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.
|
255
|
100 * The main access is through hashtable "sl_word", using the case-folded
|
|
101 * word as the key. This finds a linked list of fword_T.
|
236
|
102 */
|
|
103 typedef struct slang_S slang_T;
|
|
104 struct slang_S
|
|
105 {
|
|
106 slang_T *sl_next; /* next language */
|
|
107 char_u *sl_name; /* language name "en", "en.rare", "nl", etc. */
|
|
108 hashtab_T sl_words; /* main word table, fword_T */
|
|
109 int sl_prefcnt; /* number of prefix NRs */
|
|
110 garray_T sl_preftab; /* list of hashtables to lookup prefixes */
|
|
111 affitem_T *sl_prefzero; /* list of prefixes with zero add length */
|
243
|
112 hashtab_T sl_prewords; /* prefixes that include a word */
|
236
|
113 int sl_suffcnt; /* number of suffix NRs */
|
|
114 garray_T sl_sufftab; /* list of hashtables to lookup suffixes */
|
|
115 affitem_T *sl_suffzero; /* list of suffixes with zero add length */
|
243
|
116 char_u *sl_try; /* "TRY" from .aff file TODO: not used */
|
|
117 garray_T sl_rep; /* list of repentry_T entries from REP lines
|
|
118 * TODO not used */
|
236
|
119 char_u sl_regions[17]; /* table with up to 8 region names plus NUL */
|
|
120 sblock_T *sl_block; /* list with allocated memory blocks */
|
|
121 int sl_error; /* error while loading */
|
|
122 };
|
|
123
|
243
|
124 /* First language that is loaded, start of the linked list of loaded
|
|
125 * languages. */
|
236
|
126 static slang_T *first_lang = NULL;
|
|
127
|
|
128 /*
|
|
129 * Structure to store an addition to a basic word.
|
243
|
130 * There are many of these, keep it small!
|
236
|
131 */
|
|
132 typedef struct addword_S addword_T;
|
|
133 struct addword_S
|
|
134 {
|
|
135 addword_T *aw_next; /* next addition */
|
|
136 char_u aw_flags; /* ADD_ flags */
|
|
137 char_u aw_region; /* region for word with this addition */
|
243
|
138 char_u aw_leadlen; /* byte length of lead in aw_word */
|
|
139 char_u aw_wordlen; /* byte length of first word in aw_word */
|
|
140 char_u aw_saveb; /* saved byte where aw_word[] is truncated at
|
|
141 end of hashtable key; NUL when not using
|
|
142 hashtable */
|
236
|
143 char_u aw_word[1]; /* text, actually longer: case-folded addition
|
|
144 plus, with ADD_KEEPCAP: keep-case addition */
|
|
145 };
|
|
146
|
243
|
147 /* Get addword_T pointer from hashitem that uses aw_word */
|
|
148 static addword_T dumaw;
|
|
149 #define HI2ADDWORD(hi) ((addword_T *)((hi)->hi_key - (dumaw.aw_word - (char_u *)&dumaw)))
|
|
150
|
236
|
151 /*
|
|
152 * Structure to store a basic word.
|
243
|
153 * There are many of these, keep it small!
|
255
|
154 * The list of prefix and suffix NRs is stored after "fw_word" to avoid the
|
|
155 * need for two extra pointers.
|
236
|
156 */
|
|
157 typedef struct fword_S fword_T;
|
|
158 struct fword_S
|
|
159 {
|
243
|
160 fword_T *fw_next; /* same basic word with different caps and/or
|
|
161 * affixes */
|
|
162 addword_T *fw_adds; /* first addword_T entry */
|
236
|
163 short_u fw_flags; /* BWF_ flags */
|
243
|
164 char_u fw_region; /* region bits */
|
|
165 char_u fw_prefixcnt; /* number of prefix NRs */
|
|
166 char_u fw_suffixcnt; /* number of suffix NRs */
|
|
167 char_u fw_word[1]; /* actually longer:
|
|
168 * 0: case folded word or keep-case word when
|
|
169 * (flags & BWF_KEEPCAP)
|
|
170 * + word length + 1: list of prefix NRs
|
|
171 * + fw_prefixcnt [* 2]: list of suffix NRs
|
|
172 */
|
236
|
173 };
|
|
174
|
|
175 /* Get fword_T pointer from hashitem that uses fw_word */
|
|
176 static fword_T dumfw;
|
|
177 #define HI2FWORD(hi) ((fword_T *)((hi)->hi_key - (dumfw.fw_word - (char_u *)&dumfw)))
|
|
178
|
|
179 #define REGION_ALL 0xff
|
|
180
|
|
181
|
|
182 /*
|
|
183 * Structure used in "b_langp", filled from 'spelllang'.
|
|
184 */
|
|
185 typedef struct langp_S
|
|
186 {
|
|
187 slang_T *lp_slang; /* info for this language (NULL for last one) */
|
|
188 int lp_region; /* bitmask for region or REGION_ALL */
|
|
189 } langp_T;
|
|
190
|
|
191 #define LANGP_ENTRY(ga, i) (((langp_T *)(ga).ga_data) + (i))
|
|
192
|
|
193 #define SP_OK 0
|
|
194 #define SP_BAD 1
|
|
195 #define SP_RARE 2
|
|
196 #define SP_LOCAL 3
|
|
197
|
|
198 /* flags used for basic words in the spell file */
|
|
199 #define BWF_VALID 0x01 /* word is valid without additions */
|
|
200 #define BWF_REGION 0x02 /* region byte follows */
|
|
201 #define BWF_ONECAP 0x04 /* first letter must be capital */
|
|
202 #define BWF_SUFFIX 0x08 /* has suffix NR list */
|
|
203 #define BWF_SECOND 0x10 /* second flags byte follows */
|
|
204
|
|
205 #define BWF_ADDS 0x0100 /* there are additions */
|
|
206 #define BWF_PREFIX 0x0200 /* has prefix NR list */
|
|
207 #define BWF_ALLCAP 0x0400 /* all letters must be capital (not used
|
|
208 for single-letter words) */
|
|
209 #define BWF_KEEPCAP 0x0800 /* Keep case as-is */
|
249
|
210 #define BWF_ADDS_M 0x1000 /* there are more than 255 additions */
|
236
|
211
|
243
|
212 #define BWF_ADDHASH 0x8000 /* Internal: use hashtab for additions */
|
|
213
|
|
214 #define NOWC_KEY (char_u *)"x" /* hashtab key used for additions without
|
|
215 any word character */
|
|
216
|
236
|
217 /* flags used for addition in the spell file */
|
|
218 #define ADD_REGION 0x02 /* region byte follows */
|
|
219 #define ADD_ONECAP 0x04 /* first letter must be capital */
|
249
|
220 #define ADD_LEADLEN 0x10 /* there is a leadlen byte */
|
|
221 #define ADD_COPYLEN 0x20 /* there is a copylen byte */
|
236
|
222 #define ADD_ALLCAP 0x40 /* all letters must be capital (not used
|
|
223 for single-letter words) */
|
|
224 #define ADD_KEEPCAP 0x80 /* fixed case */
|
|
225
|
|
226 /* Translate ADD_ flags to BWF_ flags.
|
|
227 * (Needed to keep ADD_ flags in one byte.) */
|
|
228 #define ADD2BWF(x) (((x) & 0x0f) | (((x) & 0xf0) << 4))
|
|
229
|
255
|
230 #define VIMSPELLMAGIC "VIMspell04" /* string at start of Vim spell file */
|
236
|
231 #define VIMSPELLMAGICL 10
|
|
232
|
|
233 /*
|
|
234 * Structure to store info for word matching.
|
|
235 */
|
|
236 typedef struct matchinf_S
|
|
237 {
|
|
238 langp_T *mi_lp; /* info for language and region */
|
|
239 slang_T *mi_slang; /* info for the language */
|
243
|
240
|
|
241 /* pointers to original text to be checked */
|
236
|
242 char_u *mi_line; /* start of line containing word */
|
|
243 char_u *mi_word; /* start of word being checked */
|
|
244 char_u *mi_end; /* first non-word char after mi_word */
|
243
|
245 char_u *mi_wend; /* end of matching word (is mi_end
|
236
|
246 * or further) */
|
243
|
247 char_u *mi_fend; /* next char to be added to mi_fword */
|
|
248
|
|
249 /* case-folded text */
|
|
250 char_u mi_fword[MAXWLEN + 1]; /* mi_word case-folded */
|
|
251 int mi_fendlen; /* byte length of first word in
|
|
252 mi_fword */
|
|
253 int mi_faddlen; /* byte length of text in mi_fword
|
|
254 after first word */
|
|
255 char_u *mi_cword; /* word to check, points in mi_fword */
|
|
256 char_u *mi_awend; /* after next word, to check for
|
|
257 addition (NULL when not done yet) */
|
|
258 int mi_did_awend; /* did compute mi_awend */
|
|
259
|
|
260 /* others */
|
236
|
261 int mi_result; /* result so far: SP_BAD, SP_OK, etc. */
|
|
262 int mi_capflags; /* BWF_ONECAP BWF_ALLCAP BWF_KEEPCAP */
|
|
263 } matchinf_T;
|
|
264
|
|
265 static int word_match __ARGS((matchinf_T *mip));
|
|
266 static int check_adds __ARGS((matchinf_T *mip, fword_T *fw, int req_pref, int req_suf));
|
243
|
267 static void fill_awend __ARGS((matchinf_T *mip));
|
|
268 static void fold_addchars __ARGS((matchinf_T *mip, int addlen));
|
|
269 static int supports_affix __ARGS((int cnt, char_u *afflist, int afflistlen, int nr));
|
236
|
270 static int prefix_match __ARGS((matchinf_T *mip));
|
243
|
271 static int noprefix_match __ARGS((matchinf_T *mip, char_u *pword, char_u *cstart, affitem_T *ai));
|
236
|
272 static int suffix_match __ARGS((matchinf_T *mip));
|
|
273 static int match_caps __ARGS((int flags, char_u *caseword, matchinf_T *mip, char_u *cword, char_u *end));
|
|
274 static slang_T *slang_alloc __ARGS((char_u *lang));
|
|
275 static void slang_free __ARGS((slang_T *lp));
|
|
276 static slang_T *spell_load_lang __ARGS((char_u *lang));
|
|
277 static void spell_load_file __ARGS((char_u *fname, void *cookie));
|
|
278 static void *getroom __ARGS((slang_T *lp, int *bl_used, int len));
|
|
279 static int find_region __ARGS((char_u *rp, char_u *region));
|
|
280 static int captype __ARGS((char_u *word, char_u *end));
|
|
281
|
|
282 /*
|
|
283 * Main spell-checking function.
|
|
284 * "ptr" points to the start of a word.
|
|
285 * "*attrp" is set to the attributes for a badly spelled word. For a non-word
|
|
286 * or when it's OK it remains unchanged.
|
|
287 * This must only be called when 'spelllang' is not empty.
|
|
288 * Returns the length of the word in bytes, also when it's OK, so that the
|
|
289 * caller can skip over the word.
|
|
290 */
|
|
291 int
|
|
292 spell_check(wp, line, ptr, attrp)
|
|
293 win_T *wp; /* current window */
|
|
294 char_u *line; /* start of line where "ptr" points into */
|
|
295 char_u *ptr;
|
|
296 int *attrp;
|
|
297 {
|
|
298 matchinf_T mi; /* Most things are put in "mi" so that it can
|
|
299 be passed to functions quickly. */
|
|
300
|
|
301 /* Find the end of the word. We already know that *ptr is a word char. */
|
|
302 mi.mi_word = ptr;
|
|
303 mi.mi_end = ptr;
|
|
304 do
|
|
305 {
|
|
306 mb_ptr_adv(mi.mi_end);
|
|
307 } while (*mi.mi_end != NUL && spell_iswordc(mi.mi_end));
|
|
308
|
|
309 /* A word starting with a number is always OK. */
|
|
310 if (*ptr >= '0' && *ptr <= '9')
|
|
311 return (int)(mi.mi_end - ptr);
|
|
312
|
243
|
313 /* Make case-folded copy of the word. */
|
255
|
314 (void)spell_casefold(ptr, mi.mi_end - ptr, mi.mi_fword, MAXWLEN + 1);
|
236
|
315 mi.mi_cword = mi.mi_fword;
|
243
|
316 mi.mi_fendlen = STRLEN(mi.mi_fword);
|
|
317 mi.mi_faddlen = 0;
|
|
318 mi.mi_fend = mi.mi_end;
|
|
319
|
|
320 /* Check the caps type of the word. */
|
|
321 mi.mi_capflags = captype(ptr, mi.mi_end);
|
|
322
|
|
323 /* The word is bad unless we recognize it. */
|
236
|
324 mi.mi_result = SP_BAD;
|
|
325 mi.mi_wend = mi.mi_end;
|
243
|
326
|
|
327 mi.mi_awend = NULL;
|
|
328 mi.mi_did_awend = FALSE;
|
236
|
329 mi.mi_line = line;
|
|
330
|
|
331 /*
|
|
332 * Loop over the languages specified in 'spelllang'.
|
|
333 * We check them all, because a matching word may have additions that are
|
|
334 * longer than an already found matching word.
|
|
335 */
|
|
336 for (mi.mi_lp = LANGP_ENTRY(wp->w_buffer->b_langp, 0);
|
|
337 mi.mi_lp->lp_slang != NULL; ++mi.mi_lp)
|
|
338 {
|
|
339 /*
|
|
340 * Check for a matching word.
|
|
341 * If not found or wrong region try removing prefixes (and then
|
|
342 * suffixes).
|
|
343 * If still not found or wrong region try removing suffixes.
|
|
344 */
|
|
345 mi.mi_slang = mi.mi_lp->lp_slang;
|
|
346 if (!word_match(&mi) || mi.mi_result != SP_OK)
|
|
347 if (!prefix_match(&mi) || mi.mi_result != SP_OK)
|
|
348 suffix_match(&mi);
|
|
349 }
|
|
350
|
|
351 if (mi.mi_result != SP_OK)
|
|
352 {
|
|
353 if (mi.mi_result == SP_BAD)
|
|
354 *attrp = highlight_attr[HLF_SPB];
|
|
355 else if (mi.mi_result == SP_RARE)
|
|
356 *attrp = highlight_attr[HLF_SPR];
|
|
357 else
|
|
358 *attrp = highlight_attr[HLF_SPL];
|
|
359 }
|
|
360
|
|
361 return (int)(mi.mi_wend - ptr);
|
|
362 }
|
|
363
|
|
364 /*
|
243
|
365 * Check if the word "mip->mi_word" matches.
|
|
366 * "mip->mi_fword" is the same word case-folded;
|
|
367 *
|
|
368 * This checks the word as a whole and for prefixes that include a word.
|
|
369 *
|
|
370 * Note that when called mi_fword only contains the word up to mip->mi_end,
|
|
371 * but when checking additions it gets longer.
|
236
|
372 */
|
|
373 static int
|
|
374 word_match(mip)
|
|
375 matchinf_T *mip;
|
|
376 {
|
243
|
377 hash_T fhash = hash_hash(mip->mi_fword);
|
236
|
378 hashitem_T *hi;
|
|
379 fword_T *fw;
|
|
380 int valid = FALSE;
|
243
|
381 char_u *p;
|
|
382 char_u pword[MAXWLEN + 1];
|
|
383 int charlen;
|
|
384 int capflags_save;
|
|
385 affitem_T *ai;
|
|
386 char_u *cstart;
|
|
387 int addlen;
|
|
388 int n;
|
|
389 char_u *save_end;
|
|
390 int cc;
|
|
391
|
|
392 hi = hash_lookup(&mip->mi_slang->sl_words, mip->mi_fword, fhash);
|
|
393 if (!HASHITEM_EMPTY(hi))
|
|
394 {
|
|
395 /*
|
|
396 * Find a basic word for which the case of "mi_word" is correct.
|
|
397 * If it is, check additions and use the longest one.
|
|
398 */
|
|
399 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
400 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
401 mip->mi_word, mip->mi_end))
|
|
402 valid |= check_adds(mip, fw, -1, -1);
|
|
403 }
|
236
|
404
|
|
405 /*
|
243
|
406 * Try finding a matching preword for "mip->mi_word". These are
|
|
407 * prefixes that have a non-word character after a word character:
|
|
408 * "d'", "de-", "'s-", "l'de-". But not "'s".
|
|
409 * Also need to do this when a matching word was already found, because we
|
|
410 * might find a longer match this way (French: "qu" and "qu'a-t-elle").
|
255
|
411 * The check above may have added characters to mi_fword, thus we need to
|
|
412 * truncate it after the basic word for the hash lookup.
