga_data) + aff->af_rep.ga_len;
|
|
2600 rp->re_from = vim_strsave(items[1]);
|
|
2601 rp->re_to = vim_strsave(items[2]);
|
|
2602 ++aff->af_rep.ga_len;
|
|
2603 }
|
|
2604 else if (p_verbose > 0)
|
|
2605 smsg((char_u *)_("Unrecognized item in %s line %d: %s"),
|
|
2606 fname, lnum, items[0]);
|
|
2607 }
|
|
2608
|
|
2609 }
|
|
2610
|
|
2611 vim_free(pc);
|
|
2612 fclose(fd);
|
|
2613 return aff;
|
|
2614 }
|
|
2615
|
|
2616 /*
|
240
|
2617 * Return TRUE if string "s" contains a non-ASCII character (128 or higher).
|
|
2618 * When "s" is NULL FALSE is returned.
|
|
2619 */
|
|
2620 static int
|
|
2621 has_non_ascii(s)
|
|
2622 char_u *s;
|
|
2623 {
|
|
2624 char_u *p;
|
|
2625
|
|
2626 if (s != NULL)
|
|
2627 for (p = s; *p != NUL; ++p)
|
|
2628 if (*p >= 128)
|
|
2629 return TRUE;
|
|
2630 return FALSE;
|
|
2631 }
|
|
2632
|
|
2633 /*
|
236
|
2634 * Free the structure filled by spell_read_aff().
|
|
2635 */
|
|
2636 static void
|
|
2637 spell_free_aff(aff)
|
|
2638 afffile_T *aff;
|
|
2639 {
|
|
2640 hashtab_T *ht;
|
|
2641 hashitem_T *hi;
|
|
2642 int todo;
|
|
2643 int i;
|
|
2644 repentry_T *rp;
|
|
2645 affheader_T *ah;
|
|
2646
|
|
2647 vim_free(aff->af_enc);
|
|
2648 vim_free(aff->af_try);
|
|
2649
|
|
2650 for (ht = &aff->af_pref; ; ht = &aff->af_suff)
|
|
2651 {
|
|
2652 todo = ht->ht_used;
|
|
2653 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
2654 {
|
|
2655 if (!HASHITEM_EMPTY(hi))
|
|
2656 {
|
|
2657 --todo;
|
|
2658 ah = HI2AH(hi);
|
|
2659 free_affixentries(ah->ah_first);
|
|
2660 vim_free(ah);
|
|
2661 }
|
|
2662 }
|
|
2663 if (ht == &aff->af_suff)
|
|
2664 break;
|
|
2665 }
|
|
2666 hash_clear(&aff->af_pref);
|
|
2667 hash_clear(&aff->af_suff);
|
|
2668
|
|
2669 for (i = 0; i < aff->af_rep.ga_len; ++i)
|
|
2670 {
|
|
2671 rp = ((repentry_T *)aff->af_rep.ga_data) + i;
|
|
2672 vim_free(rp->re_from);
|
|
2673 vim_free(rp->re_to);
|
|
2674 }
|
|
2675 ga_clear(&aff->af_rep);
|
|
2676
|
|
2677 vim_free(aff);
|
|
2678 }
|
|
2679
|
|
2680 /*
|
|
2681 * Read a dictionary ".dic" file.
|
|
2682 * Returns OK or FAIL;
|
|
2683 * Each entry in the hashtab_T is a dicword_T.
|
|
2684 */
|
|
2685 static int
|
240
|
2686 spell_read_dic(ht, fname, conv, ascii)
|
236
|
2687 hashtab_T *ht;
|
|
2688 char_u *fname;
|
|
2689 vimconv_T *conv; /* info for encoding conversion */
|
240
|
2690 int ascii; /* only accept ASCII words */
|
236
|
2691 {
|
|
2692 char_u line[MAXLINELEN];
|
|
2693 char_u *p;
|
|
2694 dicword_T *dw;
|
|
2695 char_u *pc;
|
|
2696 char_u *w;
|
|
2697 int l;
|
|
2698 hash_T hash;
|
|
2699 hashitem_T *hi;
|
|
2700 FILE *fd;
|
|
2701 int lnum = 1;
|
|
2702
|
|
2703 fd = fopen((char *)fname, "r");
|
|
2704 if (fd == NULL)
|
|
2705 {
|
|
2706 EMSG2(_(e_notopen), fname);
|
|
2707 return FAIL;
|
|
2708 }
|
|
2709
|
|
2710 smsg((char_u *)_("Reading dictionary file %s..."), fname);
|
|
2711 out_flush();
|
|
2712
|
|
2713 /* Read and ignore the first line: word count. */
|
|
2714 (void)vim_fgets(line, MAXLINELEN, fd);
|
|
2715 if (!isdigit(*skipwhite(line)))
|
|
2716 EMSG2(_("E760: No word count in %s"), fname);
|
|
2717
|
|
2718 /*
|
|
2719 * Read all the lines in the file one by one.
|
|
2720 * The words are converted to 'encoding' here, before being added to
|
|
2721 * the hashtable.
|
|
2722 */
|
|
2723 while (!vim_fgets(line, MAXLINELEN, fd))
|
|
2724 {
|
|
2725 ++lnum;
|
|
2726
|
|
2727 /* Remove CR, LF and white space from end. */
|
|
2728 l = STRLEN(line);
|
|
2729 while (l > 0 && line[l - 1] <= ' ')
|
|
2730 --l;
|
|
2731 if (l == 0)
|
|
2732 continue; /* empty line */
|
|
2733 line[l] = NUL;
|
|
2734
|
|
2735 /* Find the optional affix names. */
|
|
2736 p = vim_strchr(line, '/');
|
|
2737 if (p != NULL)
|
|
2738 *p++ = NUL;
|
|
2739
|
240
|
2740 /* Skip non-ASCII words when "ascii" is TRUE. */
|
|
2741 if (ascii && has_non_ascii(line))
|
|
2742 continue;
|
|
2743
|
236
|
2744 /* Convert from "SET" to 'encoding' when needed. */
|
|
2745 if (conv->vc_type != CONV_NONE)
|
|
2746 {
|
|
2747 pc = string_convert(conv, line, NULL);
|
|
2748 w = pc;
|
|
2749 }
|
|
2750 else
|
|
2751 {
|
|
2752 pc = NULL;
|
|
2753 w = line;
|
|
2754 }
|
|
2755
|
|
2756 dw = (dicword_T *)alloc_clear((unsigned)sizeof(dicword_T)
|
|
2757 + STRLEN(w));
|
|
2758 if (dw == NULL)
|
|
2759 break;
|
|
2760 STRCPY(dw->dw_word, w);
|
|
2761 vim_free(pc);
|
|
2762
|
|
2763 hash = hash_hash(dw->dw_word);
|
|
2764 hi = hash_lookup(ht, dw->dw_word, hash);
|
|
2765 if (!HASHITEM_EMPTY(hi))
|
|
2766 smsg((char_u *)_("Duplicate word in %s line %d: %s"),
|
|
2767 fname, lnum, line);
|
|
2768 else
|
|
2769 hash_add_item(ht, hi, dw->dw_word, hash);
|
|
2770
|
|
2771 if (p != NULL)
|
|
2772 dw->dw_affnm = vim_strsave(p);
|
|
2773 }
|
|
2774
|
|
2775 fclose(fd);
|
|
2776 return OK;
|
|
2777 }
|
|
2778
|
|
2779 /*
|
|
2780 * Free the structure filled by spell_read_dic().
|
|
2781 */
|
|
2782 static void
|
|
2783 spell_free_dic(dic)
|
|
2784 hashtab_T *dic;
|
|
2785 {
|
|
2786 int todo;
|
|
2787 dicword_T *dw;
|
|
2788 hashitem_T *hi;
|
|
2789
|
|
2790 todo = dic->ht_used;
|
|
2791 for (hi = dic->ht_array; todo > 0; ++hi)
|
|
2792 {
|
|
2793 if (!HASHITEM_EMPTY(hi))
|
|
2794 {
|
|
2795 --todo;
|
|
2796 dw = HI2DW(hi);
|
|
2797 vim_free(dw->dw_affnm);
|
|
2798 vim_free(dw);
|
|
2799 }
|
|
2800 }
|
|
2801 hash_clear(dic);
|
|
2802 }
|
|
2803
|
|
2804 /*
|
|
2805 * Take the affixes read by spell_read_aff() and add them to the new list.
|
|
2806 * Attempts to re-use the same number for identical affixes (ignoring the
|
|
2807 * condition, since we remove that). That is especially important when using
|
|
2808 * multiple regions.
|
|
2809 * Returns OK or FAIL;
|
|
2810 */
|
|
2811 static int
|
240
|
2812 get_new_aff(oldaff, gap, prefix)
|
236
|
2813 hashtab_T *oldaff; /* hashtable with affheader_T */
|
|
2814 garray_T *gap; /* table with new affixes */
|
240
|
2815 int prefix; /* TRUE when doing prefixes, FALSE for
|
|
2816 suffixes */
|
236
|
2817 {
|
|
2818 int oldtodo;
|
|
2819 affheader_T *oldah, *newah, *gapah;
|
|
2820 affentry_T *oldae, *newae;
|
|
2821 hashitem_T *oldhi;
|
|
2822 hashitem_T *hi;
|
|
2823 hashtab_T condht; /* conditions already found */
|
|
2824 char_u condkey[MAXLINELEN];
|
|
2825 int newnr;
|
|
2826 int gapnr;
|
|
2827 int retval = OK;
|
|
2828 char_u *p;
|
|
2829 garray_T tga;
|
243
|
2830 int preword;
|
236
|
2831
|
|
2832 /*
|
|
2833 * Loop over all the old affix names.
|
|
2834 */
|
|
2835 oldtodo = oldaff->ht_used;
|
|
2836 for (oldhi = oldaff->ht_array; oldtodo > 0 && retval == OK; ++oldhi)
|
|
2837 {
|
|
2838 if (!HASHITEM_EMPTY(oldhi))
|
|
2839 {
|
|
2840 --oldtodo;
|
|
2841 oldah = (affheader_T *)oldhi->hi_key;
|
|
2842
|
|
2843 /* Put entries with the same condition under the same new affix
|
|
2844 * nr in "tga". Use hashtable "condht" to find them. */
|
|
2845 ga_init2(&tga, sizeof(affheader_T), 10);
|
|
2846 hash_init(&condht);
|
|
2847
|
|
2848 /*
|
|
2849 * Loop over all affixes with the same name.
|
|
2850 * The affixes with the same condition will get the same number,
|
|
2851 * since they can be used with the same words.
|
|
2852 * 1. build the lists of new affentry_T, with the headers in "tga".
|
|
2853 * 2. Check if some of the lists already exist in "gap", re-use
|
|
2854 * their number.
|
|
2855 * 3. Assign the new numbers to the old affixes.
|
|
2856 */
|
|
2857
|
|
2858 /* 1. build the lists of new affentry_T. */
|
|
2859 for (oldae = oldah->ah_first; oldae != NULL && retval == OK;
|
|
2860 oldae = oldae->ae_next)
|
|
2861 {
|
243
|
2862 preword = FALSE;
|
236
|
2863 oldae->ae_add_nw = NULL;
|
240
|
2864 oldae->ae_add_pw = NULL;
|
236
|
2865 if (oldae->ae_add != NULL)
|
|
2866 {
|
240
|
2867 /* Check for non-word characters in the affix. If there
|
243
|
2868 * is one a suffix will be turned into an addition, a
|
|
2869 * prefix may be turned into a leadstring.
|
236
|
2870 * This is stored with the old affix, that is where
|
|
2871 * trans_affixes() will check. */
|
|
2872 for (p = oldae->ae_add; *p != NUL; mb_ptr_adv(p))
|
|
2873 if (!spell_iswordc(p))
|
240
|
2874 {
|
|
2875 oldae->ae_add_nw = p;
|
236
|
2876 break;
|
240
|
2877 }
|
|
2878
|
|
2879 if (prefix && oldae->ae_add_nw != NULL)
|
|
2880 {
|
243
|
2881 /* If a prefix has non-word characters special
|
|
2882 * treatment is necessary:
|
|
2883 * - If it has only non-word characters it becomes a
|
|
2884 * leadstring.
|
|
2885 * - If it has a sequence of word characters followed
|
|
2886 * by a non-word char it becomes a "preword": "d'",
|
|
2887 * "de-", "d'ai", etc.
|
|
2888 * - if it has another mix of word and non-word
|
|
2889 * characters the part before the last word char
|
|
2890 * becomes a leadstring: "'d", etc.
|
|
2891 */
|
240
|
2892 for (p = oldae->ae_add; *p != NUL; mb_ptr_adv(p))
|
|
2893 if (spell_iswordc(p))
|
|
2894 {
|
|
2895 oldae->ae_add_pw = p;
|
|
2896 break;
|
|
2897 }
|
|
2898 if (oldae->ae_add_pw != NULL)
|
|
2899 {
|
|
2900 /* Mixed prefix, set ae_add_nw to first non-word
|
|
2901 * char after ae_add_pw (if there is one). */
|
|
2902 oldae->ae_add_nw = NULL;
|
|
2903 for ( ; *p != NUL; mb_ptr_adv(p))
|
|
2904 if (!spell_iswordc(p))
|
|
2905 {
|
|
2906 oldae->ae_add_nw = p;
|
|
2907 break;
|
|
2908 }
|
243
|
2909 if (oldae->ae_add_nw != NULL)
|
|
2910 {
|
|
2911 preword = TRUE;
|
|
2912 oldae->ae_add_pw = NULL;
|
|
2913 oldae->ae_add_nw = NULL;
|
|
2914 }
|
240
|
2915 }
|
|
2916 }
|
236
|
2917 }
|
|
2918
|
|
2919 if (oldae->ae_cond == NULL)
|
|
2920 /* hashtable requires a non-empty key */
|
|
2921 STRCPY(condkey, "---");
|
|
2922 else
|
|
2923 STRCPY(condkey, oldae->ae_cond);
|
|
2924
|
|
2925 /* Look for an existing list with this name and condition. */
|
|
2926 hi = hash_find(&condht, condkey);
|
|
2927 if (!HASHITEM_EMPTY(hi))
|
|
2928 /* Match with existing affix, use that one. */
|
|
2929 newnr = HI2AS(hi)->as_nr;
|
|
2930 else
|
|
2931 {
|
|
2932 /* Add a new affix number. */
|
|
2933 newnr = tga.ga_len;
|
|
2934 if (ga_grow(&tga, 1) == FAIL)
|
|
2935 retval = FAIL;
|
|
2936 else
|
|
2937 {
|
|
2938 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2939 newah->ah_combine = oldah->ah_combine;
|
|
2940 newah->ah_first = NULL;
|
|
2941 ++tga.ga_len;
|
|
2942
|
|
2943 /* Add the new list to the condht hashtable. */
|
|
2944 add_affhash(&condht, condkey, newnr);
|
|
2945 }
|
|
2946 }
|
|
2947
|
|
2948 /* Add the new affentry_T to the list. */
|
|
2949 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2950 newae = (affentry_T *)alloc_clear((unsigned)sizeof(affentry_T));
|
|
2951 if (newae == NULL)
|
|
2952 retval = FAIL;
|
|
2953 else
|
|
2954 {
|
|
2955 newae->ae_next = newah->ah_first;
|
|
2956 newah->ah_first = newae;
|
|
2957 if (oldae->ae_chop == NULL)
|
|
2958 newae->ae_chop = NULL;
|
|
2959 else
|
|
2960 newae->ae_chop = vim_strsave(oldae->ae_chop);
|
|
2961 if (oldae->ae_add == NULL)
|
|
2962 newae->ae_add = NULL;
|
|
2963 else
|
|
2964 newae->ae_add = vim_strsave(oldae->ae_add);
|
243
|
2965 newae->ae_preword = preword;
|
236
|
2966
|
|
2967 /* The condition is not copied, since the new affix is
|
|
2968 * only used for words where the condition matches. */
|
|
2969 }
|
|
2970 }
|
|
2971
|
|
2972 /* 2. Check if some of the lists already exist, re-use their
|
|
2973 * number. Otherwise add the list to "gap". */
|
|
2974 for (newnr = 0; newnr < tga.ga_len; ++newnr)
|
|
2975 {
|
|
2976 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
2977 for (gapnr = 0; gapnr < gap->ga_len; ++gapnr)
|
|
2978 {
|
|
2979 gapah = ((affheader_T *)gap->ga_data) + gapnr;
|
243
|
2980 if (newah->ah_combine == gapah->ah_combine
|
|
2981 && same_affentries(newah, gapah))
|
236
|
2982 /* Found an existing affheader_T entry with same
|
|
2983 * affentry_T list, use its number. */
|
|
2984 break;
|
|
2985 }
|
|
2986
|
|
2987 newah->ah_affnr = gapnr;
|
|
2988 if (gapnr == gap->ga_len)
|
|
2989 {
|
|
2990 /* This is a new affentry_T list, add it. */
|
|
2991 if (ga_grow(gap, 1) == FAIL)
|
|
2992 retval = FAIL;
|
|
2993 else
|
|
2994 {
|
|
2995 *(((affheader_T *)gap->ga_data) + gap->ga_len) = *newah;
|
|
2996 ++gap->ga_len;
|
|
2997 }
|
|
2998 }
|
|
2999 else
|
|
3000 {
|
|
3001 /* free unused affentry_T list */
|
|
3002 free_affixentries(newah->ah_first);
|
|
3003 }
|
|
3004 }
|
|
3005
|
|
3006 /* 3. Assign the new affix numbers to the old affixes. */
|
|
3007 for (oldae = oldah->ah_first; oldae != NULL && retval == OK;
|
|
3008 oldae = oldae->ae_next)
|
|
3009 {
|
|
3010 if (oldae->ae_cond == NULL)
|
|
3011 /* hashtable requires a non-empty key */
|
|
3012 STRCPY(condkey, "---");
|
|
3013 else
|
|
3014 STRCPY(condkey, oldae->ae_cond);
|
|
3015
|
|
3016 /* Look for an existing affix with this name and condition. */
|
|
3017 hi = hash_find(&condht, condkey);
|
|
3018 if (!HASHITEM_EMPTY(hi))
|
|
3019 /* Match with existing affix, use that one. */
|
|
3020 newnr = HI2AS(hi)->as_nr;
|
|
3021 else
|
|
3022 {
|
|
3023 EMSG(_(e_internal));
|
|
3024 retval = FAIL;
|
|
3025 }
|
|
3026 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
3027 oldae->ae_affnr = newah->ah_affnr;
|
|
3028 }
|
|
3029
|
|
3030 ga_clear(&tga);
|
|
3031 clear_affhash(&condht);
|
|
3032 }
|
|
3033 }
|
|
3034
|
|
3035 return retval;
|
|
3036 }
|
|
3037
|
|
3038 /*
|
|
3039 * Return TRUE if the affentry_T lists for "ah1" and "ah2" contain the same
|
|
3040 * items, ignoring the order.
|
|
3041 * Only compares the chop and add strings, not the condition.
|
|
3042 */
|
|
3043 static int
|
|
3044 same_affentries(ah1, ah2)
|
|
3045 affheader_T *ah1;
|
|
3046 affheader_T *ah2;
|
|
3047 {
|
|
3048 affentry_T *ae1, *ae2;
|
|
3049
|
|
3050 /* Check the length of the lists first. */
|
|
3051 ae2 = ah2->ah_first;
|
|
3052 for (ae1 = ah1->ah_first; ae1 != NULL; ae1 = ae1->ae_next)
|
|
3053 {
|
|
3054 if (ae2 == NULL)
|
|
3055 return FALSE; /* "ah1" list is longer */
|
|
3056 ae2 = ae2->ae_next;
|
|
3057 }
|
|
3058 if (ae2 != NULL)
|
|
3059 return FALSE; /* "ah2" list is longer */
|
|
3060
|
|
3061 /* Check that each entry in "ah1" appears in "ah2". */
|
|
3062 for (ae1 = ah1->ah_first; ae1 != NULL; ae1 = ae1->ae_next)
|
|
3063 {
|
|
3064 for (ae2 = ah2->ah_first; ae2 != NULL; ae2 = ae2->ae_next)
|
|
3065 {
|
|
3066 if ((ae1->ae_chop == NULL) == (ae2->ae_chop == NULL)
|
|
3067 && (ae1->ae_add == NULL) == (ae2->ae_add == NULL)
|
|
3068 && (ae1->ae_chop == NULL
|
|
3069 || STRCMP(ae1->ae_chop, ae2->ae_chop) == 0)
|
|
3070 && (ae1->ae_add == NULL
|
|
3071 || STRCMP(ae1->ae_add, ae2->ae_add) == 0))
|
|
3072 break;
|
|
3073 }
|
|
3074 if (ae2 == NULL)
|
|
3075 return FALSE;
|
|
3076 }
|
|
3077
|
|
3078 return TRUE;
|
|
3079 }
|
|
3080
|
|
3081 /*
|
|
3082 * Add a chop/add or cond hashtable entry.
|
|
3083 */
|
|
3084 static void
|
|
3085 add_affhash(ht, key, newnr)
|
|
3086 hashtab_T *ht;
|
|
3087 char_u *key;
|
|
3088 int newnr;
|
|
3089 {
|
|
3090 affhash_T *as;
|
|
3091
|
|
3092 as = (affhash_T *)alloc((unsigned)sizeof(affhash_T) + STRLEN(key));
|
|
3093 if (as != NULL)
|
|
3094 {
|
|
3095 as->as_nr = newnr;
|
|
3096 STRCPY(as->as_word, key);
|
|
3097 hash_add(ht, as->as_word);
|
|
3098 }
|
|
3099 }
|
|
3100
|
|
3101 /*
|
|
3102 * Clear the chop/add hashtable used to detect identical affixes.
|
|
3103 */
|
|
3104 static void
|
|
3105 clear_affhash(ht)
|
|
3106 hashtab_T *ht;
|
|
3107 {
|
|
3108 int todo;
|
|
3109 hashitem_T *hi;
|
|
3110
|
|
3111 todo = ht->ht_used;
|
|
3112 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
3113 {
|
|
3114 if (!HASHITEM_EMPTY(hi))
|
|
3115 {
|
|
3116 --todo;
|
|
3117 vim_free(HI2AS(hi));
|
|
3118 }
|
|
3119 }
|
|
3120 hash_clear(ht);
|
|
3121 }
|
|
3122
|
|
3123 /*
|
|
3124 * Translate list of affix names for an old word to affix numbers in a new
|
|
3125 * basic word.
|
|
3126 * This checks if the conditions match with the old word. The result is that
|
|
3127 * the new affix does not need to store the condition.
|
|
3128 */
|
|
3129 static void
|
|
3130 trans_affixes(dw, bw, oldaff, newwords)
|
|
3131 dicword_T *dw; /* old word */
|
|
3132 basicword_T *bw; /* basic word */
|
|
3133 afffile_T *oldaff; /* affixes for "oldwords" */
|
|
3134 hashtab_T *newwords; /* table with words */
|
|
3135 {
|
|
3136 char_u key[2];
|
|
3137 char_u *p;
|
|
3138 char_u *affnm;
|
|
3139 garray_T *gap;
|
|
3140 hashitem_T *aff_hi;
|
|
3141 affheader_T *ah;
|
|
3142 affentry_T *ae;
|
|
3143 regmatch_T regmatch;
|
|
3144 int i;
|
|
3145 basicword_T *nbw;
|
|
3146 int alen;
|
|
3147 int wlen;
|
240
|
3148 garray_T suffixga; /* list of words with non-word suffixes */
|
|
3149 garray_T prefixga; /* list of words with non-word prefixes */
|
236
|
3150 char_u nword[MAXWLEN];
|
|
3151 int flags;
|
|
3152 int n;
|
|
3153
|
240
|
3154 ga_init2(&suffixga, (int)sizeof(basicword_T *), 5);
|
|
3155 ga_init2(&prefixga, (int)sizeof(basicword_T *), 5);
|
236
|
3156
|
|
3157 /* Loop over all the affix names of the old word. */
|
|
3158 key[1] = NUL;
|
|
3159 for (affnm = dw->dw_affnm; *affnm != NUL; ++affnm)
|
|
3160 {
|
|
3161 key[0] = *affnm;
|
|
3162 aff_hi = hash_find(&oldaff->af_pref, key);
|
|
3163 if (!HASHITEM_EMPTY(aff_hi))
|
|
3164 gap = &bw->bw_prefix; /* found a prefix */
|
|
3165 else
|
|
3166 {
|
|
3167 gap = &bw->bw_suffix; /* must be a suffix */
|
|
3168 aff_hi = hash_find(&oldaff->af_suff, key);
|
|
3169 if (HASHITEM_EMPTY(aff_hi))
|
|
3170 {
|
|
3171 smsg((char_u *)_("No affix entry '%s' for word %s"),
|
|
3172 key, dw->dw_word);
|
|
3173 continue;
|
|
3174 }
|
|
3175 }
|
|
3176
|
|
3177 /* Loop over all the affix entries for this affix name. */
|
|
3178 ah = HI2AH(aff_hi);
|
|
3179 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
|
3180 {
|
243
|
3181 /* Setup for regexp matching. Note that we don't ignore case.
|
|
3182 * This is weird, because he rules in an .aff file don't care
|
|
3183 * about case, but it's necessary for compatibility with Myspell.
|
|
3184 */
|
236
|
3185 regmatch.regprog = ae->ae_prog;
|
243
|
3186 regmatch.rm_ic = FALSE;
|
236
|
3187 if (ae->ae_prog == NULL
|
|
3188 || vim_regexec(®match, dw->dw_word, (colnr_T)0))
|
|
3189 {
|
240
|
3190 if ((ae->ae_add_nw != NULL || ae->ae_add_pw != NULL)
|
|
3191 && (gap != &bw->bw_suffix || bw->bw_addstring == NULL))
|
236
|
3192 {
|
|
3193 /* Affix has a non-word character and isn't prepended to
|
|
3194 * leader or appended to addition. Need to use another
|
|
3195 * word with an addition. It's a copy of the basicword_T
|
|
3196 * "bw". */
|
|
3197 if (gap == &bw->bw_suffix)
|
|
3198 {
|
|
3199 alen = ae->ae_add_nw - ae->ae_add;
|
|
3200 nbw = (basicword_T *)alloc((unsigned)(
|
|
3201 sizeof(basicword_T) + STRLEN(bw->bw_word)
|
|
3202 + alen + 1));
|
|
3203 if (nbw != NULL)
|
|
3204 {
|
|
3205 *nbw = *bw;
|
|
3206 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3207 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3208
|
|
3209 /* Adding the suffix may change the caps. */
|
|
3210 STRCPY(nword, dw->dw_word);
|
|
3211 if (ae->ae_chop != NULL)
|
|
3212 {
|
|
3213 /* Remove chop string. */
|
|
3214 p = nword + STRLEN(nword);
|
|
3215 for (i = mb_charlen(ae->ae_chop); i > 0; --i)
|
|
3216 mb_ptr_back(nword, p);
|
|
3217 *p = NUL;
|
|
3218 }
|
|
3219 STRCAT(nword, ae->ae_add);
|
|
3220 flags = captype(nword, nword + STRLEN(nword));
|
|
3221 if (flags & BWF_KEEPCAP)
|
|
3222 {
|
240
|
3223 /* "caseword" excludes the addition */
|
236
|
3224 nword[STRLEN(dw->dw_word) + alen] = NUL;
|
|
3225 nbw->bw_caseword = vim_strsave(nword);
|
|
3226 }
|
|
3227 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3228 | BWF_KEEPCAP);
|
|
3229 nbw->bw_flags |= flags;
|
|
3230
|
|
3231 if (bw->bw_leadstring != NULL)
|
|
3232 nbw->bw_leadstring =
|
|
3233 vim_strsave(bw->bw_leadstring);
|
|
3234 nbw->bw_addstring = vim_strsave(ae->ae_add_nw);
|
|
3235
|
|
3236 STRCPY(nbw->bw_word, bw->bw_word);
|
|
3237 if (alen > 0 || ae->ae_chop != NULL)
|
|
3238 {
|
240
|
3239 /* Suffix starts with word character and/or
|
|
3240 * chop off something. Append it to the word.