|
236
|
413 */
|
243
|
414 cc = mip->mi_fword[mip->mi_fendlen];
|
|
415 mip->mi_fword[mip->mi_fendlen] = NUL;
|
|
416 hi = hash_lookup(&mip->mi_slang->sl_prewords, mip->mi_fword, fhash);
|
|
417 mip->mi_fword[mip->mi_fendlen] = cc;
|
|
418 if (!HASHITEM_EMPTY(hi))
|
|
419 {
|
|
420 capflags_save = mip->mi_capflags;
|
|
421
|
|
422 /* Go through the list of matching prewords. */
|
|
423 for (ai = HI2AI(hi); ai != NULL; ai = ai->ai_next)
|
|
424 {
|
|
425 /* Check that the lead string matches before the word. */
|
|
426 p = ai->ai_add + ai->ai_addlen + ai->ai_choplen + 2;
|
|
427 if (ai->ai_leadlen > 0)
|
|
428 {
|
|
429 if (mip->mi_word - mip->mi_line < ai->ai_leadlen
|
|
430 || STRNCMP(mip->mi_word - ai->ai_leadlen, p,
|
|
431 ai->ai_leadlen) != 0)
|
|
432 continue;
|
|
433 p += ai->ai_leadlen + 1; /* advance "p" to tail */
|
|
434 }
|
|
435 else
|
|
436 ++p; /* advance "p" to tail */
|
|
437
|
|
438 /* Check that the tail string matches after the word. Need
|
|
439 * to fold case first. */
|
|
440 if (ai->ai_taillen > 0)
|
|
441 {
|
|
442 if (ai->ai_taillen >= mip->mi_faddlen)
|
|
443 {
|
|
444 fold_addchars(mip, ai->ai_taillen);
|
|
445 if (ai->ai_taillen > mip->mi_faddlen)
|
|
446 continue; /* not enough chars, can't match */
|
|
447 }
|
|
448 if (STRNCMP(mip->mi_fword + mip->mi_fendlen,
|
|
449 p, ai->ai_taillen) != 0)
|
|
450 continue;
|
|
451 }
|
|
452
|
|
453 /*
|
|
454 * This preword matches. Remove the preword and check that
|
|
455 * the resulting word exits.
|
|
456 */
|
|
457
|
|
458 /* Find the place in the original word where the tail ends,
|
|
459 * needed for case checks. */
|
|
460 #ifdef FEAT_MBYTE
|
|
461 charlen = mb_charlen(p);
|
|
462 #else
|
|
463 charlen = ai->ai_taillen;
|
|
464 #endif
|
|
465 cstart = mip->mi_end;
|
|
466 for (n = 0; n < charlen; ++n)
|
|
467 mb_ptr_adv(cstart);
|
|
468
|
|
469 /* The new word starts with the chop. Then add up to the next
|
|
470 * non-word char. */
|
|
471 mch_memmove(pword, ai->ai_add + ai->ai_addlen + 1,
|
|
472 ai->ai_choplen);
|
|
473 p = mip->mi_fword + mip->mi_fendlen + ai->ai_taillen;
|
|
474 addlen = ai->ai_taillen;
|
|
475 while (spell_iswordc(p))
|
|
476 {
|
|
477 ++charlen;
|
|
478 #ifdef FEAT_MBYTE
|
|
479 addlen += (*mb_ptr2len_check)(p);
|
|
480 #else
|
|
481 ++addlen;
|
|
482 #endif
|
|
483 mb_ptr_adv(p);
|
|
484 if (addlen >= mip->mi_faddlen)
|
|
485 {
|
|
486 /* Get more folded characters in mip->mi_fword. */
|
|
487 fold_addchars(mip, addlen);
|
|
488 if (addlen >= mip->mi_faddlen)
|
|
489 break; /* not enough chars, can't match */
|
|
490 }
|
|
491 }
|
|
492 mch_memmove(pword + ai->ai_choplen,
|
|
493 mip->mi_fword + mip->mi_fendlen + ai->ai_taillen,
|
|
494 addlen - ai->ai_taillen);
|
|
495 pword[ai->ai_choplen + addlen - ai->ai_taillen] = NUL;
|
|
496
|
|
497 /* Need to set mi_end to find additions. Also set mi_fendlen
|
|
498 * and mi_faddlen. */
|
|
499 save_end = mip->mi_end;
|
|
500 while (--charlen >= 0)
|
|
501 mb_ptr_adv(mip->mi_end);
|
|
502 mip->mi_fendlen += addlen;
|
|
503 mip->mi_faddlen -= addlen;
|
|
504
|
|
505 /* Find the word "pword", caseword "cstart". */
|
|
506 n = noprefix_match(mip, pword, cstart, ai);
|
|
507 mip->mi_end = save_end;
|
|
508 mip->mi_fendlen -= addlen;
|
|
509 mip->mi_faddlen += addlen;
|
|
510 if (n)
|
|
511 valid = TRUE;
|
|
512
|
|
513 /* If we found a valid word, we still need to try other
|
|
514 * suffixes, because it may have an addition that's longer. */
|
|
515 }
|
|
516
|
|
517 mip->mi_capflags = capflags_save;
|
|
518 }
|
236
|
519
|
|
520 return valid;
|
|
521 }
|
|
522
|
|
523 /*
|
|
524 * Check a matching basic word for additions.
|
|
525 * Return TRUE if we have a valid match.
|
|
526 */
|
|
527 static int
|
|
528 check_adds(mip, fw, req_pref, req_suf)
|
|
529 matchinf_T *mip;
|
|
530 fword_T *fw;
|
|
531 int req_pref; /* required prefix nr, -1 if none */
|
|
532 int req_suf; /* required suffix nr, -1 if none */
|
|
533 {
|
|
534 int valid = FALSE;
|
|
535 addword_T *aw;
|
243
|
536 addword_T *naw = NULL;
|
236
|
537 char_u *p;
|
|
538 int addlen;
|
243
|
539 int cc;
|
|
540 hashitem_T *hi;
|
|
541 char_u *cp = NULL;
|
|
542 int n;
|
|
543
|
|
544 /* Check if required prefixes and suffixes are supported. These are on
|
|
545 * the basic word, not on each addition. */
|
|
546 if (req_pref >= 0 || req_suf >= 0)
|
|
547 {
|
|
548 /* Prefix NRs are stored just after the word in fw_word. */
|
|
549 cp = fw->fw_word + STRLEN(fw->fw_word) + 1;
|
|
550 if (req_pref >= 0 && !supports_affix(mip->mi_slang->sl_prefcnt,
|
|
551 cp, fw->fw_prefixcnt, req_pref))
|
|
552 return FALSE;
|
|
553 if (req_suf >= 0)
|
|
554 {
|
|
555 /* Suffix NRs are stored just after the Prefix NRs. */
|
|
556 if (fw->fw_prefixcnt > 0)
|
|
557 {
|
|
558 if (mip->mi_slang->sl_prefcnt > 256)
|
|
559 cp += fw->fw_prefixcnt * 2;
|
|
560 else
|
|
561 cp += fw->fw_prefixcnt;
|
|
562 }
|
|
563 if (!supports_affix(mip->mi_slang->sl_suffcnt,
|
|
564 cp, fw->fw_suffixcnt, req_suf))
|
|
565 return FALSE;
|
|
566 }
|
|
567 }
|
|
568
|
|
569 /* A word may be valid without an addition. */
|
|
570 if (fw->fw_flags & BWF_VALID)
|
236
|
571 {
|
|
572 valid = TRUE;
|
|
573 if (mip->mi_result != SP_OK)
|
|
574 {
|
|
575 if ((fw->fw_region & mip->mi_lp->lp_region) == 0)
|
|
576 mip->mi_result = SP_LOCAL;
|
|
577 else
|
|
578 mip->mi_result = SP_OK;
|
|
579 }
|
243
|
580 /* Set word end, required when matching a word after a preword. */
|
|
581 if (mip->mi_wend < mip->mi_end)
|
|
582 mip->mi_wend = mip->mi_end;
|
236
|
583 }
|
|
584
|
|
585 /*
|
|
586 * Check additions, both before and after the word.
|
|
587 * This may make the word longer, thus we also need to check
|
|
588 * when we already found a matching word.
|
243
|
589 * When the BWF_ADDHASH flag is present then fw_adds points to a hashtable
|
|
590 * for quick lookup. Otherwise it points to the list of all possible
|
|
591 * additions.
|
236
|
592 */
|
243
|
593 if (fw->fw_flags & BWF_ADDHASH)
|
236
|
594 {
|
243
|
595 /* Locate the text up to the next end-of-word. */
|
|
596 if (!mip->mi_did_awend)
|
|
597 fill_awend(mip);
|
|
598 if (mip->mi_awend == NULL)
|
|
599 return valid; /* there is no next word */
|
|
600
|
|
601 cc = *mip->mi_awend;
|
|
602 *mip->mi_awend = NUL;
|
|
603 hi = hash_find((hashtab_T *)fw->fw_adds,
|
|
604 mip->mi_fword + mip->mi_fendlen);
|
|
605 *mip->mi_awend = cc;
|
|
606 if (HASHITEM_EMPTY(hi))
|
|
607 return valid; /* no matching addition */
|
|
608 aw = HI2ADDWORD(hi);
|
|
609
|
|
610 /* Also check additions without word characters. If they are there,
|
|
611 * skip the first dummy entry. */
|
|
612 hi = hash_find((hashtab_T *)fw->fw_adds, NOWC_KEY);
|
|
613 if (!HASHITEM_EMPTY(hi))
|
|
614 naw = HI2ADDWORD(hi)->aw_next;
|
|
615 }
|
|
616 else
|
|
617 aw = fw->fw_adds;
|
|
618
|
|
619 for ( ; ; aw = aw->aw_next)
|
|
620 {
|
|
621 if (aw == NULL)
|
|
622 {
|
|
623 /* At end of list: may also try additions without word chars. */
|
|
624 if (naw == NULL)
|
|
625 break;
|
|
626 aw = naw;
|
|
627 naw = NULL;
|
|
628 }
|
|
629
|
236
|
630 if (aw->aw_leadlen > 0)
|
|
631 {
|
|
632 /* There is a leader, verify that it matches. */
|
|
633 if (aw->aw_leadlen > mip->mi_word - mip->mi_line
|
|
634 || STRNCMP(mip->mi_word - aw->aw_leadlen,
|
|
635 aw->aw_word, aw->aw_leadlen) != 0)
|
|
636 continue;
|
|
637 if (mip->mi_word - aw->aw_leadlen > mip->mi_line)
|
|
638 {
|
|
639 /* There must not be a word character just before the
|
|
640 * leader. */
|
|
641 p = mip->mi_word - aw->aw_leadlen;
|
|
642 mb_ptr_back(mip->mi_line, p);
|
|
643 if (spell_iswordc(p))
|
|
644 continue;
|
|
645 }
|
|
646 /* Leader matches. Addition is rest of "aw_word". */
|
|
647 p = aw->aw_word + aw->aw_leadlen;
|
|
648 }
|
|
649 else
|
|
650 /* No leader, use whole of "aw_word" for addition. */
|
|
651 p = aw->aw_word;
|
|
652
|
|
653 addlen = aw->aw_wordlen - aw->aw_leadlen;
|
|
654 if (addlen > 0)
|
|
655 {
|
|
656 /* Check for matching addition and no word character after it.
|
|
657 * First make sure we have enough case-folded chars to compare
|
|
658 * with. */
|
243
|
659 if (addlen >= mip->mi_faddlen)
|
|
660 fold_addchars(mip, addlen);
|
|
661
|
|
662 /* Put back the saved char, if needed. */
|
|
663 if (aw->aw_saveb != NUL)
|
236
|
664 {
|
243
|
665 cp = p + STRLEN(p);
|
|
666 *cp = aw->aw_saveb;
|
236
|
667 }
|
243
|
668 n = STRNCMP(mip->mi_fword + mip->mi_fendlen, p, addlen);
|
|
669 if (aw->aw_saveb != NUL)
|
|
670 *cp = NUL;
|
|
671
|
|
672 if (n != 0 || (mip->mi_fword[mip->mi_fendlen + addlen] != NUL
|
|
673 && spell_iswordc(mip->mi_fword + mip->mi_fendlen + addlen)))
|
236
|
674 continue;
|
|
675
|
|
676 /* Compute the length in the original word, before case folding. */
|
|
677 #ifdef FEAT_MBYTE
|
|
678 if (has_mbyte)
|
|
679 {
|
|
680 int l;
|
|
681
|
|
682 p = mip->mi_end;
|
243
|
683 for (l = 0; l < addlen; l += (*mb_ptr2len_check)(mip->mi_fword
|
|
684 + mip->mi_fendlen + l))
|
236
|
685 mb_ptr_adv(p);
|
|
686 addlen = p - mip->mi_end;
|
|
687 }
|
|
688 #endif
|
|
689
|
|
690 /* Check case of the addition. */
|
|
691 if (!match_caps(ADD2BWF(aw->aw_flags),
|
|
692 aw->aw_word + aw->aw_wordlen + 1, mip,
|
|
693 mip->mi_end, mip->mi_end + addlen))
|
|
694 continue;
|
|
695 }
|
|
696
|
|
697 /* Match! Use the new length if it's longer. */
|
|
698 if (mip->mi_wend < mip->mi_end + addlen)
|
|
699 mip->mi_wend = mip->mi_end + addlen;
|
|
700
|
|
701 valid = TRUE;
|
|
702 if (mip->mi_result != SP_OK)
|
|
703 {
|
|
704 if ((aw->aw_region & mip->mi_lp->lp_region) == 0)
|
|
705 mip->mi_result = SP_LOCAL;
|
|
706 else
|
|
707 mip->mi_result = SP_OK;
|
|
708 }
|
|
709 }
|
|
710
|
|
711 return valid;
|
|
712 }
|
|
713
|
|
714 /*
|
243
|
715 * Locate the text up to the next end-of-word after mip->mi_end.
|
|
716 */
|
|
717 static void
|
|
718 fill_awend(mip)
|
|
719 matchinf_T *mip;
|
|
720 {
|
|
721 char_u *p = mip->mi_end;
|
|
722 int addlen = 0;
|
|
723 int find_word = TRUE;
|
|
724
|
|
725 mip->mi_did_awend = TRUE;
|
|
726 if (mip->mi_faddlen == 0)
|
|
727 fold_addchars(mip, 0); /* need to fold first char */
|
|
728
|
|
729 /* 1: find_word == TRUE: skip over non-word characters after mi_end.
|
|
730 * 2: find_word == FALSE: skip over following word characters. */
|
|
731 for (p = mip->mi_fword + mip->mi_fendlen; *p != NUL; mb_ptr_adv(p))
|
|
732 {
|
|
733 if (spell_iswordc(p) == find_word)
|
|
734 {
|
|
735 if (!find_word)
|
|
736 break; /* done */
|
|
737 find_word = !find_word;
|
|
738 }
|
|
739 #ifdef FEAT_MBYTE
|
|
740 addlen += (*mb_ptr2len_check)(p);
|
|
741 #else
|
|
742 ++addlen;
|
|
743 #endif
|
|
744 if (addlen >= mip->mi_faddlen)
|
|
745 fold_addchars(mip, addlen); /* need to fold more chars */
|
|
746 }
|
|
747
|
|
748 /* If there are extra chars store the result. */
|
|
749 if (addlen != 0)
|
|
750 mip->mi_awend = p;
|
|
751 }
|
|
752
|
|
753 /*
|
|
754 * Fold enough characters of the checked text to be able to compare with an
|
|
755 * addition of length "addlen" plus one character (to be able to check the
|
|
756 * next character to be a non-word char).
|
|
757 * When there are not enough characters (end of line) mip->mi_faddlen will be
|
|
758 * smaller than "addlen".
|
|
759 */
|
|
760 static void
|
|
761 fold_addchars(mip, addlen)
|
|
762 matchinf_T *mip;
|
|
763 int addlen;
|
|
764 {
|
|
765 int l;
|
|
766 char_u *p = mip->mi_fword + mip->mi_fendlen;
|
|
767
|
|
768 while (mip->mi_faddlen <= addlen)
|
|
769 {
|
|
770 if (*mip->mi_fend == NUL) /* end of the line */
|
|
771 {
|
|
772 p[mip->mi_faddlen] = NUL;
|
|
773 break;
|
|
774 }
|
|
775 #ifdef FEAT_MBYTE
|
|
776 if (has_mbyte)
|
|
777 l = (*mb_ptr2len_check)(mip->mi_fend);
|
|
778 else
|
|
779 #endif
|
|
780 l = 1;
|
255
|
781 (void)spell_casefold(mip->mi_fend, l, p + mip->mi_faddlen,
|
243
|
782 MAXWLEN - mip->mi_fendlen - mip->mi_faddlen);
|
|
783 mip->mi_fend += l;
|
|
784 mip->mi_faddlen += STRLEN(p + mip->mi_faddlen);
|
|
785 }
|
|
786 }
|
|
787
|
|
788 /*
|
|
789 * Return TRUE if affix "nr" appears in affix list "afflist[afflistlen]".
|
236
|
790 */
|
|
791 static int
|
243
|
792 supports_affix(cnt, afflist, afflistlen, nr)
|
|
793 int cnt; /* total affix NR count */
|
|
794 char_u *afflist;
|
|
795 int afflistlen; /* affix count in "afflist" */
|
236
|
796 int nr;
|
|
797 {
|
243
|
798 char_u *pc = afflist;
|
236
|
799 int i;
|
243
|
800 int nr_msb, nr_lsb;
|
236
|
801
|
|
802 if (cnt <= 256)
|
|
803 {
|
243
|
804 /* one byte affix numbers */
|
|
805 for (i = afflistlen; --i >= 0; )
|
236
|
806 if (*pc++ == nr)
|
|
807 return TRUE;
|
|
808 }
|
|
809 else
|
|
810 {
|
243
|
811 /* two byte affix numbers, MSB first */
|
|
812 nr_msb = (unsigned)nr >> 8;
|
|
813 nr_lsb = nr & 0xff;
|
|
814 for (i = afflistlen; --i >= 0; )
|
|
815 {
|
|
816 if (*pc++ == nr_msb && *pc == nr_lsb)
|
236
|
817 return TRUE;
|
243
|
818 ++pc;
|
|
819 }
|
236
|
820 }
|
|
821 return FALSE;
|
|
822 }
|
|
823
|
|
824 /*
|
|
825 * Try finding a match for "mip->mi_cword" by removing prefixes.
|
|
826 */
|
|
827 static int
|
|
828 prefix_match(mip)
|
|
829 matchinf_T *mip;
|
|
830 {
|
|
831 int len = 0;
|
|
832 int charlen = 0;
|
|
833 int cc;
|
|
834 affitem_T *ai;
|
|
835 char_u pword[MAXWLEN + 1];
|
|
836 hashtab_T *ht;
|
|
837 hashitem_T *hi;
|
|
838 int found_valid = FALSE;
|
|
839 int cstart_charlen = 0;
|
|
840 char_u *cstart = mip->mi_word;
|
|
841 int capflags_save = mip->mi_capflags;
|
|
842
|
|
843 /*
|
|
844 * Check for prefixes with different character lengths.
|
|
845 * Start with zero length (only chop off).
|
|
846 */
|
|
847 for (charlen = 0; charlen <= mip->mi_slang->sl_preftab.ga_len; ++charlen)
|
|
848 {
|
|
849 if (charlen > 0)
|
|
850 {
|
|
851 #ifdef FEAT_MBYTE
|
|
852 if (has_mbyte)
|
243
|
853 len += (*mb_ptr2len_check)(mip->mi_cword + len);
|
236
|
854 else
|
|
855 #endif
|
|
856 len += 1;
|
|
857 }
|
|
858 if (mip->mi_cword[len] == NUL) /* end of word, no prefix possible */
|
|
859 break;
|
|
860
|
|
861 if (charlen == 0)
|
|
862 ai = mip->mi_slang->sl_prefzero;
|
|
863 else
|
|
864 {
|
|
865 /* Get pointer to hashtab for prefix of this many chars. */
|
|
866 ht = ((hashtab_T *)mip->mi_slang->sl_preftab.ga_data) + charlen - 1;
|
|
867 if (ht->ht_used == 0)
|
|
868 continue;
|
|
869
|
|
870 cc = mip->mi_cword[len];
|
|
871 mip->mi_cword[len] = NUL;
|
|
872 hi = hash_find(ht, mip->mi_cword);
|
|
873 mip->mi_cword[len] = cc;
|
|
874
|
|
875 if (HASHITEM_EMPTY(hi))
|
|
876 ai = NULL;
|
|
877 else
|
|
878 ai = HI2AI(hi);
|
|
879 }
|
|
880
|
|
881 /* Loop over all matching prefixes. */
|
|
882 for ( ; ai != NULL; ai = ai->ai_next)
|
|
883 {
|
243
|
884 /* Create the basic word from the chop string and the word after
|
|
885 * the matching add string. */
|
|
886 mch_memmove(pword, ai->ai_add + ai->ai_addlen + 1, ai->ai_choplen);
|
|
887 mch_memmove(pword + ai->ai_choplen, mip->mi_cword + ai->ai_addlen,
|
|
888 mip->mi_fendlen - ai->ai_addlen);
|
|
889 pword[mip->mi_fendlen - ai->ai_addlen] = NUL;
|
236
|
890
|
|
891 /* Adjust the word start for case checks, we only check the
|
|
892 * part after the prefix. */
|
|
893 while (cstart_charlen < charlen)
|
|
894 {
|
|
895 mb_ptr_adv(cstart);
|
|
896 ++cstart_charlen;
|
|
897 }
|
|
898
|
243
|
899 /* Find the word "pword", caseword "cstart". */
|
|
900 found_valid |= noprefix_match(mip, pword, cstart, ai);
|
|
901
|
|
902 if (found_valid && mip->mi_result == SP_OK)
|
236
|
903 {
|
243
|
904 /* Found a valid word, no need to try other suffixes. */
|
|
905 mip->mi_capflags = capflags_save;
|
|
906 return TRUE;
|
236
|
907 }
|
|
908 }
|
|
909 }
|
|
910
|
|
911 mip->mi_capflags = capflags_save;
|
|
912 return FALSE;
|
|
913 }
|
|
914
|
|
915 /*
|
243
|
916 * Check for matching word after removing a prefix.
|
|
917 * Return TRUE if found.
|
|
918 */
|
|
919 static int
|
|
920 noprefix_match(mip, pword, cstart, ai)
|
|
921 matchinf_T *mip;
|
|
922 char_u *pword; /* case-folded word */
|
|
923 char_u *cstart; /* original word after removed prefix */
|
|
924 affitem_T *ai; /* the prefix item */
|
|
925 {
|
|
926 hashitem_T *hi;
|
|
927 fword_T *fw;
|
|
928 int found_valid = FALSE;
|
|
929 char_u *word;
|
|
930 int i;
|
|
931 int fendlen;
|
|
932
|
|
933 /* Removing the prefix may change the caps, e.g. for
|
|
934 * "deAlf" removing "de" makes it ONECAP. */
|
|
935 mip->mi_capflags = captype(cstart, mip->mi_end);
|
|
936
|
|
937 /* Find the basic word. */
|
|
938 hi = hash_find(&mip->mi_slang->sl_words, pword);
|
|
939 if (!HASHITEM_EMPTY(hi))
|
|
940 {
|
|
941 /* Check if the word supports this prefix. */
|
|
942 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
943 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
944 cstart, mip->mi_end))
|
|
945 found_valid |= check_adds(mip, fw, ai->ai_nr, -1);
|
|
946
|
|
947 if (found_valid && mip->mi_result == SP_OK)
|
|
948 /* Found a valid word, no need to try other suffixes. */
|
|
949 return TRUE;
|
|
950 }
|
|
951
|
|
952 /* No matching basic word without prefix. When combining is
|
|
953 * allowed try with suffixes. */
|
|
954 if (ai->ai_flags & AFF_COMBINE)
|
|
955 {
|
|
956 /* Pass the word with prefix removed to suffix_match(). */
|
|
957 mip->mi_cword = pword;
|
|
958 word = mip->mi_word;
|
|
959 mip->mi_word = cstart;
|
|
960 fendlen = mip->mi_fendlen;
|
|
961 mip->mi_fendlen = STRLEN(pword);
|
|
962 i = suffix_match(mip);
|
|
963 mip->mi_cword = mip->mi_fword;
|
|
964 mip->mi_word = word;
|
|
965 mip->mi_fendlen = fendlen;
|
|
966 if (i)
|
|
967 return TRUE;
|
|
968 }
|
|
969
|
|
970 return FALSE;
|
|
971 }
|
|
972
|
|
973 /*
|
236
|
974 * Try finding a match for "mip->mi_cword" by removing suffixes.
|
|
975 */
|
|
976 static int
|
|
977 suffix_match(mip)
|
|
978 matchinf_T *mip;
|
|
979 {
|
|
980 char_u *sufp;
|
243
|
981 char_u *endw = mip->mi_cword + mip->mi_fendlen;
|
|
982 int endw_c = *endw;
|
236
|
983 int charlen;
|
|
984 affitem_T *ai;
|
|
985 char_u pword[MAXWLEN + 1];
|
|
986 fword_T *fw;
|
|
987 hashtab_T *ht;
|
|
988 hashitem_T *hi;
|
|
989 int tlen;
|
|
990 int cend_charlen = 0;
|
|
991 char_u *cend = mip->mi_end;
|
|
992 int found_valid = FALSE;
|
|
993 int capflags_save = mip->mi_capflags;
|
|
994
|
|
995 /*
|
|
996 * Try suffixes of different length, starting with an empty suffix (chop
|
|
997 * only, thus adds something).
|
|
998 * Stop checking if there are no suffixes with so many characters.
|
|
999 */
|
243
|
1000 sufp = endw;
|
255
|
1001 *endw = NUL; /* truncate after possible suffix */
|
|
1002
|
236
|
1003 for (charlen = 0; charlen <= mip->mi_slang->sl_sufftab.ga_len; ++charlen)
|
|
1004 {
|
|
1005 /* Move the pointer to the possible suffix back one character, unless
|
|
1006 * doing the first round (empty suffix). */
|
|
1007 if (charlen > 0)
|
|
1008 {
|
|
1009 mb_ptr_back(mip->mi_cword, sufp);
|
|
1010 if (sufp <= mip->mi_cword) /* start of word, no suffix possible */
|
|
1011 break;
|
|
1012 }
|
|
1013
|
|
1014 if (charlen == 0)
|
|
1015 ai = mip->mi_slang->sl_suffzero;
|
|
1016 else
|
|
1017 {
|
|
1018 /* Get pointer to hashtab for suffix of this many chars. */
|
|
1019 ht = ((hashtab_T *)mip->mi_slang->sl_sufftab.ga_data) + charlen - 1;
|
|
1020 if (ht->ht_used == 0)
|
|
1021 continue;
|
|
1022
|
|
1023 hi = hash_find(ht, sufp);
|
|
1024 if (HASHITEM_EMPTY(hi))
|
|
1025 ai = NULL;
|
|
1026 else
|
|
1027 ai = HI2AI(hi);
|
|
1028 }
|
|
1029
|
|
1030 if (ai != NULL)
|
|
1031 {
|
|
1032 /* Found a list of matching suffixes. Now check that there is one
|
|
1033 * we can use. */
|
|
1034 tlen = sufp - mip->mi_cword; /* length of word without suffix */
|
|
1035 mch_memmove(pword, mip->mi_cword, tlen);
|
255
|
1036 *endw = endw_c;
|
236
|
1037
|
|
1038 for ( ; ai != NULL; ai = ai->ai_next)
|
|
1039 {
|
|
1040 /* Found a matching suffix. Create the basic word by removing
|
|
1041 * the suffix and adding the chop string. */
|
|
1042 if (ai->ai_choplen == 0)
|
|
1043 pword[tlen] = NUL;
|
|
1044 else
|
243
|
1045 mch_memmove(pword + tlen, ai->ai_add + ai->ai_addlen + 1,
|
|
1046 ai->ai_choplen + 1);
|
236
|
1047
|
|
1048 /* Find the basic word. */
|
|
1049 hi = hash_find(&mip->mi_slang->sl_words, pword);
|
|
1050 if (!HASHITEM_EMPTY(hi))
|
|
1051 {
|
|
1052 /* Adjust the end for case checks, we only check the part
|
|
1053 * before the suffix. */
|
|
1054 while (cend_charlen < charlen)
|
|
1055 {
|
|
1056 mb_ptr_back(mip->mi_word, cend);
|
|
1057 ++cend_charlen;
|
|
1058 }
|
|
1059
|
|
1060 /* Removing the suffix may change the caps, e.g. for
|
|
1061 * "UFOs" removing 's' makes it ALLCAP. */
|
|
1062 mip->mi_capflags = captype(mip->mi_word, cend);
|
|
1063
|
|
1064 /* Check if the word supports this suffix. */
|
|
1065 for (fw = HI2FWORD(hi); fw != NULL; fw = fw->fw_next)
|
|
1066 if (match_caps(fw->fw_flags, fw->fw_word, mip,
|
|
1067 mip->mi_word, cend))
|
|
1068 found_valid |= check_adds(mip, fw, -1, ai->ai_nr);
|
|
1069
|
|
1070 if (found_valid && mip->mi_result == SP_OK)
|
|
1071 {
|
|
1072 /* Found a valid word, no need to try other suffixes. */
|
|
1073 mip->mi_capflags = capflags_save;
|
|
1074 return TRUE;
|
|
1075 }
|
|
1076 }
|
|
1077 }
|
255
|
1078
|
|
1079 *endw = NUL; /* truncate after possible suffix */
|
236
|
1080 }
|
|
1081 }
|
|
1082
|
255
|
1083 *endw = endw_c;
|
236
|
1084 mip->mi_capflags = capflags_save;
|
|
1085 return FALSE;
|
|
1086 }
|
|
1087
|
|
1088 /*
|
|
1089 * Return TRUE if case of "cword" meets the requirements of case flags
|
|
1090 * "flags".
|
|
1091 */
|
|
1092 static int
|
|
1093 match_caps(flags, caseword, mip, cword, end)
|
|
1094 int flags; /* flags required by basic word or addition */
|
|
1095 char_u *caseword; /* word with case as required */
|
|
1096 matchinf_T *mip;
|
|
1097 char_u *cword; /* word to compare against "caseword" */
|
|
1098 char_u *end; /* end of "cword" */
|
|
1099 {
|
|
1100 char_u *p;
|
|
1101 int c;
|
|
1102 int len;
|
|
1103 int capflags = mip->mi_capflags; /* flags of checked word */
|
|
1104 int past_second;
|
|
1105
|
|
1106 if ((capflags & BWF_KEEPCAP) == 0 && end > mip->mi_end)
|
|
1107 {
|
243
|
1108 /* If "end" is past "mip->mi_end" we need to adjust the caps type for
|
|
1109 * characters after the basic word. */
|
236
|
1110 #ifdef FEAT_MBYTE
|
|
1111 past_second = (mip->mi_word + (*mb_ptr2len_check)(mip->mi_word)
|
|
1112 < mip->mi_end);
|
|
1113 #else
|
|
1114 past_second = mip->mi_word + 1 < mip->mi_end;
|
|
1115 #endif
|
|
1116 for (p = mip->mi_end; p < end; )
|
|
1117 {
|
|
1118 if (!spell_iswordc(p))
|
|
1119 mb_ptr_adv(p);
|
|
1120 else
|
|
1121 {
|
|
1122 #ifdef FEAT_MBYTE
|
|
1123 if (has_mbyte)
|
|
1124 c = mb_ptr2char_adv(&p);
|
|
1125 else
|
|
1126 #endif
|
|
1127 c = *p++;
|
255
|
1128 if (spell_isupper(c))
|
236
|
1129 {
|
|
1130 if (capflags == 0 || (capflags & BWF_ONECAP))
|
|
1131 {
|
|
1132 capflags = BWF_KEEPCAP; /* lU or UlU */
|
|
1133 break;
|
|
1134 }
|
|
1135 }
|
|
1136 else
|
|
1137 {
|
|
1138 if (capflags & BWF_ALLCAP)
|
|
1139 {
|
|
1140 if (past_second)
|
|
1141 {
|
|
1142 capflags = BWF_KEEPCAP; /* UUl */
|
|
1143 break;
|
|
1144 }
|
|
1145 capflags = BWF_ONECAP; /* Uu */
|
|
1146 }
|
|
1147 }
|
|
1148 past_second = TRUE;
|
|
1149 }
|
|
1150 }
|
|
1151 }
|
|
1152
|
|
1153 if (capflags == BWF_ALLCAP)
|
|
1154 return TRUE; /* All caps is always OK. */
|
|
1155
|
|
1156 if (flags & BWF_KEEPCAP)
|
|
1157 {
|
|
1158 len = STRLEN(caseword);
|
|
1159 return (len == end - cword && STRNCMP(caseword, cword, len) == 0);
|
|
1160 }
|
|
1161
|
|
1162 if (flags & BWF_ALLCAP)
|
|
1163 return FALSE; /* need ALLCAP, already checked above */
|
|
1164
|
|
1165 if (flags & BWF_ONECAP)
|
|
1166 return capflags == BWF_ONECAP;
|
|
1167
|
|
1168 return capflags != BWF_KEEPCAP; /* no case check, only KEEPCAP is bad */
|
|
1169 }
|
|
1170
|
|
1171 /*
|
|
1172 * Move to next spell error.
|
|
1173 * Return OK if found, FAIL otherwise.
|
|
1174 */
|
|
1175 int
|
|
1176 spell_move_to(dir, allwords)
|
|
1177 int dir; /* FORWARD or BACKWARD */
|
|
1178 int allwords; /* TRUE for "[s" and "]s" */
|
|
1179 {
|
249
|
1180 linenr_T lnum;
|
|
1181 pos_T found_pos;
|
236
|
1182 char_u *line;
|
|
1183 char_u *p;
|
|
1184 int wc;
|
|
1185 int nwc;
|
|
1186 int attr = 0;
|
|
1187 int len;
|
249
|
1188 int has_syntax = syntax_present(curbuf);
|
|
1189 int col;
|
|
1190 int can_spell;
|
236
|
1191
|
|
1192 if (!curwin->w_p_spell || *curwin->w_buffer->b_p_spl == NUL)
|
|
1193 {
|
|
1194 EMSG(_("E756: Spell checking not enabled"));
|
|
1195 return FAIL;
|
|
1196 }
|
|
1197
|
249
|
1198 /*
|
|
1199 * Start looking for bad word at the start of the line, because we can't
|
|
1200 * start halfway a word, we don't know where it starts or ends.
|
|
1201 *
|
|
1202 * When searching backwards, we continue in the line to find the last
|
|
1203 * bad word (in the cursor line: before the cursor).