|
|
3241 * Add new word entry. */
|
236
|
3242 wlen = STRLEN(nbw->bw_word);
|
|
3243 if (ae->ae_chop != NULL)
|
|
3244 wlen -= STRLEN(ae->ae_chop);
|
|
3245 mch_memmove(nbw->bw_word + wlen, ae->ae_add,
|
|
3246 alen);
|
|
3247 nbw->bw_word[wlen + alen] = NUL;
|
|
3248 add_to_wordlist(newwords, nbw);
|
|
3249 }
|
|
3250 else
|
|
3251 /* Basic word is the same, link "nbw" after
|
|
3252 * "bw". */
|
|
3253 bw->bw_next = nbw;
|
|
3254
|
|
3255 /* Remember this word, we need to set bw_prefix
|
240
|
3256 * and bw_prefix later. */
|
|
3257 if (ga_grow(&suffixga, 1) == OK)
|
|
3258 ((basicword_T **)suffixga.ga_data)
|
|
3259 [suffixga.ga_len++] = nbw;
|
|
3260 }
|
|
3261 }
|
|
3262 else if (ae->ae_add_nw == NULL)
|
|
3263 {
|
|
3264 /* Prefix that starts with non-word char(s) and may be
|
|
3265 * followed by word chars: Make a leadstring and
|
|
3266 * prepend word chars before the word. */
|
|
3267 alen = STRLEN(ae->ae_add_pw);
|
|
3268 nbw = (basicword_T *)alloc((unsigned)(
|
|
3269 sizeof(basicword_T) + STRLEN(bw->bw_word)
|
|
3270 + alen + 1));
|
|
3271 if (nbw != NULL)
|
|
3272 {
|
|
3273 *nbw = *bw;
|
|
3274 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3275 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3276
|
|
3277 /* Adding the prefix may change the caps. */
|
|
3278 STRCPY(nword, ae->ae_add);
|
|
3279 p = dw->dw_word;
|
|
3280 if (ae->ae_chop != NULL)
|
|
3281 /* Skip chop string. */
|
|
3282 for (i = mb_charlen(ae->ae_chop); i > 0; --i)
|
|
3283 mb_ptr_adv( p);
|
|
3284 STRCAT(nword, p);
|
|
3285
|
|
3286 flags = captype(nword, nword + STRLEN(nword));
|
|
3287 if (flags & BWF_KEEPCAP)
|
|
3288 /* "caseword" excludes the addition */
|
|
3289 nbw->bw_caseword = vim_strsave(nword
|
|
3290 + (ae->ae_add_pw - ae->ae_add));
|
|
3291 else
|
|
3292 nbw->bw_caseword = NULL;
|
|
3293 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3294 | BWF_KEEPCAP);
|
|
3295 nbw->bw_flags |= flags;
|
|
3296
|
|
3297 if (bw->bw_addstring != NULL)
|
|
3298 nbw->bw_addstring =
|
|
3299 vim_strsave(bw->bw_addstring);
|
|
3300 else
|
|
3301 nbw->bw_addstring = NULL;
|
|
3302 nbw->bw_leadstring = vim_strnsave(ae->ae_add,
|
|
3303 ae->ae_add_pw - ae->ae_add);
|
|
3304
|
|
3305 if (alen > 0 || ae->ae_chop != NULL)
|
|
3306 {
|
|
3307 /* Prefix ends in word character and/or chop
|
|
3308 * off something. Prepend it to the word.
|
|
3309 * Add new word entry. */
|
|
3310 STRCPY(nbw->bw_word, ae->ae_add_pw);
|
|
3311 p = bw->bw_word;
|
|
3312 if (ae->ae_chop != NULL)
|
|
3313 p += STRLEN(ae->ae_chop);
|
|
3314 STRCAT(nbw->bw_word, p);
|
|
3315 add_to_wordlist(newwords, nbw);
|
|
3316 }
|
|
3317 else
|
|
3318 {
|
|
3319 /* Basic word is the same, link "nbw" after
|
|
3320 * "bw". */
|
|
3321 STRCPY(nbw->bw_word, bw->bw_word);
|
|
3322 bw->bw_next = nbw;
|
|
3323 }
|
|
3324
|
|
3325 /* Remember this word, we need to set bw_suffix
|
236
|
3326 * and bw_suffix later. */
|
240
|
3327 if (ga_grow(&prefixga, 1) == OK)
|
|
3328 ((basicword_T **)prefixga.ga_data)
|
|
3329 [prefixga.ga_len++] = nbw;
|
236
|
3330 }
|
|
3331 }
|
|
3332 else
|
|
3333 {
|
240
|
3334 /* Prefix with both non-word and word characters: Turn
|
|
3335 * prefix into basic word, original word becomes an
|
|
3336 * addstring. */
|
|
3337
|
|
3338 /* Fold-case the word characters in the prefix into
|
|
3339 * nword[]. */
|
|
3340 alen = 0;
|
|
3341 for (p = ae->ae_add_pw; p < ae->ae_add_nw; p += n)
|
|
3342 {
|
|
3343 #ifdef FEAT_MBYTE
|
|
3344 n = (*mb_ptr2len_check)(p);
|
|
3345 #else
|
|
3346 n = 1;
|
|
3347 #endif
|
|
3348 (void)str_foldcase(p, n, nword + alen,
|
|
3349 MAXWLEN - alen);
|
|
3350 alen += STRLEN(nword + alen);
|
|
3351 }
|
|
3352
|
|
3353 /* Allocate a new word entry. */
|
|
3354 nbw = (basicword_T *)alloc((unsigned)(
|
|
3355 sizeof(basicword_T) + alen + 1));
|
|
3356 if (nbw != NULL)
|
|
3357 {
|
|
3358 *nbw = *bw;
|
|
3359 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3360 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3361
|
|
3362 mch_memmove(nbw->bw_word, nword, alen);
|
|
3363 nbw->bw_word[alen] = NUL;
|
|
3364
|
|
3365 /* Use the cap type of the prefix. */
|
|
3366 alen = ae->ae_add_nw - ae->ae_add_pw;
|
|
3367 mch_memmove(nword, ae->ae_add_pw, alen);
|
|
3368 nword[alen] = NUL;
|
|
3369 flags = captype(nword, nword + STRLEN(nword));
|
|
3370 if (flags & BWF_KEEPCAP)
|
|
3371 nbw->bw_caseword = vim_strsave(nword);
|
|
3372 else
|
|
3373 nbw->bw_caseword = NULL;
|
|
3374 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3375 | BWF_KEEPCAP);
|
|
3376 nbw->bw_flags |= flags;
|
|
3377
|
|
3378 /* The addstring is the prefix after the word
|
|
3379 * characters, the original word excluding "chop",
|
|
3380 * plus any addition. */
|
|
3381 STRCPY(nword, ae->ae_add_nw);
|
|
3382 p = bw->bw_word;
|
|
3383 if (ae->ae_chop != NULL)
|
|
3384 p += STRLEN(ae->ae_chop);
|
|
3385 STRCAT(nword, p);
|
|
3386 if (bw->bw_addstring != NULL)
|
|
3387 STRCAT(nword, bw->bw_addstring);
|
|
3388 nbw->bw_addstring = vim_strsave(nword);
|
|
3389
|
|
3390 if (ae->ae_add_pw > ae->ae_add)
|
|
3391 nbw->bw_leadstring = vim_strnsave(ae->ae_add,
|
|
3392 ae->ae_add_pw - ae->ae_add);
|
|
3393 else
|
|
3394 nbw->bw_leadstring = NULL;
|
|
3395
|
|
3396 add_to_wordlist(newwords, nbw);
|
|
3397
|
|
3398 /* Remember this word, we need to set bw_suffix
|
|
3399 * and bw_suffix later. */
|
|
3400 if (ga_grow(&prefixga, 1) == OK)
|
|
3401 ((basicword_T **)prefixga.ga_data)
|
|
3402 [prefixga.ga_len++] = nbw;
|
|
3403 }
|
236
|
3404 }
|
|
3405 }
|
|
3406 else
|
|
3407 {
|
|
3408 /* Affix applies to this word, add the related affix
|
|
3409 * number. But only if it's not there yet. And keep the
|
|
3410 * list sorted, so that we can compare it later. */
|
|
3411 for (i = 0; i < gap->ga_len; ++i)
|
|
3412 {
|
|
3413 n = ((short_u *)gap->ga_data)[i];
|
|
3414 if (n >= ae->ae_affnr)
|
|
3415 {
|
|
3416 if (n == ae->ae_affnr)
|
|
3417 i = -1;
|
|
3418 break;
|
|
3419 }
|
|
3420 }
|
|
3421 if (i >= 0 && ga_grow(gap, 1) == OK)
|
|
3422 {
|
|
3423 if (i < gap->ga_len)
|
|
3424 mch_memmove(((short_u *)gap->ga_data) + i + 1,
|
|
3425 ((short_u *)gap->ga_data) + i,
|
|
3426 sizeof(short_u) * (gap->ga_len - i));
|
|
3427 ((short_u *)gap->ga_data)[i] = ae->ae_affnr;
|
|
3428 ++gap->ga_len;
|
|
3429 }
|
|
3430 }
|
|
3431 }
|
|
3432 }
|
|
3433 }
|
|
3434
|
|
3435 /*
|
|
3436 * For the words that we added for suffixes with non-word characters: Use
|
|
3437 * the prefix list of the main word.
|
|
3438 */
|
240
|
3439 for (i = 0; i < suffixga.ga_len; ++i)
|
236
|
3440 {
|
240
|
3441 nbw = ((basicword_T **)suffixga.ga_data)[i];
|
236
|
3442 if (ga_grow(&nbw->bw_prefix, bw->bw_prefix.ga_len) == OK)
|
|
3443 {
|
|
3444 mch_memmove(nbw->bw_prefix.ga_data, bw->bw_prefix.ga_data,
|
|
3445 bw->bw_prefix.ga_len * sizeof(short_u));
|
|
3446 nbw->bw_prefix.ga_len = bw->bw_prefix.ga_len;
|
|
3447 }
|
|
3448 }
|
|
3449
|
240
|
3450 /*
|
|
3451 * For the words that we added for prefixes with non-word characters: Use
|
|
3452 * the suffix list of the main word.
|
|
3453 */
|
|
3454 for (i = 0; i < prefixga.ga_len; ++i)
|
|
3455 {
|
|
3456 nbw = ((basicword_T **)prefixga.ga_data)[i];
|
|
3457 if (ga_grow(&nbw->bw_suffix, bw->bw_suffix.ga_len) == OK)
|
|
3458 {
|
|
3459 mch_memmove(nbw->bw_suffix.ga_data, bw->bw_suffix.ga_data,
|
|
3460 bw->bw_suffix.ga_len * sizeof(short_u));
|
|
3461 nbw->bw_suffix.ga_len = bw->bw_suffix.ga_len;
|
|
3462 }
|
|
3463 }
|
|
3464
|
|
3465 ga_clear(&suffixga);
|
|
3466 ga_clear(&prefixga);
|
236
|
3467 }
|
|
3468
|
|
3469 /*
|
|
3470 * Go over all words in "oldwords" and change the old affix names to the new
|
|
3471 * affix numbers, check the conditions, fold case, extract the basic word and
|
|
3472 * additions.