|
|
1204 */
|
|
1205 lnum = curwin->w_cursor.lnum;
|
|
1206 found_pos.lnum = 0;
|
236
|
1207
|
|
1208 while (!got_int)
|
|
1209 {
|
249
|
1210 line = ml_get(lnum);
|
|
1211 p = line;
|
|
1212 wc = FALSE;
|
|
1213
|
236
|
1214 while (*p != NUL)
|
|
1215 {
|
|
1216 nwc = spell_iswordc(p);
|
|
1217 if (!wc && nwc)
|
|
1218 {
|
249
|
1219 /* When searching backward don't search after the cursor. */
|
|
1220 if (dir == BACKWARD
|
|
1221 && lnum == curwin->w_cursor.lnum
|
|
1222 && (colnr_T)(p - line) >= curwin->w_cursor.col)
|
|
1223 break;
|
|
1224
|
236
|
1225 /* start of word */
|
|
1226 len = spell_check(curwin, line, p, &attr);
|
249
|
1227
|
236
|
1228 if (attr != 0)
|
|
1229 {
|
249
|
1230 /* We found a bad word. Check the attribute. */
|
|
1231 /* TODO: check for syntax @Spell cluster. */
|
|
1232 if (allwords || attr == highlight_attr[HLF_SPB])
|
236
|
1233 {
|
249
|
1234 /* When searching forward only accept a bad word after
|
|
1235 * the cursor. */
|
|
1236 if (dir == BACKWARD
|
|
1237 || lnum > curwin->w_cursor.lnum
|
|
1238 || (lnum == curwin->w_cursor.lnum
|
|
1239 && (colnr_T)(p - line)
|
|
1240 > curwin->w_cursor.col))
|
|
1241 {
|
|
1242 if (has_syntax)
|
|
1243 {
|
|
1244 col = p - line;
|
|
1245 (void)syn_get_id(lnum, (colnr_T)col,
|
|
1246 FALSE, &can_spell);
|
|
1247
|
|
1248 /* have to get the line again, a multi-line
|
|
1249 * regexp may make it invalid */
|
|
1250 line = ml_get(lnum);
|
|
1251 p = line + col;
|
|
1252 }
|
|
1253 else
|
|
1254 can_spell = TRUE;
|
|
1255
|
|
1256 if (can_spell)
|
|
1257 {
|
|
1258 found_pos.lnum = lnum;
|
|
1259 found_pos.col = p - line;
|
|
1260 #ifdef FEAT_VIRTUALEDIT
|
|
1261 found_pos.coladd = 0;
|
|
1262 #endif
|
|
1263 if (dir == FORWARD)
|
|
1264 {
|
|
1265 /* No need to search further. */
|
|
1266 curwin->w_cursor = found_pos;
|
|
1267 return OK;
|
|
1268 }
|
|
1269 }
|
|
1270 }
|
236
|
1271 }
|
249
|
1272 attr = 0;
|
236
|
1273 }
|
|
1274 p += len;
|
|
1275 if (*p == NUL)
|
|
1276 break;
|
|
1277 nwc = FALSE;
|
|
1278 }
|
|
1279
|
|
1280 /* advance to next character */
|
|
1281 mb_ptr_adv(p);
|
|
1282 wc = nwc;
|
|
1283 }
|
|
1284
|
|
1285 /* Advance to next line. */
|
249
|
1286 if (dir == BACKWARD)
|
|
1287 {
|
|
1288 if (found_pos.lnum != 0)
|
|
1289 {
|
|
1290 /* Use the last match in the line. */
|
|
1291 curwin->w_cursor = found_pos;
|
|
1292 return OK;
|
|
1293 }
|
|
1294 if (lnum == 1)
|
|
1295 return FAIL;
|
|
1296 --lnum;
|
|
1297 }
|
|
1298 else
|
|
1299 {
|
|
1300 if (lnum == curbuf->b_ml.ml_line_count)
|
|
1301 return FAIL;
|
|
1302 ++lnum;
|
|
1303 }
|
236
|
1304
|
|
1305 line_breakcheck();
|
|
1306 }
|
|
1307
|
|
1308 return FAIL; /* interrupted */
|
|
1309 }
|
|
1310
|
|
1311 /*
|
|
1312 * Load word list for "lang" from a Vim spell file.
|
|
1313 * "lang" must be the language without the region: "en" or "en-rare".
|
|
1314 */
|
|
1315 static slang_T *
|
|
1316 spell_load_lang(lang)
|
|
1317 char_u *lang;
|
|
1318 {
|
|
1319 slang_T *lp;
|
|
1320 char_u fname_enc[80];
|
|
1321 char_u *p;
|
|
1322 int r;
|
|
1323
|
|
1324 lp = slang_alloc(lang);
|
|
1325 if (lp != NULL)
|
|
1326 {
|
|
1327 /* Find all spell files for "lang" in 'runtimepath' and load them.
|
|
1328 * Use 'encoding', except that we use "latin1" for "latin9". */
|
|
1329 #ifdef FEAT_MBYTE
|
|
1330 if (STRLEN(p_enc) < 60 && STRCMP(p_enc, "iso-8859-15") != 0)
|
|
1331 p = p_enc;
|
|
1332 else
|
|
1333 #endif
|
|
1334 p = (char_u *)"latin1";
|
272
|
1335 vim_snprintf((char *)fname_enc, sizeof(fname_enc),
|
|
1336 "spell/%s.%s.spl", lang, p);
|
236
|
1337
|
|
1338 r = do_in_runtimepath(fname_enc, TRUE, spell_load_file, lp);
|
240
|
1339 if (r == FAIL && !lp->sl_error)
|
|
1340 {
|
|
1341 /* Try loading the ASCII version. */
|
272
|
1342 vim_snprintf((char *)fname_enc, sizeof(fname_enc),
|
|
1343 "spell/%s.ascii.spl", lang);
|
240
|
1344
|
|
1345 r = do_in_runtimepath(fname_enc, TRUE, spell_load_file, lp);
|
|
1346 }
|
236
|
1347 if (r == FAIL || lp->sl_error)
|
|
1348 {
|
|
1349 slang_free(lp);
|
|
1350 lp = NULL;
|
|
1351 if (r == FAIL)
|
|
1352 smsg((char_u *)_("Warning: Cannot find word list \"%s\""),
|
|
1353 fname_enc + 6);
|
|
1354 }
|
|
1355 else
|
|
1356 {
|
|
1357 lp->sl_next = first_lang;
|
|
1358 first_lang = lp;
|
|
1359 }
|
|
1360 }
|
|
1361
|
|
1362 return lp;
|
|
1363 }
|
|
1364
|
|
1365 /*
|
|
1366 * Allocate a new slang_T.
|
|
1367 * Caller must fill "sl_next".
|
|
1368 */
|
|
1369 static slang_T *
|
|
1370 slang_alloc(lang)
|
|
1371 char_u *lang;
|
|
1372 {
|
|
1373 slang_T *lp;
|
|
1374
|
|
1375 lp = (slang_T *)alloc(sizeof(slang_T));
|
|
1376 if (lp != NULL)
|
|
1377 {
|
|
1378 lp->sl_name = vim_strsave(lang);
|
|
1379 hash_init(&lp->sl_words);
|
|
1380 ga_init2(&lp->sl_preftab, sizeof(hashtab_T), 4);
|
243
|
1381 hash_init(&lp->sl_prewords);
|
236
|
1382 ga_init2(&lp->sl_sufftab, sizeof(hashtab_T), 4);
|
|
1383 lp->sl_prefzero = NULL;
|
|
1384 lp->sl_suffzero = NULL;
|
|
1385 lp->sl_try = NULL;
|
|
1386 ga_init2(&lp->sl_rep, sizeof(repentry_T), 4);
|
|
1387 lp->sl_regions[0] = NUL;
|
|
1388 lp->sl_block = NULL;
|
|
1389 lp->sl_error = FALSE;
|
|
1390 }
|
|
1391 return lp;
|
|
1392 }
|
|
1393
|
|
1394 /*
|
|
1395 * Free the contents of an slang_T and the structure itself.
|
|
1396 */
|
|
1397 static void
|
|
1398 slang_free(lp)
|
|
1399 slang_T *lp;
|
|
1400 {
|
|
1401 sblock_T *sp;
|
|
1402 int i;
|
243
|
1403 fword_T *fw;
|
|
1404 int todo;
|
|
1405 hashitem_T *hi;
|
236
|
1406
|
|
1407 vim_free(lp->sl_name);
|
243
|
1408
|
|
1409 /* The words themselves are in memory blocks referenced by "sl_block".
|
|
1410 * Only the hashtables for additions need to be cleared. */
|
|
1411 todo = lp->sl_words.ht_used;
|
|
1412 for (hi = lp->sl_words.ht_array; todo > 0; ++hi)
|
|
1413 {
|
|
1414 if (!HASHITEM_EMPTY(hi))
|
|
1415 {
|
|
1416 --todo;
|
|
1417 fw = HI2FWORD(hi);
|
|
1418 if (fw->fw_flags & BWF_ADDHASH)
|
|
1419 hash_clear((hashtab_T *)fw->fw_adds);
|
|
1420 }
|
|
1421 }
|
236
|
1422 hash_clear(&lp->sl_words);
|
243
|
1423
|
236
|
1424 for (i = 0; i < lp->sl_preftab.ga_len; ++i)
|
|
1425 hash_clear(((hashtab_T *)lp->sl_preftab.ga_data) + i);
|
|
1426 ga_clear(&lp->sl_preftab);
|
243
|
1427 hash_clear(&lp->sl_prewords);
|
236
|
1428 for (i = 0; i < lp->sl_sufftab.ga_len; ++i)
|
|
1429 hash_clear(((hashtab_T *)lp->sl_sufftab.ga_data) + i);
|
|
1430 ga_clear(&lp->sl_sufftab);
|
|
1431 ga_clear(&lp->sl_rep);
|
|
1432 vim_free(lp->sl_try);
|
|
1433 while (lp->sl_block != NULL)
|
|
1434 {
|
|
1435 sp = lp->sl_block;
|
|
1436 lp->sl_block = sp->sb_next;
|
|
1437 vim_free(sp);
|
|
1438 }
|
|
1439 vim_free(lp);
|
|
1440 }
|
|
1441
|
|
1442 /*
|
|
1443 * Load one spell file into an slang_T.
|
|
1444 * Invoked through do_in_runtimepath().
|
|
1445 */
|
|
1446 static void
|
|
1447 spell_load_file(fname, cookie)
|
|
1448 char_u *fname;
|
|
1449 void *cookie; /* points to the slang_T to be filled */
|
|
1450 {
|
|
1451 slang_T *lp = cookie;
|
|
1452 FILE *fd;
|
|
1453 char_u buf[MAXWLEN + 1];
|
|
1454 char_u cbuf[MAXWLEN + 1];
|
|
1455 char_u fbuf[MAXWLEN + 1];
|
243
|
1456 char_u affixbuf[256 * 2 * 2]; /* max 2 * 256 affix nrs of 2 bytes */
|
236
|
1457 char_u *p;
|
|
1458 int itm;
|
|
1459 int i;
|
|
1460 int affcount;
|
|
1461 int affnr;
|
|
1462 int affflags;
|
|
1463 int affitemcnt;
|
243
|
1464 int prefixcnt, suffixcnt;
|
236
|
1465 int bl_used = SBLOCKSIZE;
|
|
1466 int widx;
|
240
|
1467 int prefm = 0; /* 1 if <= 256 prefixes, sizeof(short_u) otherw. */
|
|
1468 int suffm = 0; /* 1 if <= 256 suffixes, sizeof(short_u) otherw. */
|
236
|
1469 int wlen;
|
|
1470 int flags;
|
|
1471 affitem_T *ai, *ai2, **aip;
|
|
1472 int round;
|
|
1473 char_u *save_sourcing_name = sourcing_name;
|
|
1474 linenr_T save_sourcing_lnum = sourcing_lnum;
|
255
|
1475 int cnt, ccnt;
|
236
|
1476 int choplen;
|
|
1477 int addlen;
|
|
1478 int leadlen;
|
|
1479 int wordcount;
|
|
1480 fword_T *fw, *fw2;
|
|
1481 garray_T *gap;
|
|
1482 hashtab_T *ht;
|
|
1483 hashitem_T *hi;
|
|
1484 hash_T hash;
|
|
1485 int adds;
|
243
|
1486 addword_T *aw, *naw;
|
236
|
1487 int flen;
|
243
|
1488 int xlen;
|
255
|
1489 char_u *fol;
|
236
|
1490
|
|
1491 fd = fopen((char *)fname, "r");
|
|
1492 if (fd == NULL)
|
|
1493 {
|
|
1494 EMSG2(_(e_notopen), fname);
|
255
|
1495 goto endFAIL;
|
236
|
1496 }
|
|
1497
|
|
1498 /* Set sourcing_name, so that error messages mention the file name. */
|
|
1499 sourcing_name = fname;
|
|
1500 sourcing_lnum = 0;
|
|
1501
|
255
|
1502 /* <HEADER>: <fileID> <regioncnt> <regionname> ...
|
|
1503 * <charflagslen> <charflags> <fcharslen> <fchars> */
|
236
|
1504 for (i = 0; i < VIMSPELLMAGICL; ++i)
|
|
1505 buf[i] = getc(fd); /* <fileID> */
|
|
1506 if (STRNCMP(buf, VIMSPELLMAGIC, VIMSPELLMAGICL) != 0)
|
|
1507 {
|
|
1508 EMSG(_("E757: Wrong file ID in spell file"));
|
255
|
1509 goto endFAIL;
|
236
|
1510 }
|
|
1511
|
|
1512 cnt = getc(fd); /* <regioncnt> */
|
255
|
1513 if (cnt < 0)
|
236
|
1514 {
|
|
1515 truncerr:
|
|
1516 EMSG(_("E758: Truncated spell file"));
|
255
|
1517 goto endFAIL;
|
236
|
1518 }
|
|
1519 if (cnt > 8)
|
|
1520 {
|
|
1521 formerr:
|
|
1522 EMSG(_("E759: Format error in spell file"));
|
255
|
1523 goto endFAIL;
|
236
|
1524 }
|
|
1525 for (i = 0; i < cnt; ++i)
|
|
1526 {
|
|
1527 lp->sl_regions[i * 2] = getc(fd); /* <regionname> */
|
|
1528 lp->sl_regions[i * 2 + 1] = getc(fd);
|
|
1529 }
|
|
1530 lp->sl_regions[cnt * 2] = NUL;
|
|
1531
|
255
|
1532 cnt = getc(fd); /* <charflagslen> */
|
|
1533 if (cnt > 0)
|
|
1534 {
|
|
1535 p = (char_u *)getroom(lp, &bl_used, cnt);
|
|
1536 if (p == NULL)
|
|
1537 goto endFAIL;
|
|
1538 for (i = 0; i < cnt; ++i)
|
|
1539 p[i] = getc(fd); /* <charflags> */
|
|
1540
|
|
1541 ccnt = (getc(fd) << 8) + getc(fd); /* <fcharslen> */
|
|
1542 if (ccnt <= 0)
|
|
1543 goto formerr;
|
|
1544 fol = (char_u *)getroom(lp, &bl_used, ccnt + 1);
|
|
1545 if (fol == NULL)
|
|
1546 goto endFAIL;
|
|
1547 for (i = 0; i < ccnt; ++i)
|
|
1548 fol[i] = getc(fd); /* <fchars> */
|
|
1549 fol[i] = NUL;
|
|
1550
|
|
1551 /* Set the word-char flags and fill spell_isupper() table. */
|
|
1552 if (set_spell_charflags(p, cnt, fol) == FAIL)
|
|
1553 goto formerr;
|
|
1554 }
|
|
1555 else
|
|
1556 {
|
|
1557 /* When <charflagslen> is zero then <fcharlen> must also be zero. */
|
|
1558 cnt = (getc(fd) << 8) + getc(fd);
|
|
1559 if (cnt != 0)
|
|
1560 goto formerr;
|
|
1561 }
|
|
1562
|
|
1563 /* round 1: <PREFIXLIST>: <affcount> <affix> ...
|
|
1564 * round 2: <SUFFIXLIST>: <affcount> <affix> ... */
|
236
|
1565 for (round = 1; round <= 2; ++round)
|
|
1566 {
|
|
1567 affcount = (getc(fd) << 8) + getc(fd); /* <affcount> */
|
|
1568 if (affcount < 0)
|
|
1569 goto truncerr;
|
|
1570 if (round == 1)
|
|
1571 {
|
|
1572 gap = &lp->sl_preftab;
|
|
1573 aip = &lp->sl_prefzero;
|
|
1574 lp->sl_prefcnt = affcount;
|
243
|
1575 prefm = affcount > 256 ? 2 : 1;
|
236
|
1576 }
|
|
1577 else
|
|
1578 {
|
|
1579 gap = &lp->sl_sufftab;
|
|
1580 aip = &lp->sl_suffzero;
|
|
1581 lp->sl_suffcnt = affcount;
|
243
|
1582 suffm = affcount > 256 ? 2 : 1;
|
236
|
1583 }
|
|
1584
|
|
1585 /*
|
|
1586 * For each affix NR there can be several affixes.
|
|
1587 */
|
|
1588 for (affnr = 0; affnr < affcount; ++affnr)
|
|
1589 {
|
243
|
1590 /* <affix>: <affitemcnt> <affitem> ... */
|
236
|
1591 affitemcnt = (getc(fd) << 8) + getc(fd); /* <affitemcnt> */
|
|
1592 if (affitemcnt < 0)
|
|
1593 goto truncerr;
|
|
1594 for (itm = 0; itm < affitemcnt; ++itm)
|
|
1595 {
|
243
|
1596 /* <affitem>: <affflags> <affchoplen> <affchop>
|
|
1597 * <affaddlen> <affadd> */
|
|
1598 affflags = getc(fd); /* <affflags> */
|
236
|
1599 choplen = getc(fd); /* <affchoplen> */
|
255
|
1600 if (choplen < 0)
|
236
|
1601 goto truncerr;
|
|
1602 if (choplen >= MAXWLEN)
|
|
1603 goto formerr;
|
|
1604 for (i = 0; i < choplen; ++i) /* <affchop> */
|
|
1605 buf[i] = getc(fd);
|
|
1606 buf[i] = NUL;
|
|
1607 addlen = getc(fd); /* <affaddlen> */
|
255
|
1608 if (addlen < 0)
|
236
|
1609 goto truncerr;
|
243
|
1610 if (affflags & AFF_PREWORD)
|
|
1611 xlen = addlen + 2; /* space for lead and trail string */
|
|
1612 else
|
|
1613 xlen = 0;
|
|
1614
|
236
|
1615 /* Get room to store the affitem_T, chop and add strings. */
|
255
|
1616 ai = (affitem_T *)getroom(lp, &bl_used,
|
243
|
1617 sizeof(affitem_T) + addlen + choplen + 1 + xlen);
|
255
|
1618 if (ai == NULL)
|
|
1619 goto endFAIL;
|
|
1620
|
236
|
1621 ai->ai_nr = affnr;
|
243
|
1622 ai->ai_flags = affflags;
|
236
|
1623 ai->ai_choplen = choplen;
|
|
1624 ai->ai_addlen = addlen;
|
|
1625
|
243
|
1626 /* Chop string is at ai_add[ai_addlen + 1]. */
|
|
1627 p = ai->ai_add + addlen + 1;
|
236
|
1628 STRCPY(p, buf);
|
|
1629
|
|
1630 p = ai->ai_add;
|
|
1631 for (i = 0; i < addlen; ++i) /* <affadd> */
|
|
1632 p[i] = getc(fd);
|
|
1633 p[i] = NUL;
|
|
1634
|
243
|
1635 if (affflags & AFF_PREWORD)
|
236
|
1636 {
|
243
|
1637 int l, leadoff, trailoff;
|
|
1638
|
|
1639 /*
|
255
|
1640 * A preword is a prefix that's recognized as a word: it
|
|
1641 * contains a word characters folled by a non-word
|
|
1642 * character.
|
|
1643 * <affadd> is the whole prefix. Separate lead and trail
|
|
1644 * string, put the word itself at ai_add, so that it can
|
|
1645 * be used as hashtable key.
|
243
|
1646 */
|
|
1647 /* lead string: up to first word char */
|
|
1648 while (*p != NUL && !spell_iswordc(p))
|
|
1649 mb_ptr_adv(p);
|
|
1650 ai->ai_leadlen = p - ai->ai_add;
|
|
1651 leadoff = addlen + choplen + 2;
|
|
1652 mch_memmove(ai->ai_add + leadoff, ai->ai_add,
|
|
1653 ai->ai_leadlen);
|
|
1654 ai->ai_add[leadoff + ai->ai_leadlen] = NUL;
|
|
1655
|
|
1656 /* trail string: after last word char */
|
|
1657 while (*p != NUL && spell_iswordc(p))
|
|
1658 mb_ptr_adv(p);
|
|
1659 trailoff = leadoff + ai->ai_leadlen + 1;
|
|
1660 STRCPY(ai->ai_add + trailoff, p);
|
|
1661 ai->ai_taillen = STRLEN(p);
|
|
1662
|
|
1663 /* word itself */
|
|
1664 l = (p - ai->ai_add) - ai->ai_leadlen;
|
|
1665 mch_memmove(ai->ai_add, ai->ai_add + ai->ai_leadlen, l);
|
|
1666 ai->ai_add[l] = NUL;
|
|
1667 hash = hash_hash(ai->ai_add);
|
|
1668 hi = hash_lookup(&lp->sl_prewords, ai->ai_add, hash);
|
236
|
1669 if (HASHITEM_EMPTY(hi))
|
|
1670 {
|
255
|
1671 /* First preword with this word, add to hashtable. */
|
243
|
1672 hash_add_item(&lp->sl_prewords, hi, ai->ai_add, hash);
|
236
|
1673 ai->ai_next = NULL;
|
|
1674 }
|
|
1675 else
|
|
1676 {
|
255
|
1677 /* There already is a preword with this word, link in
|
243
|
1678 * the list. */
|
236
|
1679 ai2 = HI2AI(hi);
|
|
1680 ai->ai_next = ai2->ai_next;
|
|
1681 ai2->ai_next = ai;
|
|
1682 }
|
|
1683 }
|
243
|
1684 else
|
|
1685 {
|
|
1686 /*
|
|
1687 * Add the affix to a hashtable. Which one depends on the
|
|
1688 * length of the added string in characters.
|
|
1689 */
|
|
1690 #ifdef FEAT_MBYTE
|
|
1691 /* Change "addlen" from length in bytes to length in
|
|
1692 * chars. */
|
|
1693 if (has_mbyte)
|
|
1694 addlen = mb_charlen(p);
|
|
1695 #endif
|
|
1696 if (addlen == 0)
|
|
1697 {
|
|
1698 /* Link in list of zero length affixes. */
|
|
1699 ai->ai_next = *aip;
|
|
1700 *aip = ai;
|
|
1701 }
|
|
1702 else
|
|
1703 {
|
|
1704 if (gap->ga_len < addlen)
|
|
1705 {
|
|
1706 /* Longer affix, need more hashtables. */
|
|
1707 if (ga_grow(gap, addlen - gap->ga_len) == FAIL)
|
255
|
1708 goto endFAIL;
|
243
|
1709
|
|
1710 /* Re-allocating ga_data means that an ht_array
|
|
1711 * pointing to ht_smallarray becomes invalid. We
|
|
1712 * can recognize this: ht_mask is at its init
|
|
1713 * value. */
|
|
1714 for (i = 0; i < gap->ga_len; ++i)
|
|
1715 {
|
|
1716 ht = ((hashtab_T *)gap->ga_data) + i;
|
|
1717 if (ht->ht_mask == HT_INIT_SIZE - 1)
|
|
1718 ht->ht_array = ht->ht_smallarray;
|
|
1719 }
|
|
1720
|
|
1721 /* Init the newly used hashtable(s). */
|
|
1722 while (gap->ga_len < addlen)
|
|
1723 {
|
|
1724 hash_init(((hashtab_T *)gap->ga_data)
|
|
1725 + gap->ga_len);
|
|
1726 ++gap->ga_len;
|
|
1727 }
|
|
1728 }
|
|
1729 ht = ((hashtab_T *)gap->ga_data) + addlen - 1;
|
|
1730 hash = hash_hash(p);
|
|
1731 hi = hash_lookup(ht, p, hash);
|
|
1732 if (HASHITEM_EMPTY(hi))
|
|
1733 {
|
|
1734 /* First affix with this "ai_add", add to
|
|
1735 * hashtable. */
|
|
1736 hash_add_item(ht, hi, p, hash);
|
|
1737 ai->ai_next = NULL;
|
|
1738 }
|
|
1739 else
|
|
1740 {
|
|
1741 /* There already is an affix with this "ai_add",
|
|
1742 * link in the list. */
|
|
1743 ai2 = HI2AI(hi);
|
|
1744 ai->ai_next = ai2->ai_next;
|
|
1745 ai2->ai_next = ai;
|
|
1746 }
|
|
1747 }
|
|
1748 }
|
236
|
1749 }
|
|
1750 }
|
|
1751 }
|
|
1752
|
|
1753 /* <SUGGEST> : <suggestlen> <more> ... */
|
|
1754 /* TODO, just skip this for now */
|
|
1755 i = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8) + getc(fd);
|
|
1756 while (i-- > 0)
|
|
1757 if (getc(fd) == EOF) /* <suggestlen> */
|
|
1758 goto truncerr;
|
|
1759
|
|
1760 /* <WORDLIST>: <wordcount> <worditem> ... */ /* <wordcount> */
|
|
1761 wordcount = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8)
|
|
1762 + getc(fd);
|
|
1763 if (wordcount < 0)
|
|
1764 goto truncerr;
|
|
1765
|
|
1766 /* Init hashtable for this number of words, so that it doesn't need to
|
|
1767 * reallocate the table halfway. */
|
|
1768 hash_lock_size(&lp->sl_words, wordcount);
|
|
1769
|
|
1770 for (widx = 0; ; ++widx)
|
|
1771 {
|
|
1772 /* <worditem>: <nr> <string> <flags> [<flags2>]
|
|
1773 * [<caselen> <caseword>]
|
|
1774 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
1775 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
1776 * [<region>]
|
|
1777 * [<addcnt> <add> ...]
|
|
1778 */
|
|
1779 /* Use <nr> bytes from the previous word. */
|
|
1780 wlen = getc(fd); /* <nr> */
|
255
|
1781 if (wlen < 0)
|
236
|
1782 {
|
|
1783 if (widx >= wordcount) /* normal way to end the file */
|
|
1784 break;
|
|
1785 goto truncerr;
|
|
1786 }
|
|
1787
|
255
|
1788 /* Read further word bytes until one below 0x20, that one must be the
|
236
|
1789 * flags. Keep this fast! */
|
|
1790 for (;;)
|
|
1791 {
|
|
1792 if ((buf[wlen] = getc(fd)) < 0x20) /* <string> */
|
|
1793 break;
|
|
1794 if (++wlen == MAXWLEN)
|
|
1795 goto formerr;
|
|
1796 }
|
|
1797 flags = buf[wlen]; /* <flags> */
|
|
1798 buf[wlen] = NUL;
|
|
1799
|
|
1800 /* Get more flags if they're there. */
|
|
1801 if (flags & BWF_SECOND)
|
|
1802 flags += getc(fd) << 8; /* <flags2> */
|
|
1803
|
|
1804 if (flags & BWF_KEEPCAP)
|
|
1805 {
|
|
1806 /* Read <caselen> and <caseword> first, its length may differ from
|
|
1807 * the case-folded word. Note: this should only happen after the
|
255
|
1808 * basic word without KEEPCAP! */
|
236
|
1809 wlen = getc(fd);
|
243
|
1810 if (wlen < 0)
|
236
|
1811 goto truncerr;
|
255
|
1812 if (wlen >= MAXWLEN)
|
|
1813 goto formerr;
|
236
|
1814 for (i = 0; i < wlen; ++i)
|
|
1815 cbuf[i] = getc(fd);
|
|
1816 cbuf[i] = NUL;
|
|
1817 }
|
|
1818
|
243
|
1819 /* Optional prefixes */
|
|
1820 p = affixbuf;
|
|
1821 if (flags & BWF_PREFIX)
|
|
1822 {
|
|
1823 cnt = getc(fd); /* <affixcnt> */
|
|
1824 if (cnt < 0)
|
|
1825 goto truncerr;
|
|
1826 prefixcnt = cnt;
|
|
1827 for (i = cnt * prefm; --i >= 0; ) /* <affixNR> */
|
|
1828 *p++ = getc(fd);
|
|
1829 }
|
|
1830 else
|
|
1831 prefixcnt = 0;
|
|
1832
|
|
1833 /* Optional suffixes */
|
|
1834 if (flags & BWF_SUFFIX)
|
|
1835 {
|
|
1836 cnt = getc(fd); /* <affixcnt> */
|
|
1837 if (cnt < 0)
|
|
1838 goto truncerr;
|
|
1839 suffixcnt = cnt;
|
|
1840 for (i = cnt * suffm; --i >= 0; ) /* <affixNR> */
|
|
1841 *p++ = getc(fd);
|
|
1842 }
|
|
1843 else
|
|
1844 suffixcnt = 0;
|
|
1845
|
|
1846 /* Find room to store the word in an fword_T. */
|
|
1847 fw = (fword_T *)getroom(lp, &bl_used, (int)sizeof(fword_T) + wlen
|
|
1848 + (p - affixbuf));
|
236
|
1849 if (fw == NULL)
|
255
|
1850 goto endFAIL;
|
236
|
1851 mch_memmove(fw->fw_word, (flags & BWF_KEEPCAP) ? cbuf : buf, wlen + 1);
|
243
|
1852
|
|
1853 /* Put the affix NRs just after the word, if any. */
|
|
1854 if (p > affixbuf)
|
|
1855 mch_memmove(fw->fw_word + wlen + 1, affixbuf, p - affixbuf);
|
|
1856
|
236
|
1857 fw->fw_flags = flags;
|
243
|
1858 fw->fw_prefixcnt = prefixcnt;
|
|
1859 fw->fw_suffixcnt = suffixcnt;
|
236
|
1860
|
255
|
1861 /* We store the word in the hashtable case-folded. For a KEEPCAP word
|
|
1862 * the entry must already exist, because fw_word can't be used as the
|
|
1863 * key, it differs from "buf"! */
|
236
|
1864 hash = hash_hash(buf);
|
|
1865 hi = hash_lookup(&lp->sl_words, buf, hash);
|
|
1866 if (HASHITEM_EMPTY(hi))
|
|
1867 {
|
|
1868 if (hash_add_item(&lp->sl_words, hi, fw->fw_word, hash) == FAIL)
|
255
|
1869 goto endFAIL;
|
236
|
1870 fw->fw_next = NULL;
|
|
1871 }
|
|
1872 else
|
|
1873 {
|
|
1874 /* Already have this basic word in the hashtable, this one will
|
243
|
1875 * have different case flags and/or affixes. */
|
236
|
1876 fw2 = HI2FWORD(hi);
|
|
1877 fw->fw_next = fw2->fw_next;
|
|
1878 fw2->fw_next = fw;
|
255
|
1879 --widx; /* don't count this one as a basic word */
|
236
|
1880 }
|
|
1881
|
|
1882 if (flags & BWF_REGION)
|
|
1883 fw->fw_region = getc(fd); /* <region> */
|
|
1884 else
|
|
1885 fw->fw_region = REGION_ALL;
|
|
1886
|
|
1887 fw->fw_adds = NULL;
|
|
1888 if (flags & BWF_ADDS)
|
|
1889 {
|
249
|
1890 if (flags & BWF_ADDS_M)
|
|
1891 adds = (getc(fd) << 8) + getc(fd); /* <addcnt> */
|
|
1892 else
|
|
1893 adds = getc(fd); /* <addcnt> */
|
255
|
1894 if (adds < 0)
|
|
1895 goto formerr;
|
236
|
1896
|
243
|
1897 if (adds > 30)
|
|
1898 {
|
255
|
1899 /* Use a hashtable to lookup the part until the next word end.
|
|
1900 * Thus for "de-bur-die" "de" is the basic word, "-bur" is key
|
|
1901 * in the addition hashtable, "-bur<NUL>die" the whole
|
|
1902 * addition and "aw_saveb" is '-'.
|
243
|
1903 * This uses more memory and involves some overhead, thus only
|
255
|
1904 * do it when there are many additions (e.g., for French). */
|
243
|
1905 ht = (hashtab_T *)getroom(lp, &bl_used, sizeof(hashtab_T));
|
|
1906 if (ht == NULL)
|
255
|
1907 goto endFAIL;
|
243
|
1908 hash_init(ht);
|
|
1909 fw->fw_adds = (addword_T *)ht;
|
|
1910 fw->fw_flags |= BWF_ADDHASH;
|
|
1911
|
|
1912 /* Preset the size of the hashtable. It's never unlocked. */
|
|
1913 hash_lock_size(ht, adds + 1);
|
|
1914 }
|
|
1915 else
|
|
1916 ht = NULL;
|
|
1917
|
255
|
1918 /*
|
|
1919 * Note: uses cbuf[] to copy bytes from previous addition.
|
|
1920 */
|
236
|
1921 while (--adds >= 0)
|
|
1922 {
|
249
|
1923 /* <add>: <addflags> <addlen> [<leadlen>] [<copylen>]
|
|
1924 * [<addstring>] [<region>] */
|
236
|
1925 flags = getc(fd); /* <addflags> */
|
|
1926 addlen = getc(fd); /* <addlen> */
|
255
|
1927 if (addlen < 0)
|
236
|
1928 goto truncerr;
|
|
1929 if (addlen >= MAXWLEN)
|
|
1930 goto formerr;
|
|
1931
|
249
|
1932 if (flags & ADD_LEADLEN)
|
255
|
1933 {
|
249
|
1934 leadlen = getc(fd); /* <leadlen> */
|
255
|
1935 if (leadlen > addlen)
|
|
1936 goto formerr;
|
|
1937 }
|
249
|
1938 else
|
|
1939 leadlen = 0;
|
|
1940
|
236
|
1941 if (addlen > 0)
|
|
1942 {
|
249
|
1943 if (flags & ADD_COPYLEN)
|
|
1944 i = getc(fd); /* <copylen> */
|
|
1945 else
|
|
1946 i = 0;
|
|
1947 for ( ; i < addlen; ++i) /* <addstring> */
|
236
|
1948 cbuf[i] = getc(fd);
|
|
1949 cbuf[i] = NUL;
|
|
1950 }
|
|
1951
|
|
1952 if (flags & ADD_KEEPCAP)
|
|
1953 {
|
|
1954 /* <addstring> is in original case, need to get
|
|
1955 * case-folded word too. */
|
255
|
1956 (void)spell_casefold(cbuf, addlen, fbuf, MAXWLEN);
|
236
|
1957 flen = addlen - leadlen + 1;
|
|
1958 addlen = STRLEN(fbuf);
|
|
1959 }
|
|
1960 else
|
|
1961 flen = 0;
|
|
1962
|
|
1963 aw = (addword_T *)getroom(lp, &bl_used,
|
|
1964 sizeof(addword_T) + addlen + flen);
|
|
1965 if (aw == NULL)
|
255
|
1966 goto endFAIL;
|
236
|
1967
|
|
1968 if (flags & ADD_KEEPCAP)
|
|
1969 {
|
|
1970 /* Put the addition in original case after the case-folded
|
|
1971 * string. */
|
|
1972 STRCPY(aw->aw_word, fbuf);
|
|
1973 STRCPY(aw->aw_word + addlen + 1, cbuf + leadlen);
|
|
1974 }
|
|
1975 else
|
|
1976 STRCPY(aw->aw_word, cbuf);
|
|
1977
|
|
1978 aw->aw_flags = flags;
|
|
1979 aw->aw_wordlen = addlen;
|
243
|
1980 aw->aw_leadlen = leadlen;
|
236
|
1981
|
|
1982 if (flags & ADD_REGION)
|
|
1983 aw->aw_region = getc(fd); /* <region> */
|
|
1984 else
|
|
1985 aw->aw_region = REGION_ALL;
|
243
|
1986
|
|
1987 if (ht == NULL)
|
|
1988 {
|
|
1989 /* Using simple linked list, put it in front. */
|
|
1990 aw->aw_next = fw->fw_adds;
|
|
1991 fw->fw_adds = aw;
|
|
1992 aw->aw_saveb = NUL;
|
|
1993 }
|
|
1994 else
|
|
1995 {
|
|
1996 /* Put addition in hashtable. For key we use the part up
|
|
1997 * to the next end-of-word. */
|
|
1998 if (leadlen == 0)
|
|
1999 {
|
|
2000 p = aw->aw_word;
|
|
2001 while (*p != NUL && !spell_iswordc(p))
|
|
2002 mb_ptr_adv(p);
|
|
2003 }
|
|
2004
|
|
2005 if (leadlen != 0 || *p == NUL)
|
|
2006 {
|
|
2007 /* Only non-word characters in addition, add it to the
|
|
2008 * list with the special key NOWC_KEY. Also do this
|
|
2009 * when there is a leadstring, it would get too
|
|
2010 * complicated. */
|
|
2011 hash = hash_hash(NOWC_KEY);
|
|
2012 hi = hash_lookup(ht, NOWC_KEY, hash);
|
|
2013 if (HASHITEM_EMPTY(hi))
|
|
2014 {
|
|
2015 /* we use a dummy item as the list header */
|
|
2016 naw = (addword_T *)getroom(lp, &bl_used,
|
|
2017 sizeof(addword_T) + STRLEN(NOWC_KEY));
|
|
2018 if (naw == NULL)
|
255
|
2019 goto endFAIL;
|
243
|
2020 STRCPY(naw->aw_word, NOWC_KEY);
|
|
2021 hash_add_item(ht, hi, naw->aw_word, hash);
|
|
2022 naw->aw_next = aw;
|
|
2023 aw->aw_next = NULL;
|
|
2024 }
|
|
2025 else
|
|
2026 {
|
|
2027 naw = HI2ADDWORD(hi);
|
|
2028 aw->aw_next = naw->aw_next;
|
|
2029 naw->aw_next = aw;
|
|
2030 }
|
|
2031 aw->aw_saveb = NUL;
|
|
2032 }
|
|
2033 else
|
|
2034 {
|
|
2035 /* Truncate at next non-word character, store that
|
|
2036 * byte in "aw_saveb". */
|
|
2037 while (*p != NUL && spell_iswordc(p))
|
|
2038 mb_ptr_adv(p);
|
|
2039 aw->aw_saveb = *p;
|
|
2040 *p = NUL;
|
|
2041 hash = hash_hash(aw->aw_word);
|
|
2042 hi = hash_lookup(ht, aw->aw_word, hash);
|
|
2043 if (HASHITEM_EMPTY(hi))
|
|
2044 {
|
|
2045 hash_add_item(ht, hi, aw->aw_word, hash);
|
|
2046 aw->aw_next = NULL;
|
|
2047 }
|
|
2048 else
|
|
2049 {
|
|
2050 naw = HI2ADDWORD(hi);
|
|
2051 aw->aw_next = naw->aw_next;
|
|
2052 naw->aw_next = aw;
|
|
2053 }
|
|
2054 }
|
|
2055 }
|
236
|
2056 }
|
|
2057 }
|
|
2058 }
|
255
|
2059 goto endOK;
|
|
2060
|
|
2061 endFAIL:
|
236
|
2062 lp->sl_error = TRUE;
|
255
|
2063
|
|
2064 endOK:
|
236
|
2065 if (fd != NULL)
|
|
2066 fclose(fd);
|
|
2067 hash_unlock(&lp->sl_words);
|
|
2068 sourcing_name = save_sourcing_name;
|
|
2069 sourcing_lnum = save_sourcing_lnum;
|
|
2070 }
|
|
2071
|
|
2072 /*
|
|
2073 * Get part of an sblock_T, at least "len" bytes long.