|
|
3473 */
|
|
3474 static int
|
|
3475 build_wordlist(newwords, oldwords, oldaff, regionmask)
|
|
3476 hashtab_T *newwords; /* basicword_T entries */
|
|
3477 hashtab_T *oldwords; /* dicword_T entries */
|
|
3478 afffile_T *oldaff; /* affixes for "oldwords" */
|
|
3479 int regionmask; /* value for bw_region */
|
|
3480 {
|
|
3481 int todo;
|
|
3482 hashitem_T *old_hi;
|
|
3483 dicword_T *dw;
|
|
3484 basicword_T *bw;
|
|
3485 char_u foldword[MAXLINELEN];
|
|
3486 int leadlen;
|
|
3487 char_u leadstring[MAXLINELEN];
|
|
3488 int addlen;
|
|
3489 char_u addstring[MAXLINELEN];
|
|
3490 int dwlen;
|
|
3491 char_u *p;
|
|
3492 int clen;
|
|
3493 int flags;
|
240
|
3494 char_u *cp = NULL;
|
236
|
3495 int l;
|
240
|
3496 char_u message[MAXLINELEN + MAXWLEN];
|
236
|
3497
|
|
3498 todo = oldwords->ht_used;
|
|
3499 for (old_hi = oldwords->ht_array; todo > 0; ++old_hi)
|
|
3500 {
|
|
3501 if (!HASHITEM_EMPTY(old_hi))
|
|
3502 {
|
|
3503 --todo;
|
|
3504 dw = HI2DW(old_hi);
|
|
3505
|
|
3506 /* This takes time, print a message now and then. */
|
240
|
3507 if ((todo & 0x3ff) == 0 || todo == (int)oldwords->ht_used - 1)
|
236
|
3508 {
|
240
|
3509 sprintf((char *)message, _("%6d todo - %s"),
|
|
3510 todo, dw->dw_word);
|
|
3511 msg_start();
|
|
3512 msg_outtrans_attr(message, 0);
|
|
3513 msg_clr_eos();
|
|
3514 msg_didout = FALSE;
|
|
3515 msg_col = 0;
|
236
|
3516 out_flush();
|
|
3517 ui_breakcheck();
|
|
3518 if (got_int)
|
|
3519 break;
|
|
3520 }
|
|
3521
|
|
3522 /* The basic words are always stored with folded case. */
|
|
3523 dwlen = STRLEN(dw->dw_word);
|
|
3524 (void)str_foldcase(dw->dw_word, dwlen, foldword, MAXLINELEN);
|
|
3525 flags = captype(dw->dw_word, dw->dw_word + dwlen);
|
|
3526
|
|
3527 /* Check for non-word characters before the word. */
|
|
3528 clen = 0;
|
|
3529 leadlen = 0;
|
|
3530 if (!spell_iswordc(foldword))
|
|
3531 {
|
|
3532 p = foldword;
|
|
3533 for (;;)
|
|
3534 {
|
|
3535 mb_ptr_adv(p);
|
|
3536 ++clen;
|
|
3537 if (*p == NUL) /* Only non-word chars (bad word!) */
|
|
3538 {
|
|
3539 if (p_verbose > 0)
|
|
3540 smsg((char_u *)_("Warning: word without word characters: \"%s\""),
|
|
3541 foldword);
|
|
3542 break;
|
|
3543 }
|
|
3544 if (spell_iswordc(p))
|
|
3545 {
|
|
3546 /* Move the leader to "leadstring" and remove it from
|
|
3547 * "foldword". */
|
|
3548 leadlen = p - foldword;
|
|
3549 mch_memmove(leadstring, foldword, leadlen);
|
|
3550 leadstring[leadlen] = NUL;
|
|
3551 mch_memmove(foldword, p, STRLEN(p) + 1);
|
|
3552 break;
|
|
3553 }
|
|
3554 }
|
|
3555 }
|
|
3556
|
|
3557 /* Check for non-word characters after word characters. */
|
|
3558 addlen = 0;
|
|
3559 for (p = foldword; spell_iswordc(p); mb_ptr_adv(p))
|
|
3560 {
|
|
3561 if (*p == NUL)
|
|
3562 break;
|
|
3563 ++clen;
|
|
3564 }
|
|
3565 if (*p != NUL)
|
|
3566 {
|
|
3567 /* Move the addition to "addstring" and truncate "foldword". */
|
|
3568 if (flags & BWF_KEEPCAP)
|
|
3569 {
|
|
3570 /* Preserve caps, need to skip the right number of
|
|
3571 * characters in the original word (case folding may
|
|
3572 * change the byte count). */
|
|
3573 l = 0;
|
|
3574 for (cp = dw->dw_word; l < clen; mb_ptr_adv(cp))
|
|
3575 ++l;
|
|
3576 addlen = STRLEN(cp);
|
|
3577 mch_memmove(addstring, cp, addlen + 1);
|
|
3578 }
|
|
3579 else
|
|
3580 {
|
|
3581 addlen = STRLEN(p);
|
|
3582 mch_memmove(addstring, p, addlen + 1);
|
|
3583 }
|
|
3584 *p = NUL;
|
|
3585 }
|
|
3586
|
|
3587 bw = (basicword_T *)alloc_clear((unsigned)sizeof(basicword_T)
|
|
3588 + STRLEN(foldword));
|
|
3589 if (bw == NULL)
|
|
3590 break;
|
|
3591 STRCPY(bw->bw_word, foldword);
|
|
3592 bw->bw_region = regionmask;
|
|
3593
|
|
3594 if (leadlen > 0)
|
|
3595 bw->bw_leadstring = vim_strsave(leadstring);
|
|
3596 else
|
|
3597 bw->bw_leadstring = NULL;
|
|
3598 if (addlen > 0)
|
|
3599 bw->bw_addstring = vim_strsave(addstring);
|
|
3600 else
|
|
3601 bw->bw_addstring = NULL;
|
|
3602
|
|
3603 add_to_wordlist(newwords, bw);
|
|
3604
|
|
3605 if (flags & BWF_KEEPCAP)
|
|
3606 {
|
|
3607 if (addlen == 0)
|
|
3608 /* use the whole word */
|
|
3609 bw->bw_caseword = vim_strsave(dw->dw_word + leadlen);
|
|
3610 else
|
|
3611 /* use only up to the addition */
|
|
3612 bw->bw_caseword = vim_strnsave(dw->dw_word + leadlen,
|
|
3613 cp - dw->dw_word - leadlen);
|
|
3614 if (bw->bw_caseword == NULL) /* out of memory */
|
|
3615 flags &= ~BWF_KEEPCAP;
|
|
3616 }
|
|
3617 bw->bw_flags = flags;
|
|
3618
|
|
3619 /* Deal with any affix names on the old word, translate them
|
|
3620 * into affix numbers. */
|
|
3621 ga_init2(&bw->bw_prefix, sizeof(short_u), 10);
|
|
3622 ga_init2(&bw->bw_suffix, sizeof(short_u), 10);
|
|
3623 if (dw->dw_affnm != NULL)
|
|
3624 trans_affixes(dw, bw, oldaff, newwords);
|
|
3625 }
|
|
3626 }
|
|
3627 if (todo > 0)
|
|
3628 return FAIL;
|
|
3629 return OK;
|
|
3630 }
|
|
3631
|
|
3632 /*
|
|
3633 * Go through the list of words and combine the ones that are identical except
|
|
3634 * for the region.
|
|
3635 */
|
|
3636 static void
|
|
3637 combine_regions(newwords)
|
|
3638 hashtab_T *newwords;
|
|
3639 {
|
|
3640 int todo;
|
|
3641 hashitem_T *hi;
|
|
3642 basicword_T *bw, *nbw, *pbw;
|
|
3643
|
|
3644 /* Loop over all basic words in the words table. */
|
|
3645 todo = newwords->ht_used;
|
|
3646 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
3647 {
|
|
3648 if (!HASHITEM_EMPTY(hi))
|
|
3649 {
|
|
3650 --todo;
|
|
3651
|
|
3652 /* Loop over the list of words for this basic word. Compare with
|
|
3653 * each following word in the same list. */
|
|
3654 for (bw = HI2BW(hi); bw != NULL; bw = bw->bw_next)
|
|
3655 {
|
|
3656 pbw = bw;
|
|
3657 for (nbw = pbw->bw_next; nbw != NULL; nbw = pbw->bw_next)
|
|
3658 {
|
|
3659 if (bw->bw_flags == nbw->bw_flags
|
|
3660 && (bw->bw_leadstring == NULL)
|
|
3661 == (nbw->bw_leadstring == NULL)
|
|
3662 && (bw->bw_addstring == NULL)
|
|
3663 == (nbw->bw_addstring == NULL)
|
|
3664 && ((bw->bw_flags & BWF_KEEPCAP) == 0
|
|
3665 || (STRCMP(bw->bw_caseword,
|
|
3666 nbw->bw_caseword) == 0))
|
|
3667 && (bw->bw_leadstring == NULL
|
|
3668 || (STRCMP(bw->bw_leadstring,
|
|
3669 nbw->bw_leadstring) == 0))
|
|
3670 && (bw->bw_addstring == NULL
|
|
3671 || (STRCMP(bw->bw_addstring,
|
|
3672 nbw->bw_addstring) == 0))
|
|
3673 && same_affixes(bw, nbw)
|
|
3674 )
|
|
3675 {
|
|
3676 /* Match, combine regions and delete "nbw". */
|
|
3677 pbw->bw_next = nbw->bw_next;
|
|
3678 bw->bw_region |= nbw->bw_region;
|
|
3679 free_basicword(nbw);
|
|
3680 }
|
|
3681 else
|
|
3682 /* No match, continue with next one. */
|
|
3683 pbw = nbw;
|
|
3684 }
|
|
3685 }
|
|
3686 }
|
|
3687 }
|
|
3688 }
|
|
3689
|
|
3690 /*
|
|
3691 * Return TRUE when the prefixes and suffixes for "bw" and "nbw" are equal.
|
|
3692 */
|
|
3693 static int
|
|
3694 same_affixes(bw, nbw)
|
|
3695 basicword_T *bw;
|
|
3696 basicword_T *nbw;
|
|
3697 {
|
|
3698 return (bw->bw_prefix.ga_len == nbw->bw_prefix.ga_len
|
|
3699 && bw->bw_suffix.ga_len == nbw->bw_suffix.ga_len
|
|
3700 && (bw->bw_prefix.ga_len == 0
|
|
3701 || vim_memcmp(bw->bw_prefix.ga_data,
|
|
3702 nbw->bw_prefix.ga_data,
|
|
3703 bw->bw_prefix.ga_len * sizeof(short_u)) == 0)
|
|
3704 && (bw->bw_suffix.ga_len == 0
|
|
3705 || vim_memcmp(bw->bw_suffix.ga_data,
|
|
3706 nbw->bw_suffix.ga_data,
|
|
3707 bw->bw_suffix.ga_len * sizeof(short_u)) == 0));
|
|
3708 }
|
|
3709
|
|
3710 /*
|
243
|
3711 * For each basic word with additions turn the suffixes into other additions
|
|
3712 * and/or new basic words. For each basic word with a leadstring turn the
|
|
3713 * prefixes into other leadstrings and/or new basic words.
|
|
3714 * The result is that no affixes apply to the additions or leadstring of a
|
|
3715 * word.
|
|
3716 * This is also needed when a word with an addition has a prefix and the word
|
|
3717 * with prefix also exists. E.g., "blurp's/D" (D is prefix "de") and
|
|
3718 * "deblurp". "deblurp" would match and no prefix would be tried.
|
236
|
3719 */
|
|
3720 static void
|
|
3721 expand_affixes(newwords, prefgap, suffgap)
|
|
3722 hashtab_T *newwords;
|
|
3723 garray_T *prefgap;
|
|
3724 garray_T *suffgap;
|
|
3725 {
|
|
3726 int todo;
|
|
3727 hashitem_T *hi;
|
|
3728 basicword_T *bw;
|
|
3729 int pi, si;
|
|
3730 affentry_T *pae, *sae;
|
|
3731 garray_T add_words;
|
|
3732 int n;
|
240
|
3733 char_u message[MAXLINELEN + MAXWLEN];
|
236
|
3734
|
|
3735 ga_init2(&add_words, sizeof(basicword_T *), 10);
|
|
3736
|
|
3737 todo = newwords->ht_used;
|
|
3738 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
3739 {
|
|
3740 if (!HASHITEM_EMPTY(hi))
|
|
3741 {
|
|
3742 --todo;
|
240
|
3743
|
|
3744 /* This takes time, print a message now and then. */
|
|
3745 if ((todo & 0x3ff) == 0 || todo == (int)newwords->ht_used - 1)
|
|
3746 {
|
|
3747 sprintf((char *)message, _("%6d todo - %s"),
|
|
3748 todo, HI2BW(hi)->bw_word);
|
|
3749 msg_start();
|
|
3750 msg_outtrans_attr(message, 0);
|
|
3751 msg_clr_eos();
|
|
3752 msg_didout = FALSE;
|
|
3753 msg_col = 0;
|
|
3754 out_flush();
|
|
3755 ui_breakcheck();
|
|
3756 if (got_int)
|
|
3757 break;
|
|
3758 }
|
|
3759
|
236
|
3760 for (bw = HI2BW(hi); bw != NULL; bw = bw->bw_next)
|
|
3761 {
|
|
3762 /*
|
|
3763 * Need to fix affixes if there is a leader or addition and
|
|
3764 * there are prefixes or suffixes.
|
|
3765 */
|
|
3766 if ((bw->bw_leadstring != NULL || bw->bw_addstring != NULL)
|
|
3767 && (bw->bw_prefix.ga_len != 0
|
|
3768 || bw->bw_suffix.ga_len != 0))
|
|
3769 {
|
|
3770 /* Loop over all prefix numbers, but first without a
|
|
3771 * prefix. */
|
|
3772 for (pi = -1; pi < bw->bw_prefix.ga_len; ++pi)
|
|
3773 {
|
|
3774 pae = NULL;
|
|
3775 if (pi >= 0)
|
|
3776 {
|
|
3777 n = ((short_u *)bw->bw_prefix.ga_data)[pi];
|
|
3778 pae = ((affheader_T *)prefgap->ga_data + n)
|
|
3779 ->ah_first;
|
|
3780 }
|
|
3781
|
|
3782 /* Loop over all entries for prefix "pi". Do it once
|
|
3783 * when there is no prefix (pi == -1). */
|
|
3784 do
|
|
3785 {
|
243
|
3786 /* Skip prewords, they don't need to be expanded. */
|
|
3787 if (pae == NULL || !pae->ae_preword)
|
|
3788 {
|
236
|
3789 /* Loop over all suffix numbers. Do without a
|
|
3790 * suffix first when there is a prefix. */
|
|
3791 for (si = (pi == -1 ? 0 : -1);
|
|
3792 si < bw->bw_suffix.ga_len; ++si)
|
|
3793 {
|
|
3794 sae = NULL;
|
|
3795 if (si >= 0)
|
|
3796 {
|
|
3797 n = ((short_u *)bw->bw_suffix.ga_data)[si];
|
|
3798 sae = ((affheader_T *)suffgap->ga_data + n)
|
|
3799 ->ah_first;
|
|
3800 }
|
|
3801
|
|
3802 /* Loop over all entries for suffix "si". Do
|
|
3803 * it once when there is no suffix (si == -1).
|
|
3804 */
|
|
3805 do
|
|
3806 {
|
|
3807 /* Expand the word for this combination of
|
|
3808 * prefixes and affixes. */
|
|
3809 expand_one_aff(bw, &add_words, pae, sae);
|
|
3810
|
|
3811 /* Advance to next suffix entry, if there
|
|
3812 * is one. */
|
|
3813 if (sae != NULL)
|
|
3814 sae = sae->ae_next;
|
|
3815 } while (sae != NULL);
|
|
3816 }
|
243
|
3817 }
|
|
3818
|
|
3819 /* Advance to next prefix entry, if there is one. */
|
|
3820 if (pae != NULL)
|
|
3821 pae = pae->ae_next;
|
236
|
3822 } while (pae != NULL);
|
|
3823 }
|
|
3824 }
|
|
3825 }
|
|
3826 }
|
|
3827 }
|
|
3828
|
|
3829 /*
|
|
3830 * Add the new words afterwards, can't change "newwords" while going over
|
|
3831 * all its items.
|
|
3832 */
|
|
3833 for (pi = 0; pi < add_words.ga_len; ++pi)
|
|
3834 add_to_wordlist(newwords, ((basicword_T **)add_words.ga_data)[pi]);
|
|
3835
|
|
3836 ga_clear(&add_words);
|
|
3837 }
|
|
3838
|
|
3839 /*
|
|
3840 * Add one word to "add_words" for basic word "bw" with additions, adding
|
|
3841 * prefix "pae" and suffix "sae". Either "pae" or "sae" can be NULL.
|
243
|
3842 * Don't do this when not necessary:
|
|
3843 * - no leadstring and adding prefix doesn't result in existing word.
|
236
|
3844 */
|
|
3845 static void
|
|
3846 expand_one_aff(bw, add_words, pae, sae)
|
|
3847 basicword_T *bw;
|
|
3848 garray_T *add_words;
|
|
3849 affentry_T *pae;
|
|
3850 affentry_T *sae;
|
|
3851 {
|
|
3852 char_u word[MAXWLEN + 1];
|
|
3853 char_u caseword[MAXWLEN + 1];
|
|
3854 int l = 0;
|
|
3855 int choplen = 0;
|
|
3856 int ll;
|
|
3857 basicword_T *nbw;
|
|
3858
|
|
3859 /* Prepend prefix to the basic word if there is a prefix and there is no
|
|
3860 * leadstring. */
|
|
3861 if (pae != NULL && bw->bw_leadstring == NULL)
|
|
3862 {
|
|
3863 if (pae->ae_add != NULL)
|
|
3864 {
|
|
3865 l = STRLEN(pae->ae_add);
|
|
3866 mch_memmove(word, pae->ae_add, l);
|
|
3867 }
|
|
3868 if (pae->ae_chop != NULL)
|
|
3869 choplen = STRLEN(pae->ae_chop);
|
|
3870 }
|
|
3871
|
|
3872 /* Copy the body of the word. */
|
|
3873 STRCPY(word + l, bw->bw_word + choplen);
|
|
3874
|
|
3875 /* Do the same for bw_caseword, if it's there. */
|
|
3876 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3877 {
|
|
3878 if (l > 0)
|
|
3879 mch_memmove(caseword, pae->ae_add, l);
|
|
3880 STRCPY(caseword + l, bw->bw_caseword + choplen);
|
|
3881 }
|
|
3882
|
|
3883 /* Append suffix to the basic word if there is a suffix and there is no
|
|
3884 * addstring. */
|
|
3885 if (sae != 0 && bw->bw_addstring == NULL)
|
|
3886 {
|
|
3887 l = STRLEN(word);
|
|
3888 if (sae->ae_chop != NULL)
|
|
3889 l -= STRLEN(sae->ae_chop);
|
|
3890 if (sae->ae_add == NULL)
|
|
3891 word[l] = NUL;
|
|
3892 else
|
|
3893 STRCPY(word + l, sae->ae_add);
|
|
3894
|
|
3895 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3896 {
|
|
3897 /* Do the same for the caseword. */
|
|
3898 l = STRLEN(caseword);
|
|
3899 if (sae->ae_chop != NULL)
|
|
3900 l -= STRLEN(sae->ae_chop);
|
|
3901 if (sae->ae_add == NULL)
|
|
3902 caseword[l] = NUL;
|
|
3903 else
|
|
3904 STRCPY(caseword + l, sae->ae_add);
|
|
3905 }
|
|
3906 }
|
|
3907
|
|
3908 nbw = (basicword_T *)alloc_clear((unsigned)
|
|
3909 sizeof(basicword_T) + STRLEN(word));
|
|
3910 if (nbw != NULL)
|
|
3911 {
|
|
3912 /* Add the new word to the list of words to be added later. */
|
|
3913 if (ga_grow(add_words, 1) == FAIL)
|
|
3914 {
|
|
3915 vim_free(nbw);
|
|
3916 return;
|
|
3917 }
|
|
3918 ((basicword_T **)add_words->ga_data)[add_words->ga_len++] = nbw;
|
|
3919
|
|
3920 /* Copy the (modified) basic word, flags and region. */
|
|
3921 STRCPY(nbw->bw_word, word);
|
|
3922 nbw->bw_flags = bw->bw_flags;
|
|
3923 nbw->bw_region = bw->bw_region;
|
|
3924
|
|
3925 /* Set the (modified) caseword. */
|
|
3926 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3927 if ((nbw->bw_caseword = vim_strsave(caseword)) == NULL)
|
|
3928 nbw->bw_flags &= ~BWF_KEEPCAP;
|
|
3929
|
|
3930 if (bw->bw_leadstring != NULL)
|
|
3931 {
|
|
3932 if (pae != NULL)
|
|
3933 {
|
|
3934 /* Prepend prefix to leadstring. */
|
|
3935 ll = STRLEN(bw->bw_leadstring);
|
|
3936 l = choplen = 0;
|
|
3937 if (pae->ae_add != NULL)
|
|
3938 l = STRLEN(pae->ae_add);
|
|
3939 if (pae->ae_chop != NULL)
|
|
3940 {
|
|
3941 choplen = STRLEN(pae->ae_chop);
|
|
3942 if (choplen > ll) /* TODO: error? */
|
|
3943 choplen = ll;
|
|
3944 }
|
|
3945 nbw->bw_leadstring = alloc((unsigned)(ll + l - choplen + 1));
|
|
3946 if (nbw->bw_leadstring != NULL)
|
|
3947 {
|
|
3948 if (l > 0)
|
|
3949 mch_memmove(nbw->bw_leadstring, pae->ae_add, l);
|
|
3950 STRCPY(nbw->bw_leadstring + l, bw->bw_leadstring + choplen);
|
|
3951 }
|
|
3952 }
|
|
3953 else
|
|
3954 nbw->bw_leadstring = vim_strsave(bw->bw_leadstring);
|
|
3955 }
|
243
|
3956 else if (bw->bw_prefix.ga_len > 0)
|
|
3957 {
|
|
3958 /* There is no leadstring, copy the list of possible prefixes. */
|
|
3959 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3960 if (ga_grow(&nbw->bw_prefix, bw->bw_prefix.ga_len) == OK)
|
|
3961 {
|
|
3962 mch_memmove(nbw->bw_prefix.ga_data, bw->bw_prefix.ga_data,
|
|
3963 bw->bw_prefix.ga_len * sizeof(short_u));
|
|
3964 nbw->bw_prefix.ga_len = bw->bw_prefix.ga_len;
|
|
3965 }
|
|
3966 }
|
236
|
3967
|
|
3968 if (bw->bw_addstring != NULL)
|
|
3969 {
|
|
3970 if (sae != NULL)
|
|
3971 {
|
|
3972 /* Append suffix to addstring. */
|
|
3973 l = STRLEN(bw->bw_addstring);
|
|
3974 if (sae->ae_chop != NULL)
|
|
3975 {
|
|
3976 l -= STRLEN(sae->ae_chop);
|
|
3977 if (l < 0) /* TODO: error? */
|
|
3978 l = 0;
|
|
3979 }
|
|
3980 if (sae->ae_add == NULL)
|
|
3981 ll = 0;
|
|
3982 else
|
|
3983 ll = STRLEN(sae->ae_add);
|
|
3984 nbw->bw_addstring = alloc((unsigned)(ll + l - choplen + 1));
|
|
3985 if (nbw->bw_addstring != NULL)
|
|
3986 {
|
|
3987 STRCPY(nbw->bw_addstring, bw->bw_addstring);
|
|
3988 if (sae->ae_add == NULL)
|
|
3989 nbw->bw_addstring[l] = NUL;
|
|
3990 else
|
|
3991 STRCPY(nbw->bw_addstring + l, sae->ae_add);
|
|
3992 }
|
|
3993 }
|
|
3994 else
|
|
3995 nbw->bw_addstring = vim_strsave(bw->bw_addstring);
|
|
3996 }
|
|
3997 }
|
|
3998 }
|
|
3999
|
|
4000 /*
|
|
4001 * Add basicword_T "*bw" to wordlist "newwords".
|
|
4002 */
|
|
4003 static void
|
|
4004 add_to_wordlist(newwords, bw)
|
|
4005 hashtab_T *newwords;
|
|
4006 basicword_T *bw;
|
|
4007 {
|
|
4008 hashitem_T *hi;
|
|
4009 basicword_T *bw2;
|
|
4010
|
|
4011 hi = hash_find(newwords, bw->bw_word);
|
|
4012 if (HASHITEM_EMPTY(hi))
|
|
4013 {
|
|
4014 /* New entry, add to hashlist. */
|
|
4015 hash_add(newwords, bw->bw_word);
|
|
4016 bw->bw_next = NULL;
|
|
4017 }
|
|
4018 else
|
|
4019 {
|
|
4020 /* Existing entry, append to list of basic words. */
|
|
4021 bw2 = HI2BW(hi);
|
|
4022 bw->bw_next = bw2->bw_next;
|
|
4023 bw2->bw_next = bw;
|
|
4024 }
|
|
4025 }
|
|
4026
|
|
4027 /*
|
|
4028 * Write a number to file "fd", MSB first, in "len" bytes.
|
|
4029 */
|
|
4030 static void
|
|
4031 put_bytes(fd, nr, len)
|
|
4032 FILE *fd;
|
|
4033 long_u nr;
|
|
4034 int len;
|
|
4035 {
|
|
4036 int i;
|
|
4037
|
|
4038 for (i = len - 1; i >= 0; --i)
|
|
4039 putc((int)(nr >> (i * 8)), fd);
|
|
4040 }
|
|
4041
|
|
4042 /*
|
243
|
4043 * Write affix info. <affitemcnt> <affitem> ...
|
236
|
4044 */
|
|
4045 static void
|
|
4046 write_affix(fd, ah)
|
|
4047 FILE *fd;
|
|
4048 affheader_T *ah;
|
|
4049 {
|
|
4050 int i = 0;
|
|
4051 affentry_T *ae;
|
|
4052 char_u *p;
|
|
4053 int round;
|
243
|
4054 int flags;
|
236
|
4055
|
|
4056 /* Count the number of entries. */
|
|
4057 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
|
4058 ++i;
|
|
4059 put_bytes(fd, (long_u)i, 2); /* <affitemcnt> */
|
|
4060
|
243
|
4061 /* <affitem>: <affflags> <affchoplen> <affchop> <affaddlen> <affadd> */
|
236
|
4062 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
243
|
4063 {
|
|
4064 flags = ah->ah_combine ? AFF_COMBINE : 0;
|
|
4065 if (ae->ae_preword)
|
|
4066 flags |= AFF_PREWORD;
|
|
4067 fputc(flags, fd); /* <affflags> */
|
|
4068
|
236
|
4069 for (round = 1; round <= 2; ++round)
|
|
4070 {
|
|
4071 p = round == 1 ? ae->ae_chop : ae->ae_add;
|
|
4072 if (p == NULL)
|
|
4073 putc(0, fd); /* <affchoplen> / <affaddlen> */
|
|
4074 else
|
|
4075 {
|
|
4076 putc(STRLEN(p), fd); /* <affchoplen> / <affaddlen> */
|
|
4077 /* <affchop> / <affadd> */
|
|
4078 fwrite(p, STRLEN(p), (size_t)1, fd);
|
|
4079 }
|
|
4080 }
|
243
|
4081 }
|
236
|
4082 }
|
|
4083
|
|
4084 /*
|
|
4085 * Write list of affix NRs: <affixcnt> <affixNR> ...
|
|
4086 */
|
|
4087 static void
|
|
4088 write_affixlist(fd, aff, bytes)
|
|
4089 FILE *fd;
|
|
4090 garray_T *aff;
|
|
4091 int bytes;
|
|
4092 {
|
|
4093 int i;
|
|
4094
|
|
4095 if (aff->ga_len > 0)
|
|
4096 {
|
|
4097 putc(aff->ga_len, fd); /* <affixcnt> */
|
|
4098 for (i = 0; i < aff->ga_len; ++i)
|
|
4099 put_bytes(fd, (long_u )((short_u *)aff->ga_data)[i], bytes);
|
|
4100 }
|
|
4101 }
|
|
4102
|
|
4103 /*
|
|
4104 * Vim spell file format: <HEADER> <PREFIXLIST> <SUFFIXLIST>
|
|
4105 * <SUGGEST> <WORDLIST>
|
|
4106 *
|
|
4107 * <HEADER>: <fileID> <regioncnt> <regionname> ...