|
|
2074 * Returns NULL when out of memory.
|
|
2075 */
|
|
2076 static void *
|
|
2077 getroom(lp, bl_used, len)
|
|
2078 slang_T *lp; /* lp->sl_block is current block or NULL */
|
|
2079 int *bl_used; /* used up from current block */
|
|
2080 int len; /* length needed */
|
|
2081 {
|
|
2082 char_u *p;
|
|
2083 sblock_T *bl = lp->sl_block;
|
|
2084
|
|
2085 if (bl == NULL || *bl_used + len > SBLOCKSIZE)
|
|
2086 {
|
|
2087 /* Allocate a block of memory. This is not freed until spell_reload()
|
|
2088 * is called. */
|
|
2089 bl = (sblock_T *)alloc((unsigned)(sizeof(sblock_T) + SBLOCKSIZE));
|
|
2090 if (bl == NULL)
|
|
2091 return NULL;
|
|
2092 bl->sb_next = lp->sl_block;
|
|
2093 lp->sl_block = bl;
|
|
2094 *bl_used = 0;
|
|
2095 }
|
|
2096
|
|
2097 p = bl->sb_data + *bl_used;
|
|
2098 *bl_used += len;
|
|
2099
|
|
2100 return p;
|
|
2101 }
|
|
2102
|
|
2103 /*
|
|
2104 * Parse 'spelllang' and set buf->b_langp accordingly.
|
|
2105 * Returns an error message or NULL.
|
|
2106 */
|
|
2107 char_u *
|
|
2108 did_set_spelllang(buf)
|
|
2109 buf_T *buf;
|
|
2110 {
|
|
2111 garray_T ga;
|
|
2112 char_u *lang;
|
|
2113 char_u *e;
|
|
2114 char_u *region;
|
|
2115 int region_mask;
|
|
2116 slang_T *lp;
|
|
2117 int c;
|
|
2118 char_u lbuf[MAXWLEN + 1];
|
|
2119
|
|
2120 ga_init2(&ga, sizeof(langp_T), 2);
|
|
2121
|
|
2122 /* loop over comma separated languages. */
|
|
2123 for (lang = buf->b_p_spl; *lang != NUL; lang = e)
|
|
2124 {
|
|
2125 e = vim_strchr(lang, ',');
|
|
2126 if (e == NULL)
|
|
2127 e = lang + STRLEN(lang);
|
240
|
2128 region = NULL;
|
236
|
2129 if (e > lang + 2)
|
|
2130 {
|
|
2131 if (e - lang >= MAXWLEN)
|
|
2132 {
|
|
2133 ga_clear(&ga);
|
|
2134 return e_invarg;
|
|
2135 }
|
|
2136 if (lang[2] == '_')
|
|
2137 region = lang + 3;
|
|
2138 }
|
|
2139
|
|
2140 for (lp = first_lang; lp != NULL; lp = lp->sl_next)
|
|
2141 if (STRNICMP(lp->sl_name, lang, 2) == 0)
|
|
2142 break;
|
|
2143
|
|
2144 if (lp == NULL)
|
|
2145 {
|
|
2146 /* Not found, load the language. */
|
|
2147 STRNCPY(lbuf, lang, e - lang);
|
|
2148 lbuf[e - lang] = NUL;
|
|
2149 if (region != NULL)
|
|
2150 mch_memmove(lbuf + 2, lbuf + 5, e - lang - 4);
|
|
2151 lp = spell_load_lang(lbuf);
|
|
2152 }
|
|
2153
|
|
2154 if (lp != NULL)
|
|
2155 {
|
|
2156 if (region == NULL)
|
|
2157 region_mask = REGION_ALL;
|
|
2158 else
|
|
2159 {
|
|
2160 /* find region in sl_regions */
|
|
2161 c = find_region(lp->sl_regions, region);
|
|
2162 if (c == REGION_ALL)
|
|
2163 {
|
|
2164 c = *e;
|
|
2165 *e = NUL;
|
|
2166 smsg((char_u *)_("Warning: region %s not supported"), lang);
|
|
2167 *e = c;
|
|
2168 region_mask = REGION_ALL;
|
|
2169 }
|
|
2170 else
|
|
2171 region_mask = 1 << c;
|
|
2172 }
|
|
2173
|
|
2174 if (ga_grow(&ga, 1) == FAIL)
|
|
2175 {
|
|
2176 ga_clear(&ga);
|
|
2177 return e_outofmem;
|
|
2178 }
|
|
2179 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = lp;
|
|
2180 LANGP_ENTRY(ga, ga.ga_len)->lp_region = region_mask;
|
|
2181 ++ga.ga_len;
|
|
2182 }
|
|
2183
|
|
2184 if (*e == ',')
|
|
2185 ++e;
|
|
2186 }
|
|
2187
|
|
2188 /* Add a NULL entry to mark the end of the list. */
|
|
2189 if (ga_grow(&ga, 1) == FAIL)
|
|
2190 {
|
|
2191 ga_clear(&ga);
|
|
2192 return e_outofmem;
|
|
2193 }
|
|
2194 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = NULL;
|
|
2195 ++ga.ga_len;
|
|
2196
|
|
2197 /* Everything is fine, store the new b_langp value. */
|
|
2198 ga_clear(&buf->b_langp);
|
|
2199 buf->b_langp = ga;
|
|
2200
|
|
2201 return NULL;
|
|
2202 }
|
|
2203
|
|
2204 /*
|
|
2205 * Find the region "region[2]" in "rp" (points to "sl_regions").
|
|
2206 * Each region is simply stored as the two characters of it's name.
|
|
2207 * Returns the index if found, REGION_ALL if not found.
|
|
2208 */
|
|
2209 static int
|
|
2210 find_region(rp, region)
|
|
2211 char_u *rp;
|
|
2212 char_u *region;
|
|
2213 {
|
|
2214 int i;
|
|
2215
|
|
2216 for (i = 0; ; i += 2)
|
|
2217 {
|
|
2218 if (rp[i] == NUL)
|
|
2219 return REGION_ALL;
|
|
2220 if (rp[i] == region[0] && rp[i + 1] == region[1])
|
|
2221 break;
|
|
2222 }
|
|
2223 return i / 2;
|
|
2224 }
|
|
2225
|
|
2226 /*
|
|
2227 * Return type of word:
|
|
2228 * w word 0
|
|
2229 * Word BWF_ONECAP
|
|
2230 * W WORD BWF_ALLCAP
|
|
2231 * WoRd wOrd BWF_KEEPCAP
|
|
2232 */
|
|
2233 static int
|
|
2234 captype(word, end)
|
|
2235 char_u *word;
|
|
2236 char_u *end;
|
|
2237 {
|
|
2238 char_u *p;
|
|
2239 int c;
|
|
2240 int firstcap;
|
|
2241 int allcap;
|
|
2242 int past_second = FALSE; /* past second word char */
|
|
2243
|
|
2244 /* find first letter */
|
|
2245 for (p = word; !spell_iswordc(p); mb_ptr_adv(p))
|
|
2246 if (p >= end)
|
|
2247 return 0; /* only non-word characters, illegal word */
|
|
2248 #ifdef FEAT_MBYTE
|
|
2249 c = mb_ptr2char_adv(&p);
|
|
2250 #else
|
|
2251 c = *p++;
|
|
2252 #endif
|
255
|
2253 firstcap = allcap = spell_isupper(c);
|
236
|
2254
|
|
2255 /*
|
|
2256 * Need to check all letters to find a word with mixed upper/lower.
|
|
2257 * But a word with an upper char only at start is a ONECAP.
|
|
2258 */
|
|
2259 for ( ; p < end; mb_ptr_adv(p))
|
|
2260 if (spell_iswordc(p))
|
|
2261 {
|
|
2262 #ifdef FEAT_MBYTE
|
|
2263 c = mb_ptr2char(p);
|
|
2264 #else
|
|
2265 c = *p;
|
|
2266 #endif
|
255
|
2267 if (!spell_isupper(c))
|
236
|
2268 {
|
|
2269 /* UUl -> KEEPCAP */
|
|
2270 if (past_second && allcap)
|
|
2271 return BWF_KEEPCAP;
|
|
2272 allcap = FALSE;
|
|
2273 }
|
|
2274 else if (!allcap)
|
|
2275 /* UlU -> KEEPCAP */
|
|
2276 return BWF_KEEPCAP;
|
|
2277 past_second = TRUE;
|
|
2278 }
|
|
2279
|
|
2280 if (allcap)
|
|
2281 return BWF_ALLCAP;
|
|
2282 if (firstcap)
|
|
2283 return BWF_ONECAP;
|
|
2284 return 0;
|
|
2285 }
|
|
2286
|
|
2287 # if defined(FEAT_MBYTE) || defined(PROTO)
|
|
2288 /*
|
|
2289 * Clear all spelling tables and reload them.
|
|
2290 * Used after 'encoding' is set.
|
|
2291 */
|
|
2292 void
|
|
2293 spell_reload()
|
|
2294 {
|
|
2295 buf_T *buf;
|
|
2296 slang_T *lp;
|
|
2297
|
|
2298 /* Initialize the table for spell_iswordc(). */
|
|
2299 init_spell_chartab();
|
|
2300
|
|
2301 /* Unload all allocated memory. */
|
|
2302 while (first_lang != NULL)
|
|
2303 {
|
|
2304 lp = first_lang;
|
|
2305 first_lang = lp->sl_next;
|
|
2306 slang_free(lp);
|
|
2307 }
|
|
2308
|
|
2309 /* Go through all buffers and handle 'spelllang'. */
|
|
2310 for (buf = firstbuf; buf != NULL; buf = buf->b_next)
|
|
2311 {
|
|
2312 ga_clear(&buf->b_langp);
|
|
2313 if (*buf->b_p_spl != NUL)
|
|
2314 did_set_spelllang(buf);
|
|
2315 }
|
|
2316 }
|
|
2317 # endif
|
|
2318
|
|
2319 /*
|
|
2320 * Recognizing words uses a two-step mechanism:
|
|
2321 * 1. Locate a basic word, made out of word characters only and separated by
|
|
2322 * non-word characters.
|
|
2323 * 2. When a basic word is found, check if (possibly required) additions
|
|
2324 * before and after the word are present.
|
|
2325 *
|
|
2326 * Both mechanisms use affixes (prefixes and suffixes) to reduce the number of
|
|
2327 * words. When no matching word was found in the hashtable the start of the
|
|
2328 * word is checked for matching prefixes and the end of the word for matching
|
|
2329 * suffixes. All matching affixes are removed and then the resulting word is
|
|
2330 * searched for. If found it is checked if it supports the used affix.
|
|
2331 */
|
|
2332
|
|
2333
|
|
2334 #if defined(FEAT_MBYTE) || defined(PROTO)
|
|
2335 /*
|
|
2336 * Functions for ":mkspell".
|
|
2337 * Only possible with the multi-byte feature.
|
|
2338 */
|
|
2339
|
|
2340 #define MAXLINELEN 300 /* Maximum length in bytes of a line in a .aff
|
|
2341 and .dic file. */
|
|
2342 /*
|
|
2343 * Main structure to store the contents of a ".aff" file.
|
|
2344 */
|
|
2345 typedef struct afffile_S
|
|
2346 {
|
|
2347 char_u *af_enc; /* "SET", normalized, alloc'ed string or NULL */
|
|
2348 char_u *af_try; /* "TRY" line in "af_enc" encoding */
|
|
2349 hashtab_T af_pref; /* hashtable for prefixes, affheader_T */
|
|
2350 hashtab_T af_suff; /* hashtable for suffixes, affheader_T */
|
|
2351 garray_T af_rep; /* list of repentry_T entries from REP lines */
|
|
2352 } afffile_T;
|
|
2353
|
|
2354 typedef struct affentry_S affentry_T;
|
|
2355
|
|
2356 /* Affix header from ".aff" file. Used for af_pref and af_suff. */
|
|
2357 typedef struct affheader_S
|
|
2358 {
|
|
2359 char_u ah_key[2]; /* key for hashtable == name of affix entry */
|
|
2360 int ah_combine;
|
|
2361 affentry_T *ah_first; /* first affix entry */
|
|
2362 short_u ah_affnr; /* used in get_new_aff() */
|
|
2363 } affheader_T;
|
|
2364
|
|
2365 #define HI2AH(hi) ((affheader_T *)(hi)->hi_key)
|
|
2366
|
|
2367 /* Affix entry from ".aff" file. Used for prefixes and suffixes. */
|
|
2368 struct affentry_S
|
|
2369 {
|
|
2370 affentry_T *ae_next; /* next affix with same name/number */
|
|
2371 char_u *ae_chop; /* text to chop off basic word (can be NULL) */
|
|
2372 char_u *ae_add; /* text to add to basic word (can be NULL) */
|
240
|
2373 char_u *ae_add_nw; /* For a suffix: first non-word char in
|
|
2374 * "ae_add"; for a prefix with only non-word
|
|
2375 * chars: equal to "ae_add", for a prefix with
|
|
2376 * word and non-word chars: first non-word
|
|
2377 * char after word char. NULL otherwise. */
|
|
2378 char_u *ae_add_pw; /* For a prefix with both word and non-word
|
|
2379 * chars: first word char. NULL otherwise. */
|
243
|
2380 char_u ae_preword; /* TRUE for a prefix with one word */
|
236
|
2381 char_u *ae_cond; /* condition (NULL for ".") */
|
|
2382 regprog_T *ae_prog; /* regexp program for ae_cond or NULL */
|
|
2383 short_u ae_affnr; /* for old affix: new affix number */
|
|
2384 };
|
|
2385
|
|
2386 /*
|
|
2387 * Structure to store a word from a ".dic" file.
|
|
2388 */
|
|
2389 typedef struct dicword_S
|
|
2390 {
|
|
2391 char_u *dw_affnm; /* original affix names */
|
|
2392 char_u dw_word[1]; /* actually longer: the word in 'encoding' */
|
|
2393 } dicword_T;
|
|
2394
|
|
2395 static dicword_T dumdw;
|
|
2396 #define HI2DW(hi) ((dicword_T *)((hi)->hi_key - (dumdw.dw_word - (char_u *)&dumdw)))
|
|
2397
|
|
2398 /*
|
|
2399 * Structure to store a basic word for the spell file.
|
|
2400 * This is used for ":mkspell", not for spell checking.