|
|
4108 *
|
249
|
4109 * <fileID> 10 bytes "VIMspell03"
|
236
|
4110 * <regioncnt> 1 byte number of regions following (8 supported)
|
|
4111 * <regionname> 2 bytes Region name: ca, au, etc.
|
|
4112 * First <regionname> is region 1.
|
|
4113 *
|
|
4114 *
|
|
4115 * <PREFIXLIST>: <affcount> <afftotcnt> <affix> ...
|
|
4116 * <SUFFIXLIST>: <affcount> <afftotcnt> <affix> ...
|
|
4117 * list of possible affixes: prefixes and suffixes.
|
|
4118 *
|
|
4119 * <affcount> 2 bytes Number of affixes (MSB comes first).
|
|
4120 * When more than 256 an affixNR is 2 bytes.
|
|
4121 * This is separate for prefixes and suffixes!
|
|
4122 * First affixNR is 0.
|
|
4123 * <afftotcnt> 2 bytes Total number of affix items (MSB comes first).
|
|
4124 *
|
243
|
4125 * <affix>: <affitemcnt> <affitem> ...
|
236
|
4126 *
|
|
4127 * <affitemcnt> 2 bytes Number of affixes with this affixNR (MSB first).
|
|
4128 *
|
243
|
4129 * <affitem>: <affflags> <affchoplen> <affchop> <affaddlen> <affadd>
|
236
|
4130 *
|
243
|
4131 * <affflags> 1 byte 0x01: prefix combines with suffix, AFF_COMBINE
|
|
4132 * 0x02: prefix includes word, AFF_PREWORD
|
|
4133 * 0x04-0x80: unset
|
236
|
4134 * <affchoplen> 1 byte Length of <affchop> in bytes.
|
|
4135 * <affchop> N bytes To be removed from basic word.
|
|
4136 * <affaddlen> 1 byte Length of <affadd> in bytes.
|
|
4137 * <affadd> N bytes To be added to basic word.
|
|
4138 *
|
|
4139 *
|
|
4140 * <SUGGEST> : <suggestlen> <more> ...
|
|
4141 *
|
|
4142 * <suggestlen> 4 bytes Length of <SUGGEST> in bytes, excluding
|
|
4143 * <suggestlen>. MSB first.
|
|
4144 * <more> To be defined.
|
|
4145 *
|
|
4146 *
|
|
4147 * <WORDLIST>: <wordcount> <worditem> ...
|
|
4148 *
|
|
4149 * <wordcount> 4 bytes Number of <worditem> following. MSB first.
|
|
4150 *
|
|
4151 * <worditem>: <nr> <string> <flags> [<flags2>]
|
|
4152 * [<caselen> <caseword>]
|
|
4153 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
4154 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
4155 * [<region>]
|
|
4156 * [<addcnt> <add> ...]
|
|
4157 *
|
|
4158 * <nr> i 1 byte Number of bytes copied from previous word.
|
|
4159 * <string> N bytes Additional bytes for word, up to byte smaller than
|
|
4160 * 0x20 (space).
|
|
4161 * Must only contain case-folded word characters.
|
243
|
4162 * <flags> 1 byte 0x01: word is valid without addition, BWF_VALID
|
|
4163 * 0x02: has region byte, BWF_REGION
|
|
4164 * 0x04: first letter must be upper-case, BWF_ONECAP
|
236
|
4165 * 0x08: has suffixes, <affixcnt> and <affixNR> follow
|
243
|
4166 * BWF_SUFFIX
|
|
4167 * 0x10: more flags, <flags2> follows next, BWF_SECOND
|
236
|
4168 * 0x20-0x80: can't be used, unset
|
243
|
4169 * <flags2> 1 byte 0x01: has additions, <addcnt> and <add> follow,
|
|
4170 * BWF_ADDS
|
236
|
4171 * 0x02: has prefixes, <affixcnt> and <affixNR> follow
|
243
|
4172 * BWF_PREFIX
|
|
4173 * 0x04: all letters must be upper-case, BWF_ALLCAP
|
|
4174 * 0x08: case must match, BWF_KEEPCAP
|
249
|
4175 * 0x10: has more than 255 additions, <addcnt> is two
|
|
4176 * bytes, BWF_ADDS_M
|
236
|
4177 * 0x10-0x80: unset
|
|
4178 * <caselen> 1 byte Length of <caseword>.
|
|
4179 * <caseword> N bytes Word with matching case.
|
|
4180 * <affixcnt> 1 byte Number of affix NRs following.
|
|
4181 * <affixNR> 1 or 2 byte Number of possible affix for this word.
|
|
4182 * When using 2 bytes MSB comes first.
|
|
4183 * <region> 1 byte Bitmask for regions in which word is valid. When
|
|
4184 * omitted it's valid in all regions.
|
|
4185 * Lowest bit is for region 1.
|
249
|
4186 * <addcnt> 1 or 2 byte Number of <add> items following.
|
236
|
4187 *
|
249
|
4188 * <add>: <addflags> <addlen> [<leadlen>] [<copylen>] [<addstring>] [<region>]
|
236
|
4189 *
|
243
|
4190 * <addflags> 1 byte 0x01: unset
|
|
4191 * 0x02: has region byte, ADD_REGION
|
|
4192 * 0x04: first letter must be upper-case, ADD_ONECAP
|
249
|
4193 * 0x08: unset
|
|
4194 * 0x10: has a <leadlen>, ADD_LEADLEN
|
|
4195 * 0x20: has a <copylen>, ADD_COPYLEN
|
243
|
4196 * 0x40: all letters must be upper-case, ADD_ALLCAP
|
|
4197 * 0x80: fixed case, <addstring> is the whole word
|
|
4198 * with matching case, ADD_KEEPCAP.
|
236
|
4199 * <addlen> 1 byte Length of <addstring> in bytes.
|
|
4200 * <leadlen> 1 byte Number of bytes at start of <addstring> that must
|
|
4201 * come before the start of the basic word.
|
249
|
4202 * <copylen> 1 byte Number of bytes copied from previous <addstring>.
|
236
|
4203 * <addstring> N bytes Word characters, before/in/after the word.
|
|
4204 *
|
|
4205 * All text characters are in 'encoding': <affchop>, <affadd>, <string>,
|
|
4206 * <caseword>> and <addstring>.
|
|
4207 * All other fields are ASCII: <regionname>
|
|
4208 * <string> is always case-folded.
|
|
4209 */
|
|
4210
|
|
4211 /*
|
|
4212 * Write the Vim spell file "fname".
|
|
4213 */
|
|
4214 static void
|
|
4215 write_vim_spell(fname, prefga, suffga, newwords, regcount, regchars)
|
|
4216 char_u *fname;
|
|
4217 garray_T *prefga; /* prefixes, affheader_T entries */
|
|
4218 garray_T *suffga; /* suffixes, affheader_T entries */
|
|
4219 hashtab_T *newwords; /* basic words, basicword_T entries */
|
|
4220 int regcount; /* number of regions */
|
|
4221 char_u *regchars; /* region names */
|
|
4222 {
|
249
|
4223 winfo_T wif;
|
236
|
4224 garray_T *gap;
|
|
4225 hashitem_T *hi;
|
|
4226 char_u **wtab;
|
|
4227 int todo;
|
|
4228 int flags, aflags;
|
249
|
4229 basicword_T *bw, *bwf, *bw2 = NULL;
|
236
|
4230 int i;
|
|
4231 int cnt;
|
|
4232 affentry_T *ae;
|
|
4233 int round;
|
|
4234 garray_T bwga;
|
|
4235
|
249
|
4236 vim_memset(&wif, 0, sizeof(winfo_T));
|
|
4237
|
|
4238 wif.wif_fd = fopen((char *)fname, "w");
|
|
4239 if (wif.wif_fd == NULL)
|
236
|
4240 {
|
|
4241 EMSG2(_(e_notopen), fname);
|
|
4242 return;
|
|
4243 }
|
|
4244
|
249
|
4245 fwrite(VIMSPELLMAGIC, VIMSPELLMAGICL, (size_t)1, wif.wif_fd);
|
236
|
4246
|
|
4247 /* write the region names if there is more than one */
|
|
4248 if (regcount > 1)
|
|
4249 {
|
249
|
4250 putc(regcount, wif.wif_fd);
|
|
4251 fwrite(regchars, (size_t)(regcount * 2), (size_t)1, wif.wif_fd);
|
|
4252 wif.wif_regionmask = (1 << regcount) - 1;
|
236
|
4253 }
|
|
4254 else
|
|
4255 {
|
249
|
4256 putc(0, wif.wif_fd);
|
|
4257 wif.wif_regionmask = 0;
|
236
|
4258 }
|
|
4259
|
|
4260 /* Write the prefix and suffix lists. */
|
|
4261 for (round = 1; round <= 2; ++round)
|
|
4262 {
|
|
4263 gap = round == 1 ? prefga : suffga;
|
249
|
4264 put_bytes(wif.wif_fd, (long_u)gap->ga_len, 2); /* <affcount> */
|
236
|
4265
|
|
4266 /* Count the total number of affix items. */
|
|
4267 cnt = 0;
|
|
4268 for (i = 0; i < gap->ga_len; ++i)
|
|
4269 for (ae = ((affheader_T *)gap->ga_data + i)->ah_first;
|
|
4270 ae != NULL; ae = ae->ae_next)
|
|
4271 ++cnt;
|
249
|
4272 put_bytes(wif.wif_fd, (long_u)cnt, 2); /* <afftotcnt> */
|
236
|
4273
|
|
4274 for (i = 0; i < gap->ga_len; ++i)
|
249
|
4275 write_affix(wif.wif_fd, (affheader_T *)gap->ga_data + i);
|
236
|
4276 }
|
|
4277
|
|
4278 /* Number of bytes used for affix NR depends on affix count. */
|
249
|
4279 wif.wif_prefm = (prefga->ga_len > 256) ? 2 : 1;
|
|
4280 wif.wif_suffm = (suffga->ga_len > 256) ? 2 : 1;
|
236
|
4281
|
|
4282 /* Write the suggest info. TODO */
|
249
|
4283 put_bytes(wif.wif_fd, 0L, 4);
|
236
|
4284
|
|
4285 /*
|
|
4286 * Write the word list. <wordcount> <worditem> ...
|
|
4287 */
|
|
4288 /* number of basic words in 4 bytes */
|
249
|
4289 put_bytes(wif.wif_fd, newwords->ht_used, 4); /* <wordcount> */
|
236
|
4290
|
|
4291 /*
|
249
|
4292 * Sort the word list, so that we can copy as many bytes as possible from
|
|
4293 * the previous word.
|
236
|
4294 */
|
|
4295 wtab = (char_u **)alloc((unsigned)(sizeof(char_u *) * newwords->ht_used));
|
|
4296 if (wtab != NULL)
|
|
4297 {
|
|
4298 /* Make a table with pointers to each word. */
|
|
4299 todo = newwords->ht_used;
|
|
4300 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
4301 if (!HASHITEM_EMPTY(hi))
|
|
4302 wtab[--todo] = hi->hi_key;
|
|
4303
|
|
4304 /* Sort. */
|
|
4305 sort_strings(wtab, (int)newwords->ht_used);
|
|
4306
|
|
4307 /* Now write each basic word to the spell file. */
|
|
4308 ga_init2(&bwga, sizeof(basicword_T *), 10);
|
240
|
4309 for (todo = 0; (long_u)todo < newwords->ht_used; ++todo)
|
236
|
4310 {
|
|
4311 bwf = KEY2BW(wtab[todo]);
|
|
4312
|
|
4313 /*
|
|
4314 * Reorder the list of basicword_T words: make a list for words
|
|
4315 * with the same case-folded word. Put them together for same
|
|
4316 * caps (ONECAP, ALLCAP and various KEEPCAP words) and same
|
|
4317 * affixes. Each list will then be put as a basic word with
|
|
4318 * additions.
|
|
4319 * This won't take much space, since the basic word is the same
|
|
4320 * every time, only its length is written.
|
|
4321 */
|
|
4322 bwga.ga_len = 0;
|
|
4323 for (bw = bwf; bw != NULL; bw = bw->bw_next)
|
|
4324 {
|
|
4325 flags = bw->bw_flags & (BWF_ONECAP | BWF_KEEPCAP | BWF_ALLCAP);
|
|
4326
|
|
4327 /* Go through the lists we found so far. Break when the case
|
|
4328 * matches. */
|
|
4329 for (i = 0; i < bwga.ga_len; ++i)
|
|
4330 {
|
|
4331 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4332 aflags = bw2->bw_flags & (BWF_ONECAP | BWF_KEEPCAP
|
|
4333 | BWF_ALLCAP);
|
|
4334 if (flags == aflags
|
|
4335 && ((flags & BWF_KEEPCAP) == 0
|
|
4336 || (STRCMP(bw->bw_caseword,
|
|
4337 bw2->bw_caseword) == 0))
|
|
4338 && same_affixes(bw, bw2))
|
|
4339 break;
|
|
4340 }
|
|
4341 if (i == bwga.ga_len)
|
|
4342 {
|
|
4343 /* No word with similar caps, make a new list. */
|
|
4344 if (ga_grow(&bwga, 1) == FAIL)
|
|
4345 break;
|
|
4346 ((basicword_T **)bwga.ga_data)[i] = bw;
|
|
4347 bw->bw_cnext = NULL;
|
|
4348 ++bwga.ga_len;
|
|
4349 }
|
|
4350 else
|
|
4351 {
|
|
4352 /* Add to list of words with similar caps. */
|
|
4353 bw->bw_cnext = bw2->bw_cnext;
|
|
4354 bw2->bw_cnext = bw;
|
|
4355 }
|
|
4356 }
|
|
4357
|
|
4358 /* Prefer the word with no caps to use as the first basic word.
|
|
4359 * At least one without KEEPCAP. */
|
|
4360 bw = NULL;
|
|
4361 for (i = 0; i < bwga.ga_len; ++i)
|
|
4362 {
|
|
4363 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4364 if (bw == NULL
|
|
4365 || (bw2->bw_flags & (BWF_ONECAP | BWF_KEEPCAP
|
|
4366 | BWF_ALLCAP)) == 0
|
|
4367 || (bw->bw_flags & BWF_KEEPCAP))
|
|
4368 bw = bw2;
|
|
4369 }
|
|
4370
|
|
4371 /* Write first basic word. If it's KEEPCAP then we need a word
|
|
4372 * without VALID flag first (makes it easier to read the list back
|
|
4373 * in). */
|
|
4374 if (bw->bw_flags & BWF_KEEPCAP)
|
249
|
4375 write_bword(&wif, bw, TRUE);
|
|
4376 write_bword(&wif, bw, FALSE);
|
236
|
4377
|
|
4378 /* Write other basic words, with different caps. */
|
|
4379 for (i = 0; i < bwga.ga_len; ++i)
|
|
4380 {
|
|
4381 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4382 if (bw2 != bw)
|
249
|
4383 write_bword(&wif, bw2, FALSE);
|
236
|
4384 }
|
|
4385 }
|
|
4386
|
|
4387 ga_clear(&bwga);
|
|
4388 }
|
|
4389
|
249
|
4390 fclose(wif.wif_fd);
|
|
4391
|
|
4392 /* Print a few statistics. */
|
|
4393 if (wif.wif_addmaxw == NULL)
|
|
4394 wif.wif_addmaxw = (char_u *)"";
|
|
4395 smsg((char_u *)_("Maximum number of adds on a word: %ld (%s)"),
|
|
4396 wif.wif_addmax, wif.wif_addmaxw);
|
|
4397 smsg((char_u *)_("Average number of adds on a word: %f"),
|
|
4398 (float)wif.wif_acount / (float)wif.wif_wcount);
|
|
4399 }
|
|
4400
|
|
4401 /*
|
|
4402 * Compare two basic words for their <addstring>.
|
|
4403 */
|
|
4404 static int
|
|
4405 #ifdef __BORLANDC__
|
|
4406 _RTLENTRYF
|
|
4407 #endif
|
|
4408 bw_compare __ARGS((const void *s1, const void *s2));
|
|
4409
|
|
4410 static int
|
|
4411 #ifdef __BORLANDC__
|
|
4412 _RTLENTRYF
|
|
4413 #endif
|
|
4414 bw_compare(s1, s2)
|
|
4415 const void *s1;
|
|
4416 const void *s2;
|
|
4417 {
|
|
4418 basicword_T *bw1 = *(basicword_T **)s1;
|
|
4419 basicword_T *bw2 = *(basicword_T **)s2;
|
|
4420 int i = 0;
|
|
4421
|
|
4422 /* compare the leadstrings */
|
|
4423 if (bw1->bw_leadstring == NULL)
|
|
4424 {
|
|
4425 if (bw2->bw_leadstring != NULL)
|
|
4426 return 1;
|
|
4427 }
|
|
4428 else if (bw2->bw_leadstring == NULL)
|
|
4429 return -1;
|
|
4430 else
|
|
4431 i = STRCMP(bw1->bw_leadstring, bw2->bw_leadstring);
|
|
4432
|
|
4433 if (i == 0)
|
|
4434 {
|
|
4435 /* leadstrings are identical, compare the addstrings */
|
|
4436 if (bw1->bw_addstring == NULL)
|
|
4437 {
|
|
4438 if (bw2->bw_addstring != NULL)
|
|
4439 return 1;
|
|
4440 }
|
|
4441 else if (bw2->bw_addstring == NULL)
|
|
4442 return -1;
|
|
4443 else
|
|
4444 i = STRCMP(bw1->bw_addstring, bw2->bw_addstring);
|
|
4445 }
|
|
4446 return i;
|
236
|
4447 }
|
|
4448
|
|
4449 /*
|
|
4450 * Write basic word, followed by any additions.
|
|
4451 *
|
|
4452 * <worditem>: <nr> <string> <flags> [<flags2>]
|
|
4453 * [<caselen> <caseword>]
|
|
4454 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
4455 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
4456 * [<region>]
|
|
4457 * [<addcnt> <add> ...]
|
|
4458 */
|
|
4459 static void
|
249
|
4460 write_bword(wif, bwf, lowcap)
|
|
4461 winfo_T *wif; /* info for writing */
|
236
|
4462 basicword_T *bwf;
|
|
4463 int lowcap; /* write KEEPKAP word as not-valid */
|
|
4464 {
|
249
|
4465 FILE *fd = wif->wif_fd;
|
236
|
4466 int flags;
|
|
4467 int aflags;
|
|
4468 int len;
|
|
4469 int leadlen, addlen;
|
249
|
4470 int copylen;
|
236
|
4471 int clen;
|
|
4472 int adds = 0;
|
|
4473 int i;
|
249
|
4474 int idx;
|
236
|
4475 basicword_T *bw, *bw2;
|
249
|
4476 basicword_T **wtab;
|
|
4477 int count;
|
|
4478 int l;
|
236
|
4479
|
|
4480 /* Check how many bytes can be copied from the previous word. */
|
|
4481 len = STRLEN(bwf->bw_word);
|
249
|
4482 if (wif->wif_prevbw == NULL)
|
236
|
4483 clen = 0;
|
|
4484 else
|
|
4485 for (clen = 0; clen < len
|
249
|
4486 && wif->wif_prevbw->bw_word[clen] == bwf->bw_word[clen]; ++clen)
|
236
|
4487 ;
|
|
4488 putc(clen, fd); /* <nr> */
|
249
|
4489 wif->wif_prevbw = bwf;
|
236
|
4490 /* <string> */
|
|
4491 if (len > clen)
|
|
4492 fwrite(bwf->bw_word + clen, (size_t)(len - clen), (size_t)1, fd);
|
|
4493
|
|
4494 /* Try to find a word without additions to use first. */
|
|
4495 bw = bwf;
|
|
4496 for (bw2 = bwf; bw2 != NULL; bw2 = bw2->bw_cnext)
|
|
4497 {
|
|
4498 if (bw2->bw_addstring != NULL || bw2->bw_leadstring != NULL)
|
|
4499 ++adds;
|
|
4500 else
|
|
4501 bw = bw2;
|
|
4502 }
|
|
4503
|
|
4504 /* Flags: If there is no leadstring and no addstring the basic word is
|
|
4505 * valid, may have prefixes, suffixes and region. */
|
|
4506 flags = bw->bw_flags;
|
|
4507 if (bw->bw_addstring == NULL && bw->bw_leadstring == NULL)
|
|
4508 {
|
|
4509 flags |= BWF_VALID;
|
|
4510
|
|
4511 /* Flags: add the region byte if the word isn't valid in all
|
|
4512 * regions. */
|
249
|
4513 if (wif->wif_regionmask != 0 && (bw->bw_region & wif->wif_regionmask)
|
|
4514 != wif->wif_regionmask)
|
236
|
4515 flags |= BWF_REGION;
|
|
4516 }
|
243
|
4517 /* Add the prefix/suffix list if there are prefixes/suffixes. */
|
|
4518 if (bw->bw_leadstring == NULL && bw->bw_prefix.ga_len > 0)
|
|
4519 flags |= BWF_PREFIX;
|
|
4520 if (bw->bw_addstring == NULL && bw->bw_suffix.ga_len > 0)
|
|
4521 flags |= BWF_SUFFIX;
|
236
|
4522
|
|
4523 /* Flags: may have additions. */
|
|
4524 if (adds > 0)
|
249
|
4525 {
|
236
|
4526 flags |= BWF_ADDS;
|
249
|
4527 if (adds >= 256)
|
|
4528 flags |= BWF_ADDS_M;
|
|
4529 }
|
236
|
4530
|
|
4531 /* The dummy word before a KEEPCAP word doesn't have any flags, they are
|
|
4532 * in the actual word that follows. */
|
|
4533 if (lowcap)
|
|
4534 flags = 0;
|
|
4535
|
|
4536 /* Flags: when the upper byte is not used we only write one flags
|
|
4537 * byte, if it's used then set an extra flag in the first byte and
|
|
4538 * also write the second byte. */
|
|
4539 if ((flags & 0xff00) == 0)
|
|
4540 putc(flags, fd); /* <flags> */
|
|
4541 else
|
|
4542 {
|
|
4543 putc(flags | BWF_SECOND, fd); /* <flags> */
|
|
4544 putc((int)((unsigned)flags >> 8), fd); /* <flags2> */
|
|
4545 }
|
|
4546
|
|
4547 /* First dummy word doesn't need anything but flags. */
|
|
4548 if (lowcap)
|
|
4549 return;
|
|
4550
|
|
4551 if (flags & BWF_KEEPCAP)
|
|
4552 {
|
|
4553 len = STRLEN(bw->bw_caseword);
|
|
4554 putc(len, fd); /* <caselen> */
|
|
4555 for (i = 0; i < len; ++i)
|
|
4556 putc(bw->bw_caseword[i], fd); /* <caseword> */
|
|
4557 }
|
|
4558
|
|
4559 /* write prefix and suffix lists: <affixcnt> <affixNR> ... */
|
|
4560 if (flags & BWF_PREFIX)
|
249
|
4561 write_affixlist(fd, &bw->bw_prefix, wif->wif_prefm);
|
236
|
4562 if (flags & BWF_SUFFIX)
|
249
|
4563 write_affixlist(fd, &bw->bw_suffix, wif->wif_suffm);
|
236
|
4564
|
|
4565 if (flags & BWF_REGION)
|
|
4566 putc(bw->bw_region, fd); /* <region> */
|
|
4567
|
249
|
4568 ++wif->wif_wcount;
|
|
4569
|
236
|
4570 /*
|
|
4571 * Additions.
|
|
4572 */
|
|
4573 if (adds > 0)
|
|
4574 {
|
249
|
4575 if (adds >= 256)
|
|
4576 put_bytes(fd, (long_u)adds, 2); /* 2 byte <addcnt> */
|
|
4577 else
|
|
4578 putc(adds, fd); /* 1 byte <addcnt> */
|
|
4579
|
|
4580 /* statistics */
|
|
4581 wif->wif_acount += adds;
|
|
4582 if (wif->wif_addmax < adds)
|
|
4583 {
|
|
4584 wif->wif_addmax = adds;
|
|
4585 wif->wif_addmaxw = bw->bw_word;
|
|
4586 }
|
|
4587
|
|
4588 /*
|
|
4589 * Sort the list of additions, so that we can copy as many bytes as
|
|
4590 * possible from the previous addstring.
|
|
4591 */
|
|
4592
|
|
4593 /* Make a table with pointers to each basic word that has additions. */
|
|
4594 wtab = (basicword_T **)alloc((unsigned)(sizeof(basicword_T *) * adds));
|
|
4595 if (wtab == NULL)
|
|
4596 return;
|
|
4597 count = 0;
|
236
|
4598 for (bw = bwf; bw != NULL; bw = bw->bw_cnext)
|
|
4599 if (bw->bw_leadstring != NULL || bw->bw_addstring != NULL)
|
249
|
4600 wtab[count++] = bw;
|
|
4601
|
|
4602 /* Sort. */
|
|
4603 qsort((void *)wtab, (size_t)count, sizeof(basicword_T *), bw_compare);
|
|
4604
|
|
4605 /* Now write each basic word to the spell file. Copy bytes from the
|
|
4606 * previous leadstring/addstring if possible. */
|
|
4607 bw2 = NULL;
|
|
4608 for (idx = 0; idx < count; ++idx)
|
|
4609 {
|
|
4610 bw = wtab[idx];
|
|
4611
|
|
4612 /* <add>: <addflags> <addlen> [<leadlen>] [<copylen>]
|
|
4613 * [<addstring>] [<region>] */
|
|
4614 copylen = 0;
|
|
4615 if (bw->bw_leadstring == NULL)
|
|
4616 leadlen = 0;
|
|
4617 else
|
236
|
4618 {
|
249
|
4619 leadlen = STRLEN(bw->bw_leadstring);
|
|
4620 if (bw2 != NULL && bw2->bw_leadstring != NULL)
|
|
4621 for ( ; copylen < leadlen; ++copylen)
|
|
4622 if (bw->bw_leadstring[copylen]
|
|
4623 != bw2->bw_leadstring[copylen])
|
|
4624 break;
|
|
4625 }
|
|
4626 if (bw->bw_addstring == NULL)
|
|
4627 addlen = 0;
|
|
4628 else
|
|
4629 {
|
|
4630 addlen = STRLEN(bw->bw_addstring);
|
|
4631 if (bw2 != NULL && copylen == leadlen
|
|
4632 && bw2->bw_addstring != NULL)
|
|
4633 {
|
|
4634 for (i = 0; i < addlen; ++i)
|
|
4635 if (bw->bw_addstring[i] != bw2->bw_addstring[i])
|
|
4636 break;
|
|
4637 copylen += i;
|
|
4638 }
|
|
4639 }
|
|
4640
|
|
4641 aflags = 0;
|
|
4642 /* Only copy bytes when it's more than one, the length itself
|
|
4643 * takes an extra byte. */
|
|
4644 if (copylen > 1)
|
|
4645 aflags |= ADD_COPYLEN;
|
|
4646 else
|
|
4647 copylen = 0;
|
|
4648
|
|
4649 if (bw->bw_flags & BWF_ONECAP)
|
|
4650 aflags |= ADD_ONECAP;
|
|
4651 if (bw->bw_flags & BWF_ALLCAP)
|
|
4652 aflags |= ADD_ALLCAP;
|
|
4653 if (bw->bw_flags & BWF_KEEPCAP)
|
|
4654 aflags |= ADD_KEEPCAP;
|
|
4655 if (wif->wif_regionmask != 0 && (bw->bw_region
|
|
4656 & wif->wif_regionmask) != wif->wif_regionmask)
|
|
4657 aflags |= ADD_REGION;
|
|
4658 if (leadlen > 0)
|
|
4659 aflags |= ADD_LEADLEN;
|
|
4660 putc(aflags, fd); /* <addflags> */
|
|
4661
|
|
4662 putc(leadlen + addlen, fd); /* <addlen> */
|
|
4663 if (aflags & ADD_LEADLEN)
|
|
4664 putc(leadlen, fd); /* <leadlen> */
|
|
4665 if (aflags & ADD_COPYLEN)
|
|
4666 putc(copylen, fd); /* <copylen> */
|
|
4667
|
|
4668 /* <addstring> */
|
|
4669 if (leadlen > copylen && bw->bw_leadstring != NULL)
|
|
4670 fwrite(bw->bw_leadstring + copylen,
|
|
4671 (size_t)(leadlen - copylen), (size_t)1, fd);
|
|
4672 if (leadlen + addlen > copylen && bw->bw_addstring != NULL)
|
|
4673 {
|
|
4674 if (copylen >= leadlen)
|
|
4675 l = copylen - leadlen;
|
236
|
4676 else
|
249
|
4677 l = 0;
|
|
4678 fwrite(bw->bw_addstring + l,
|
|
4679 (size_t)(addlen - l), (size_t)1, fd);
|
236
|
4680 }
|
249
|
4681
|
|
4682 if (aflags & ADD_REGION)
|
|
4683 putc(bw->bw_region, fd); /* <region> */
|
|
4684
|
|
4685 bw2 = bw;
|
|
4686 }
|
|
4687 vim_free(wtab);
|
236
|
4688 }
|
|
4689 }
|
|
4690
|
|
4691
|
|
4692 /*
|
|
4693 * ":mkspell outfile infile ..."