|
|
2401 */
|
|
2402 typedef struct basicword_S basicword_T;
|
|
2403 struct basicword_S
|
|
2404 {
|
|
2405 basicword_T *bw_next; /* next word with same basic word */
|
|
2406 basicword_T *bw_cnext; /* next word with same caps */
|
|
2407 int bw_flags; /* BWF_ flags */
|
|
2408 garray_T bw_prefix; /* table with prefix numbers */
|
|
2409 garray_T bw_suffix; /* table with suffix numbers */
|
|
2410 int bw_region; /* region bits */
|
255
|
2411 char_u *bw_caseword; /* keep-case word or NULL */
|
|
2412 char_u *bw_leadstring; /* must come before bw_word or NULL */
|
|
2413 char_u *bw_addstring; /* must come after bw_word or NULL */
|
236
|
2414 char_u bw_word[1]; /* actually longer: word case folded */
|
|
2415 };
|
|
2416
|
|
2417 static basicword_T dumbw;
|
|
2418 #define KEY2BW(p) ((basicword_T *)((p) - (dumbw.bw_word - (char_u *)&dumbw)))
|
|
2419 #define HI2BW(hi) KEY2BW((hi)->hi_key)
|
|
2420
|
|
2421 /* Store the affix number related with a certain string. */
|
|
2422 typedef struct affhash_S
|
|
2423 {
|
|
2424 short_u as_nr; /* the affix nr */
|
|
2425 char_u as_word[1]; /* actually longer */
|
|
2426 } affhash_T;
|
|
2427
|
|
2428 static affhash_T dumas;
|
|
2429 #define HI2AS(hi) ((affhash_T *)((hi)->hi_key - (dumas.as_word - (char_u *)&dumas)))
|
|
2430
|
249
|
2431 /* info for writing the spell file */
|
|
2432 typedef struct winfo_S
|
|
2433 {
|
|
2434 FILE *wif_fd;
|
|
2435 basicword_T *wif_prevbw; /* last written basic word */
|
|
2436 int wif_regionmask; /* regions supported */
|
|
2437 int wif_prefm; /* 1 or 2 bytes used for prefix NR */
|
|
2438 int wif_suffm; /* 1 or 2 bytes used for suffix NR */
|
|
2439 long wif_wcount; /* written word count */
|
|
2440 long wif_acount; /* written addition count */
|
|
2441 long wif_addmax; /* max number of additions on one word */
|
|
2442 char_u *wif_addmaxw; /* word with max additions */
|
|
2443 } winfo_T;
|
|
2444
|
236
|
2445
|
240
|
2446 static afffile_T *spell_read_aff __ARGS((char_u *fname, vimconv_T *conv, int ascii));
|
236
|
2447 static void spell_free_aff __ARGS((afffile_T *aff));
|
240
|
2448 static int has_non_ascii __ARGS((char_u *s));
|
|
2449 static int spell_read_dic __ARGS((hashtab_T *ht, char_u *fname, vimconv_T *conv, int ascii));
|
|
2450 static int get_new_aff __ARGS((hashtab_T *oldaff, garray_T *gap, int prefix));
|
236
|
2451 static void spell_free_dic __ARGS((hashtab_T *dic));
|
|
2452 static int same_affentries __ARGS((affheader_T *ah1, affheader_T *ah2));
|
|
2453 static void add_affhash __ARGS((hashtab_T *ht, char_u *key, int newnr));
|
|
2454 static void clear_affhash __ARGS((hashtab_T *ht));
|
|
2455 static void trans_affixes __ARGS((dicword_T *dw, basicword_T *bw, afffile_T *oldaff, hashtab_T *newwords));
|
|
2456 static int build_wordlist __ARGS((hashtab_T *newwords, hashtab_T *oldwords, afffile_T *oldaff, int regionmask));
|
255
|
2457 static basicword_T *get_basicword __ARGS((char_u *word, int asize));
|
236
|
2458 static void combine_regions __ARGS((hashtab_T *newwords));
|
|
2459 static int same_affixes __ARGS((basicword_T *bw, basicword_T *nbw));
|
255
|
2460 static int expand_affixes __ARGS((hashtab_T *newwords, garray_T *prefgap, garray_T *suffgap));
|
|
2461 static int expand_one_aff __ARGS((basicword_T *bw, garray_T *add_words, affentry_T *pae, affentry_T *sae));
|
|
2462 static int add_to_wordlist __ARGS((hashtab_T *newwords, basicword_T *bw));
|
236
|
2463 static void write_affix __ARGS((FILE *fd, affheader_T *ah));
|
|
2464 static void write_affixlist __ARGS((FILE *fd, garray_T *aff, int bytes));
|
260
|
2465 static void write_vim_spell __ARGS((char_u *fname, garray_T *prefga, garray_T *suffga, hashtab_T *newwords, int regcount, char_u *regchars, int ascii));
|
249
|
2466 static void write_bword __ARGS((winfo_T *wif, basicword_T *bw, int lowcap));
|
236
|
2467 static void free_wordtable __ARGS((hashtab_T *ht));
|
|
2468 static void free_basicword __ARGS((basicword_T *bw));
|
|
2469 static void free_affixentries __ARGS((affentry_T *first));
|
240
|
2470 static void free_affix_entry __ARGS((affentry_T *ap));
|
236
|
2471
|
|
2472 /*
|
|
2473 * Read an affix ".aff" file.
|
|
2474 * Returns an afffile_T, NULL for failure.
|
|
2475 */
|
|
2476 static afffile_T *
|
240
|
2477 spell_read_aff(fname, conv, ascii)
|
236
|
2478 char_u *fname;
|
|
2479 vimconv_T *conv; /* info for encoding conversion */
|
240
|
2480 int ascii; /* Only accept ASCII characters */
|
236
|
2481 {
|
|
2482 FILE *fd;
|
|
2483 afffile_T *aff;
|
|
2484 char_u rline[MAXLINELEN];
|
|
2485 char_u *line;
|
|
2486 char_u *pc = NULL;
|
|
2487 char_u *(items[6]);
|
|
2488 int itemcnt;
|
|
2489 char_u *p;
|
|
2490 int lnum = 0;
|
|
2491 affheader_T *cur_aff = NULL;
|
|
2492 int aff_todo = 0;
|
|
2493 hashtab_T *tp;
|
255
|
2494 char_u *low = NULL;
|
|
2495 char_u *fol = NULL;
|
|
2496 char_u *upp = NULL;
|
236
|
2497
|
|
2498 fd = fopen((char *)fname, "r");
|
|
2499 if (fd == NULL)
|
|
2500 {
|
|
2501 EMSG2(_(e_notopen), fname);
|
|
2502 return NULL;
|
|
2503 }
|
|
2504
|
|
2505 smsg((char_u *)_("Reading affix file %s..."), fname);
|
|
2506 out_flush();
|
|
2507
|
|
2508 aff = (afffile_T *)alloc_clear((unsigned)sizeof(afffile_T));
|
|
2509 if (aff == NULL)
|
|
2510 return NULL;
|
|
2511 hash_init(&aff->af_pref);
|
|
2512 hash_init(&aff->af_suff);
|
|
2513 ga_init2(&aff->af_rep, (int)sizeof(repentry_T), 20);
|
|
2514
|
|
2515 /*
|
|
2516 * Read all the lines in the file one by one.
|
|
2517 */
|
255
|
2518 while (!vim_fgets(rline, MAXLINELEN, fd) && !got_int)
|
236
|
2519 {
|
255
|
2520 line_breakcheck();
|
236
|
2521 ++lnum;
|
|
2522
|
|
2523 /* Skip comment lines. */
|
|
2524 if (*rline == '#')
|
|
2525 continue;
|
|
2526
|
|
2527 /* Convert from "SET" to 'encoding' when needed. */
|
|
2528 vim_free(pc);
|
|
2529 if (conv->vc_type != CONV_NONE)
|
|
2530 {
|
|
2531 pc = string_convert(conv, rline, NULL);
|
255
|
2532 if (pc == NULL)
|
|
2533 {
|
|
2534 smsg((char_u *)_("Conversion failure for word in %s line %d: %s"),
|
|
2535 fname, lnum, rline);
|
|
2536 continue;
|
|
2537 }
|
236
|
2538 line = pc;
|
|
2539 }
|
|
2540 else
|
|
2541 {
|
|
2542 pc = NULL;
|
|
2543 line = rline;
|
|
2544 }
|
|
2545
|
|
2546 /* Split the line up in white separated items. Put a NUL after each
|
|
2547 * item. */
|
|
2548 itemcnt = 0;
|
|
2549 for (p = line; ; )
|
|
2550 {
|
|
2551 while (*p != NUL && *p <= ' ') /* skip white space and CR/NL */
|
|
2552 ++p;
|
|
2553 if (*p == NUL)
|
|
2554 break;
|
|
2555 items[itemcnt++] = p;
|
|
2556 while (*p > ' ') /* skip until white space or CR/NL */
|
|
2557 ++p;
|
|
2558 if (*p == NUL)
|
|
2559 break;
|
|
2560 *p++ = NUL;
|
|
2561 }
|
|
2562
|
|
2563 /* Handle non-empty lines. */
|
|
2564 if (itemcnt > 0)
|
|
2565 {
|
|
2566 if (STRCMP(items[0], "SET") == 0 && itemcnt == 2
|
|
2567 && aff->af_enc == NULL)
|
|
2568 {
|
|
2569 if (aff->af_enc != NULL)
|
|
2570 smsg((char_u *)_("Duplicate SET line ignored in %s line %d: %s"),
|
|
2571 fname, lnum, line);
|
|
2572 else
|
|
2573 {
|
|
2574 /* Setup for conversion from "ENC" to 'encoding'. */
|
|
2575 aff->af_enc = enc_canonize(items[1]);
|
240
|
2576 if (aff->af_enc != NULL && !ascii
|
236
|
2577 && convert_setup(conv, aff->af_enc, p_enc) == FAIL)
|
|
2578 smsg((char_u *)_("Conversion in %s not supported: from %s to %s"),
|
|
2579 fname, aff->af_enc, p_enc);
|
|
2580 }
|
|
2581 }
|
|
2582 else if (STRCMP(items[0], "TRY") == 0 && itemcnt == 2
|
|
2583 && aff->af_try == NULL)
|
|
2584 aff->af_try = vim_strsave(items[1]);
|
|
2585 else if ((STRCMP(items[0], "PFX") == 0
|
|
2586 || STRCMP(items[0], "SFX") == 0)
|
|
2587 && aff_todo == 0
|
|
2588 && itemcnt == 4)
|
|
2589 {
|
|
2590 /* New affix letter. */
|
|
2591 cur_aff = (affheader_T *)alloc((unsigned)sizeof(affheader_T));
|
|
2592 if (cur_aff == NULL)
|
|
2593 break;
|
|
2594 cur_aff->ah_key[0] = *items[1];
|
|
2595 cur_aff->ah_key[1] = NUL;
|
|
2596 if (items[1][1] != NUL)
|
|
2597 smsg((char_u *)_("Affix name too long in %s line %d: %s"),
|
|
2598 fname, lnum, items[1]);
|
|
2599 if (*items[2] == 'Y')
|
|
2600 cur_aff->ah_combine = TRUE;
|
|
2601 else if (*items[2] == 'N')
|
|
2602 cur_aff->ah_combine = FALSE;
|
|
2603 else if (p_verbose > 0)
|
290
|
2604 {
|
|
2605 verbose_enter();
|
236
|
2606 smsg((char_u *)_("Expected Y or N in %s line %d: %s"),
|
|
2607 fname, lnum, items[2]);
|
290
|
2608 verbose_leave();
|
|
2609 }
|
236
|
2610 cur_aff->ah_first = NULL;
|
|
2611 if (*items[0] == 'P')
|
|
2612 tp = &aff->af_pref;
|
|
2613 else
|
|
2614 tp = &aff->af_suff;
|
|
2615 if (!HASHITEM_EMPTY(hash_find(tp, cur_aff->ah_key)))
|
|
2616 smsg((char_u *)_("Duplicate affix in %s line %d: %s"),
|
|
2617 fname, lnum, items[1]);
|
|
2618 else
|
|
2619 hash_add(tp, cur_aff->ah_key);
|
|
2620
|
|
2621 aff_todo = atoi((char *)items[3]);
|
|
2622 }
|
|
2623 else if ((STRCMP(items[0], "PFX") == 0
|
|
2624 || STRCMP(items[0], "SFX") == 0)
|
|
2625 && aff_todo > 0
|
|
2626 && STRCMP(cur_aff->ah_key, items[1]) == 0
|
|
2627 && itemcnt == 5)
|
|
2628 {
|
|
2629 affentry_T *aff_entry;
|
|
2630
|
|
2631 /* New item for an affix letter. */
|
|
2632 --aff_todo;
|
|
2633 aff_entry = (affentry_T *)alloc_clear(
|
|
2634 (unsigned)sizeof(affentry_T));
|
|
2635 if (aff_entry == NULL)
|
|
2636 break;
|
240
|
2637
|
236
|
2638 if (STRCMP(items[2], "0") != 0)
|
|
2639 aff_entry->ae_chop = vim_strsave(items[2]);
|
|
2640 if (STRCMP(items[3], "0") != 0)
|
|
2641 aff_entry->ae_add = vim_strsave(items[3]);
|
|
2642 if (STRCMP(items[4], ".") != 0)
|
|
2643 {
|
|
2644 char_u buf[MAXLINELEN];
|
|
2645
|
|
2646 aff_entry->ae_cond = vim_strsave(items[4]);
|
|
2647 if (*items[0] == 'P')
|
|
2648 sprintf((char *)buf, "^%s", items[4]);
|
|
2649 else
|
|
2650 sprintf((char *)buf, "%s$", items[4]);
|
|
2651 aff_entry->ae_prog = vim_regcomp(buf, RE_MAGIC + RE_STRING);
|
|
2652 }
|
240
|
2653
|
|
2654 if (ascii && (has_non_ascii(aff_entry->ae_chop)
|
|
2655 || has_non_ascii(aff_entry->ae_add)))
|
|
2656 {
|
|
2657 /* Don't use an affix entry with non-ASCII characters when
|
|
2658 * "ascii" is TRUE. */
|
|
2659 free_affix_entry(aff_entry);
|
|
2660 }
|
|
2661 else
|
|
2662 {
|
|
2663 aff_entry->ae_next = cur_aff->ah_first;
|
|
2664 cur_aff->ah_first = aff_entry;
|
|
2665 }
|
236
|
2666 }
|
255
|
2667 else if (STRCMP(items[0], "FOL") == 0 && itemcnt == 2)
|
|
2668 {
|
|
2669 if (fol != NULL)
|
|
2670 smsg((char_u *)_("Duplicate FOL in %s line %d"),
|
|
2671 fname, lnum);
|
|
2672 else
|
|
2673 fol = vim_strsave(items[1]);
|
|
2674 }
|
|
2675 else if (STRCMP(items[0], "LOW") == 0 && itemcnt == 2)
|
|
2676 {
|
|
2677 if (low != NULL)
|
|
2678 smsg((char_u *)_("Duplicate LOW in %s line %d"),
|
|
2679 fname, lnum);
|
|
2680 else
|
|
2681 low = vim_strsave(items[1]);
|
|
2682 }
|
|
2683 else if (STRCMP(items[0], "UPP") == 0 && itemcnt == 2)
|
|
2684 {
|
|
2685 if (upp != NULL)
|
|
2686 smsg((char_u *)_("Duplicate UPP in %s line %d"),
|
|
2687 fname, lnum);
|
|
2688 else
|
|
2689 upp = vim_strsave(items[1]);
|
|
2690 }
|
236
|
2691 else if (STRCMP(items[0], "REP") == 0 && itemcnt == 2)
|
|
2692 /* Ignore REP count */;
|
|
2693 else if (STRCMP(items[0], "REP") == 0 && itemcnt == 3)
|
|
2694 {
|
|
2695 repentry_T *rp;
|
|
2696
|
|
2697 /* REP item */
|
|
2698 if (ga_grow(&aff->af_rep, 1) == FAIL)
|
|
2699 break;
|
|
2700 rp = ((repentry_T *)aff->af_rep.ga_data) + aff->af_rep.ga_len;
|
|
2701 rp->re_from = vim_strsave(items[1]);
|
|
2702 rp->re_to = vim_strsave(items[2]);
|
|
2703 ++aff->af_rep.ga_len;
|
|
2704 }
|
|
2705 else if (p_verbose > 0)
|
290
|
2706 {
|
|
2707 verbose_enter();
|
236
|
2708 smsg((char_u *)_("Unrecognized item in %s line %d: %s"),
|
|
2709 fname, lnum, items[0]);
|
290
|
2710 verbose_leave();
|
|
2711 }
|
236
|
2712 }
|
|
2713
|
|
2714 }
|
|
2715
|
255
|
2716 if (fol != NULL || low != NULL || upp != NULL)
|
|
2717 {
|
260
|
2718 /* Don't write a word table for an ASCII file, so that we don't check
|
|
2719 * for conflicts with a word table that matches 'encoding'. */
|
|
2720 if (!ascii)
|
|
2721 {
|
|
2722 if (fol == NULL || low == NULL || upp == NULL)
|
|
2723 smsg((char_u *)_("Missing FOL/LOW/UPP line in %s"), fname);
|
|
2724 else
|
|
2725 set_spell_chartab(fol, low, upp);
|
|
2726 }
|
255
|
2727
|
|
2728 vim_free(fol);
|
|
2729 vim_free(low);
|
|
2730 vim_free(upp);
|
|
2731 }
|
|
2732
|
236
|
2733 vim_free(pc);
|
|
2734 fclose(fd);
|
|
2735 return aff;
|
|
2736 }
|
|
2737
|
|
2738 /*
|
240
|
2739 * Return TRUE if string "s" contains a non-ASCII character (128 or higher).