|
|
4694 */
|
|
4695 void
|
|
4696 ex_mkspell(eap)
|
|
4697 exarg_T *eap;
|
|
4698 {
|
|
4699 int fcount;
|
|
4700 char_u **fnames;
|
|
4701 char_u fname[MAXPATHL];
|
|
4702 char_u wfname[MAXPATHL];
|
|
4703 afffile_T *(afile[8]);
|
|
4704 hashtab_T dfile[8];
|
|
4705 int i;
|
|
4706 int len;
|
|
4707 char_u region_name[16];
|
|
4708 struct stat st;
|
|
4709 int round;
|
|
4710 vimconv_T conv;
|
240
|
4711 int ascii = FALSE;
|
|
4712 char_u *arg = eap->arg;
|
|
4713
|
|
4714 if (STRNCMP(arg, "-ascii", 6) == 0)
|
|
4715 {
|
|
4716 ascii = TRUE;
|
|
4717 arg = skipwhite(arg + 6);
|
|
4718 }
|
|
4719
|
|
4720 /* Expand all the remaining arguments (e.g., $VIMRUNTIME). */
|
|
4721 if (get_arglist_exp(arg, &fcount, &fnames) == FAIL)
|
236
|
4722 return;
|
|
4723 if (fcount < 2)
|
|
4724 EMSG(_(e_invarg)); /* need at least output and input names */
|
|
4725 else if (fcount > 9)
|
|
4726 EMSG(_("E754: Only up to 8 regions supported"));
|
|
4727 else
|
|
4728 {
|
|
4729 /* Check for overwriting before doing things that may take a lot of
|
|
4730 * time. */
|
240
|
4731 sprintf((char *)wfname, "%s.%s.spl", fnames[0],
|
|
4732 ascii ? (char_u *)"ascii" : p_enc);
|
236
|
4733 if (!eap->forceit && mch_stat((char *)wfname, &st) >= 0)
|
|
4734 {
|
|
4735 EMSG(_(e_exists));
|
|
4736 goto theend;
|
|
4737 }
|
|
4738 if (mch_isdir(fnames[0]))
|
|
4739 {
|
|
4740 EMSG2(_(e_isadir2), fnames[0]);
|
|
4741 goto theend;
|
|
4742 }
|
|
4743
|
|
4744 /*
|
|
4745 * Init the aff and dic pointers.
|
|
4746 * Get the region names if there are more than 2 arguments.
|
|
4747 */
|
|
4748 for (i = 1; i < fcount; ++i)
|
|
4749 {
|
|
4750 afile[i - 1] = NULL;
|
|
4751 hash_init(&dfile[i - 1]);
|
|
4752 if (fcount > 2)
|
|
4753 {
|
|
4754 len = STRLEN(fnames[i]);
|
|
4755 if (STRLEN(gettail(fnames[i])) < 5 || fnames[i][len - 3] != '_')
|
|
4756 {
|
|
4757 EMSG2(_("E755: Invalid region in %s"), fnames[i]);
|
|
4758 goto theend;
|
|
4759 }
|
|
4760 else
|
|
4761 {
|
|
4762 region_name[(i - 1) * 2] = TOLOWER_ASC(fnames[i][len - 2]);
|
|
4763 region_name[(i - 1) * 2 + 1] =
|
|
4764 TOLOWER_ASC(fnames[i][len - 1]);
|
|
4765 }
|
|
4766 }
|
|
4767 }
|
|
4768
|
|
4769 /*
|
|
4770 * Read all the .aff and .dic files.
|
|
4771 * Text is converted to 'encoding'.
|
|
4772 */
|
|
4773 for (i = 1; i < fcount; ++i)
|
|
4774 {
|
|
4775 /* Read the .aff file. Will init "conv" based on the "SET" line. */
|
|
4776 conv.vc_type = CONV_NONE;
|
|
4777 sprintf((char *)fname, "%s.aff", fnames[i]);
|
240
|
4778 if ((afile[i - 1] = spell_read_aff(fname, &conv, ascii)) == NULL)
|
236
|
4779 break;
|
|
4780
|
|
4781 /* Read the .dic file. */
|
|
4782 sprintf((char *)fname, "%s.dic", fnames[i]);
|
240
|
4783 if (spell_read_dic(&dfile[i - 1], fname, &conv, ascii) == FAIL)
|
236
|
4784 break;
|
|
4785
|
|
4786 /* Free any conversion stuff. */
|
|
4787 convert_setup(&conv, NULL, NULL);
|
|
4788 }
|
|
4789
|
|
4790 /* Process the data when all the files could be read. */
|
|
4791 if (i == fcount)
|
|
4792 {
|
|
4793 garray_T prefga;
|
|
4794 garray_T suffga;
|
|
4795 garray_T *gap;
|
|
4796 hashtab_T newwords;
|
|
4797
|
|
4798 /*
|
|
4799 * Combine all the affixes into one new affix list. This is done
|
|
4800 * for prefixes and suffixes separately.
|
|
4801 * We need to do this for each region, try to re-use the same
|
|
4802 * affixes.
|
|
4803 * Since we number the new affix entries, a growarray will do. In
|
|
4804 * the affheader_T the ah_key is unused.
|
|
4805 */
|
|
4806 MSG(_("Combining affixes..."));
|
|
4807 out_flush();
|
|
4808 for (round = 1; round <= 2; ++round)
|
|
4809 {
|
|
4810 gap = round == 1 ? &prefga : &suffga;
|
|
4811 ga_init2(gap, sizeof(affheader_T), 50);
|
|
4812 for (i = 1; i < fcount; ++i)
|
|
4813 get_new_aff(round == 1 ? &afile[i - 1]->af_pref
|
240
|
4814 : &afile[i - 1]->af_suff,
|
|
4815 gap, round == 1);
|
236
|
4816 }
|
|
4817
|
|
4818 /*
|
|
4819 * Go over all words and:
|
|
4820 * - change the old affix names to the new affix numbers
|
|
4821 * - check the conditions
|
|
4822 * - fold case
|
|
4823 * - extract the basic word and additions.
|
|
4824 * Do this for each region.
|
|
4825 */
|
|
4826 MSG(_("Building word list..."));
|
|
4827 out_flush();
|
|
4828 hash_init(&newwords);
|
|
4829
|
|
4830 for (i = 1; i < fcount; ++i)
|
|
4831 build_wordlist(&newwords, &dfile[i - 1], afile[i - 1],
|
|
4832 1 << (i - 1));
|
|
4833
|
|
4834 if (fcount > 2)
|
|
4835 {
|
|
4836 /* Combine words for the different regions into one. */
|
|
4837 MSG(_("Combining regions..."));
|
|
4838 out_flush();
|
|
4839 combine_regions(&newwords);
|
|
4840 }
|
|
4841
|
|
4842 /*
|
|
4843 * Affixes on a word with additions are clumsy, would require
|
|
4844 * inefficient searching. Turn the affixes into additions and/or
|
|
4845 * the expanded word.
|
|
4846 */
|
|
4847 MSG(_("Processing words..."));
|
|
4848 out_flush();
|
|
4849 expand_affixes(&newwords, &prefga, &suffga);
|
|
4850
|
|
4851 /* Write the info in the spell file. */
|
|
4852 smsg((char_u *)_("Writing spell file %s..."), wfname);
|
|
4853 out_flush();
|
|
4854 write_vim_spell(wfname, &prefga, &suffga, &newwords,
|
|
4855 fcount - 1, region_name);
|
|
4856 MSG(_("Done!"));
|
|
4857 out_flush();
|
|
4858
|
|
4859 /* Free the allocated stuff. */
|
|
4860 free_wordtable(&newwords);
|
|
4861 for (round = 1; round <= 2; ++round)
|
|
4862 {
|
|
4863 gap = round == 1 ? &prefga: &suffga;
|
|
4864 for (i = 0; i < gap->ga_len; ++i)
|
|
4865 free_affixentries(((affheader_T *)gap->ga_data + i)
|
|
4866 ->ah_first);
|
|
4867 ga_clear(gap);
|
|
4868 }
|
|
4869 }
|
|
4870
|
|
4871 /* Free the .aff and .dic file structures. */
|
|
4872 for (i = 1; i < fcount; ++i)
|
|
4873 {
|
|
4874 if (afile[i - 1] != NULL)
|
|
4875 spell_free_aff(afile[i - 1]);
|
|
4876 spell_free_dic(&dfile[i - 1]);
|
|
4877 }
|
|
4878 }
|
|
4879
|
|
4880 theend:
|
|
4881 FreeWild(fcount, fnames);
|
|
4882 }
|
|
4883
|
|
4884 static void
|
|
4885 free_wordtable(ht)
|
|
4886 hashtab_T *ht;
|
|
4887 {
|
|
4888 int todo;
|
|
4889 basicword_T *bw, *nbw;
|
|
4890 hashitem_T *hi;
|
|
4891
|
|
4892 todo = ht->ht_used;
|
|
4893 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
4894 {
|
|
4895 if (!HASHITEM_EMPTY(hi))
|
|
4896 {
|
|
4897 --todo;
|
|
4898 for (bw = HI2BW(hi); bw != NULL; bw = nbw)
|
|
4899 {
|
|
4900 nbw = bw->bw_next;
|
|
4901 free_basicword(bw);
|
|
4902 }
|
|
4903 }
|
|
4904 }
|
|
4905 }
|
|
4906
|
|
4907 /*
|
|
4908 * Free a basicword_T and what it contains.
|
|
4909 */
|
|
4910 static void
|
|
4911 free_basicword(bw)
|
|
4912 basicword_T *bw;
|
|
4913 {
|
|
4914 ga_clear(&bw->bw_prefix);
|
|
4915 ga_clear(&bw->bw_suffix);
|
|
4916 vim_free(bw->bw_caseword);
|
|
4917 vim_free(bw->bw_leadstring);
|
|
4918 vim_free(bw->bw_addstring);
|
|
4919 vim_free(bw);
|
|
4920 }
|
|
4921
|
|
4922 /*
|
240
|
4923 * Free a list of affentry_T and what they contain.
|
236
|
4924 */
|
|
4925 static void
|
|
4926 free_affixentries(first)
|
|
4927 affentry_T *first;
|
|
4928 {
|
|
4929 affentry_T *ap, *an;
|
|
4930
|
|
4931 for (ap = first; ap != NULL; ap = an)
|
|
4932 {
|
|
4933 an = ap->ae_next;
|
240
|
4934 free_affix_entry(ap);
|
236
|
4935 }
|
|
4936 }
|
|
4937
|
240
|
4938 /*
|
|
4939 * Free one affentry_T and what it contains.
|
|
4940 */
|
|
4941 static void
|
|
4942 free_affix_entry(ap)
|
|
4943 affentry_T *ap;
|
|
4944 {
|
|
4945 vim_free(ap->ae_chop);
|
|
4946 vim_free(ap->ae_add);
|
|
4947 vim_free(ap->ae_cond);
|
|
4948 vim_free(ap->ae_prog);
|
|
4949 vim_free(ap);
|
|
4950 }
|
|
4951
|
236
|
4952 #endif /* FEAT_MBYTE */
|
|
4953
|
|
4954 #endif /* FEAT_SYN_HL */
|