|
|
2740 * When "s" is NULL FALSE is returned.
|
|
2741 */
|
|
2742 static int
|
|
2743 has_non_ascii(s)
|
|
2744 char_u *s;
|
|
2745 {
|
|
2746 char_u *p;
|
|
2747
|
|
2748 if (s != NULL)
|
|
2749 for (p = s; *p != NUL; ++p)
|
|
2750 if (*p >= 128)
|
|
2751 return TRUE;
|
|
2752 return FALSE;
|
|
2753 }
|
|
2754
|
|
2755 /*
|
236
|
2756 * Free the structure filled by spell_read_aff().
|
|
2757 */
|
|
2758 static void
|
|
2759 spell_free_aff(aff)
|
|
2760 afffile_T *aff;
|
|
2761 {
|
|
2762 hashtab_T *ht;
|
|
2763 hashitem_T *hi;
|
|
2764 int todo;
|
|
2765 int i;
|
|
2766 repentry_T *rp;
|
|
2767 affheader_T *ah;
|
|
2768
|
|
2769 vim_free(aff->af_enc);
|
|
2770 vim_free(aff->af_try);
|
|
2771
|
|
2772 for (ht = &aff->af_pref; ; ht = &aff->af_suff)
|
|
2773 {
|
|
2774 todo = ht->ht_used;
|
|
2775 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
2776 {
|
|
2777 if (!HASHITEM_EMPTY(hi))
|
|
2778 {
|
|
2779 --todo;
|
|
2780 ah = HI2AH(hi);
|
|
2781 free_affixentries(ah->ah_first);
|
|
2782 vim_free(ah);
|
|
2783 }
|
|
2784 }
|
|
2785 if (ht == &aff->af_suff)
|
|
2786 break;
|
|
2787 }
|
|
2788 hash_clear(&aff->af_pref);
|
|
2789 hash_clear(&aff->af_suff);
|
|
2790
|
|
2791 for (i = 0; i < aff->af_rep.ga_len; ++i)
|
|
2792 {
|
|
2793 rp = ((repentry_T *)aff->af_rep.ga_data) + i;
|
|
2794 vim_free(rp->re_from);
|
|
2795 vim_free(rp->re_to);
|
|
2796 }
|
|
2797 ga_clear(&aff->af_rep);
|
|
2798
|
|
2799 vim_free(aff);
|
|
2800 }
|
|
2801
|
|
2802 /*
|
|
2803 * Read a dictionary ".dic" file.
|
|
2804 * Returns OK or FAIL;
|
|
2805 * Each entry in the hashtab_T is a dicword_T.
|
|
2806 */
|
|
2807 static int
|
240
|
2808 spell_read_dic(ht, fname, conv, ascii)
|
236
|
2809 hashtab_T *ht;
|
|
2810 char_u *fname;
|
|
2811 vimconv_T *conv; /* info for encoding conversion */
|
240
|
2812 int ascii; /* only accept ASCII words */
|
236
|
2813 {
|
|
2814 char_u line[MAXLINELEN];
|
|
2815 char_u *p;
|
|
2816 dicword_T *dw;
|
|
2817 char_u *pc;
|
|
2818 char_u *w;
|
|
2819 int l;
|
|
2820 hash_T hash;
|
|
2821 hashitem_T *hi;
|
|
2822 FILE *fd;
|
|
2823 int lnum = 1;
|
|
2824
|
|
2825 fd = fopen((char *)fname, "r");
|
|
2826 if (fd == NULL)
|
|
2827 {
|
|
2828 EMSG2(_(e_notopen), fname);
|
|
2829 return FAIL;
|
|
2830 }
|
|
2831
|
|
2832 smsg((char_u *)_("Reading dictionary file %s..."), fname);
|
|
2833 out_flush();
|
|
2834
|
|
2835 /* Read and ignore the first line: word count. */
|
|
2836 (void)vim_fgets(line, MAXLINELEN, fd);
|
|
2837 if (!isdigit(*skipwhite(line)))
|
|
2838 EMSG2(_("E760: No word count in %s"), fname);
|
|
2839
|
|
2840 /*
|
|
2841 * Read all the lines in the file one by one.
|
|
2842 * The words are converted to 'encoding' here, before being added to
|
|
2843 * the hashtable.
|
|
2844 */
|
255
|
2845 while (!vim_fgets(line, MAXLINELEN, fd) && !got_int)
|
236
|
2846 {
|
255
|
2847 line_breakcheck();
|
236
|
2848 ++lnum;
|
|
2849
|
|
2850 /* Remove CR, LF and white space from end. */
|
|
2851 l = STRLEN(line);
|
|
2852 while (l > 0 && line[l - 1] <= ' ')
|
|
2853 --l;
|
|
2854 if (l == 0)
|
|
2855 continue; /* empty line */
|
|
2856 line[l] = NUL;
|
|
2857
|
|
2858 /* Find the optional affix names. */
|
|
2859 p = vim_strchr(line, '/');
|
|
2860 if (p != NULL)
|
|
2861 *p++ = NUL;
|
|
2862
|
240
|
2863 /* Skip non-ASCII words when "ascii" is TRUE. */
|
|
2864 if (ascii && has_non_ascii(line))
|
|
2865 continue;
|
|
2866
|
236
|
2867 /* Convert from "SET" to 'encoding' when needed. */
|
|
2868 if (conv->vc_type != CONV_NONE)
|
|
2869 {
|
|
2870 pc = string_convert(conv, line, NULL);
|
255
|
2871 if (pc == NULL)
|
|
2872 {
|
|
2873 smsg((char_u *)_("Conversion failure for word in %s line %d: %s"),
|
|
2874 fname, lnum, line);
|
|
2875 continue;
|
|
2876 }
|
236
|
2877 w = pc;
|
|
2878 }
|
|
2879 else
|
|
2880 {
|
|
2881 pc = NULL;
|
|
2882 w = line;
|
|
2883 }
|
|
2884
|
|
2885 dw = (dicword_T *)alloc_clear((unsigned)sizeof(dicword_T)
|
|
2886 + STRLEN(w));
|
|
2887 if (dw == NULL)
|
255
|
2888 {
|
|
2889 vim_free(pc);
|
236
|
2890 break;
|
255
|
2891 }
|
236
|
2892 STRCPY(dw->dw_word, w);
|
|
2893 vim_free(pc);
|
|
2894
|
|
2895 hash = hash_hash(dw->dw_word);
|
|
2896 hi = hash_lookup(ht, dw->dw_word, hash);
|
|
2897 if (!HASHITEM_EMPTY(hi))
|
|
2898 smsg((char_u *)_("Duplicate word in %s line %d: %s"),
|
|
2899 fname, lnum, line);
|
|
2900 else
|
|
2901 hash_add_item(ht, hi, dw->dw_word, hash);
|
|
2902
|
|
2903 if (p != NULL)
|
|
2904 dw->dw_affnm = vim_strsave(p);
|
|
2905 }
|
|
2906
|
|
2907 fclose(fd);
|
|
2908 return OK;
|
|
2909 }
|
|
2910
|
|
2911 /*
|
|
2912 * Free the structure filled by spell_read_dic().
|
|
2913 */
|
|
2914 static void
|
|
2915 spell_free_dic(dic)
|
|
2916 hashtab_T *dic;
|
|
2917 {
|
|
2918 int todo;
|
|
2919 dicword_T *dw;
|
|
2920 hashitem_T *hi;
|
|
2921
|
|
2922 todo = dic->ht_used;
|
|
2923 for (hi = dic->ht_array; todo > 0; ++hi)
|
|
2924 {
|
|
2925 if (!HASHITEM_EMPTY(hi))
|
|
2926 {
|
|
2927 --todo;
|
|
2928 dw = HI2DW(hi);
|
|
2929 vim_free(dw->dw_affnm);
|
|
2930 vim_free(dw);
|
|
2931 }
|
|
2932 }
|
|
2933 hash_clear(dic);
|
|
2934 }
|
|
2935
|
|
2936 /*
|
|
2937 * Take the affixes read by spell_read_aff() and add them to the new list.
|
|
2938 * Attempts to re-use the same number for identical affixes (ignoring the
|
|
2939 * condition, since we remove that). That is especially important when using
|
|
2940 * multiple regions.
|
|
2941 * Returns OK or FAIL;
|
|
2942 */
|
|
2943 static int
|
240
|
2944 get_new_aff(oldaff, gap, prefix)
|
236
|
2945 hashtab_T *oldaff; /* hashtable with affheader_T */
|
|
2946 garray_T *gap; /* table with new affixes */
|
240
|
2947 int prefix; /* TRUE when doing prefixes, FALSE for
|
|
2948 suffixes */
|
236
|
2949 {
|
|
2950 int oldtodo;
|
|
2951 affheader_T *oldah, *newah, *gapah;
|
|
2952 affentry_T *oldae, *newae;
|
|
2953 hashitem_T *oldhi;
|
|
2954 hashitem_T *hi;
|
|
2955 hashtab_T condht; /* conditions already found */
|
|
2956 char_u condkey[MAXLINELEN];
|
|
2957 int newnr;
|
|
2958 int gapnr;
|
|
2959 int retval = OK;
|
|
2960 char_u *p;
|
|
2961 garray_T tga;
|
243
|
2962 int preword;
|
236
|
2963
|
|
2964 /*
|
|
2965 * Loop over all the old affix names.
|
|
2966 */
|
|
2967 oldtodo = oldaff->ht_used;
|
|
2968 for (oldhi = oldaff->ht_array; oldtodo > 0 && retval == OK; ++oldhi)
|
|
2969 {
|
|
2970 if (!HASHITEM_EMPTY(oldhi))
|
|
2971 {
|
|
2972 --oldtodo;
|
|
2973 oldah = (affheader_T *)oldhi->hi_key;
|
|
2974
|
|
2975 /* Put entries with the same condition under the same new affix
|
|
2976 * nr in "tga". Use hashtable "condht" to find them. */
|
|
2977 ga_init2(&tga, sizeof(affheader_T), 10);
|
|
2978 hash_init(&condht);
|
|
2979
|
|
2980 /*
|
|
2981 * Loop over all affixes with the same name.
|
|
2982 * The affixes with the same condition will get the same number,
|
|
2983 * since they can be used with the same words.
|
|
2984 * 1. build the lists of new affentry_T, with the headers in "tga".
|
|
2985 * 2. Check if some of the lists already exist in "gap", re-use
|
|
2986 * their number.
|
|
2987 * 3. Assign the new numbers to the old affixes.
|
|
2988 */
|
|
2989
|
|
2990 /* 1. build the lists of new affentry_T. */
|
|
2991 for (oldae = oldah->ah_first; oldae != NULL && retval == OK;
|
|
2992 oldae = oldae->ae_next)
|
|
2993 {
|
243
|
2994 preword = FALSE;
|
236
|
2995 oldae->ae_add_nw = NULL;
|
240
|
2996 oldae->ae_add_pw = NULL;
|
236
|
2997 if (oldae->ae_add != NULL)
|
|
2998 {
|
240
|
2999 /* Check for non-word characters in the affix. If there
|
243
|
3000 * is one a suffix will be turned into an addition, a
|
|
3001 * prefix may be turned into a leadstring.
|
236
|
3002 * This is stored with the old affix, that is where
|
|
3003 * trans_affixes() will check. */
|
|
3004 for (p = oldae->ae_add; *p != NUL; mb_ptr_adv(p))
|
|
3005 if (!spell_iswordc(p))
|
240
|
3006 {
|
|
3007 oldae->ae_add_nw = p;
|
236
|
3008 break;
|
240
|
3009 }
|
|
3010
|
|
3011 if (prefix && oldae->ae_add_nw != NULL)
|
|
3012 {
|
243
|
3013 /* If a prefix has non-word characters special
|
|
3014 * treatment is necessary:
|
|
3015 * - If it has only non-word characters it becomes a
|
|
3016 * leadstring.
|
|
3017 * - If it has a sequence of word characters followed
|
|
3018 * by a non-word char it becomes a "preword": "d'",
|
|
3019 * "de-", "d'ai", etc.
|
|
3020 * - if it has another mix of word and non-word
|
|
3021 * characters the part before the last word char
|
|
3022 * becomes a leadstring: "'d", etc.
|
|
3023 */
|
240
|
3024 for (p = oldae->ae_add; *p != NUL; mb_ptr_adv(p))
|
|
3025 if (spell_iswordc(p))
|
|
3026 {
|
|
3027 oldae->ae_add_pw = p;
|
|
3028 break;
|
|
3029 }
|
|
3030 if (oldae->ae_add_pw != NULL)
|
|
3031 {
|
|
3032 /* Mixed prefix, set ae_add_nw to first non-word
|
|
3033 * char after ae_add_pw (if there is one). */
|
|
3034 oldae->ae_add_nw = NULL;
|
|
3035 for ( ; *p != NUL; mb_ptr_adv(p))
|
|
3036 if (!spell_iswordc(p))
|
|
3037 {
|
|
3038 oldae->ae_add_nw = p;
|
|
3039 break;
|
|
3040 }
|
243
|
3041 if (oldae->ae_add_nw != NULL)
|
|
3042 {
|
|
3043 preword = TRUE;
|
|
3044 oldae->ae_add_pw = NULL;
|
|
3045 oldae->ae_add_nw = NULL;
|
|
3046 }
|
240
|
3047 }
|
|
3048 }
|
236
|
3049 }
|
|
3050
|
|
3051 if (oldae->ae_cond == NULL)
|
|
3052 /* hashtable requires a non-empty key */
|
|
3053 STRCPY(condkey, "---");
|
|
3054 else
|
|
3055 STRCPY(condkey, oldae->ae_cond);
|
|
3056
|
|
3057 /* Look for an existing list with this name and condition. */
|
|
3058 hi = hash_find(&condht, condkey);
|
|
3059 if (!HASHITEM_EMPTY(hi))
|
|
3060 /* Match with existing affix, use that one. */
|
|
3061 newnr = HI2AS(hi)->as_nr;
|
|
3062 else
|
|
3063 {
|
|
3064 /* Add a new affix number. */
|
|
3065 newnr = tga.ga_len;
|
|
3066 if (ga_grow(&tga, 1) == FAIL)
|
|
3067 retval = FAIL;
|
|
3068 else
|
|
3069 {
|
|
3070 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
3071 newah->ah_combine = oldah->ah_combine;
|
|
3072 newah->ah_first = NULL;
|
|
3073 ++tga.ga_len;
|
|
3074
|
|
3075 /* Add the new list to the condht hashtable. */
|
|
3076 add_affhash(&condht, condkey, newnr);
|
|
3077 }
|
|
3078 }
|
|
3079
|
|
3080 /* Add the new affentry_T to the list. */
|
|
3081 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
3082 newae = (affentry_T *)alloc_clear((unsigned)sizeof(affentry_T));
|
|
3083 if (newae == NULL)
|
|
3084 retval = FAIL;
|
|
3085 else
|
|
3086 {
|
|
3087 newae->ae_next = newah->ah_first;
|
|
3088 newah->ah_first = newae;
|
|
3089 if (oldae->ae_chop == NULL)
|
|
3090 newae->ae_chop = NULL;
|
|
3091 else
|
|
3092 newae->ae_chop = vim_strsave(oldae->ae_chop);
|
|
3093 if (oldae->ae_add == NULL)
|
|
3094 newae->ae_add = NULL;
|
|
3095 else
|
|
3096 newae->ae_add = vim_strsave(oldae->ae_add);
|
243
|
3097 newae->ae_preword = preword;
|
236
|
3098
|
|
3099 /* The condition is not copied, since the new affix is
|
|
3100 * only used for words where the condition matches. */
|
|
3101 }
|
|
3102 }
|
|
3103
|
|
3104 /* 2. Check if some of the lists already exist, re-use their
|
|
3105 * number. Otherwise add the list to "gap". */
|
|
3106 for (newnr = 0; newnr < tga.ga_len; ++newnr)
|
|
3107 {
|
|
3108 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
3109 for (gapnr = 0; gapnr < gap->ga_len; ++gapnr)
|
|
3110 {
|
|
3111 gapah = ((affheader_T *)gap->ga_data) + gapnr;
|
243
|
3112 if (newah->ah_combine == gapah->ah_combine
|
|
3113 && same_affentries(newah, gapah))
|
236
|
3114 /* Found an existing affheader_T entry with same
|
|
3115 * affentry_T list, use its number. */
|
|
3116 break;
|
|
3117 }
|
|
3118
|
|
3119 newah->ah_affnr = gapnr;
|
|
3120 if (gapnr == gap->ga_len)
|
|
3121 {
|
|
3122 /* This is a new affentry_T list, add it. */
|
|
3123 if (ga_grow(gap, 1) == FAIL)
|
|
3124 retval = FAIL;
|
|
3125 else
|
|
3126 {
|
|
3127 *(((affheader_T *)gap->ga_data) + gap->ga_len) = *newah;
|
|
3128 ++gap->ga_len;
|
|
3129 }
|
|
3130 }
|
|
3131 else
|
|
3132 {
|
|
3133 /* free unused affentry_T list */
|
|
3134 free_affixentries(newah->ah_first);
|
|
3135 }
|
|
3136 }
|
|
3137
|
|
3138 /* 3. Assign the new affix numbers to the old affixes. */
|
|
3139 for (oldae = oldah->ah_first; oldae != NULL && retval == OK;
|
|
3140 oldae = oldae->ae_next)
|
|
3141 {
|
|
3142 if (oldae->ae_cond == NULL)
|
|
3143 /* hashtable requires a non-empty key */
|
|
3144 STRCPY(condkey, "---");
|
|
3145 else
|
|
3146 STRCPY(condkey, oldae->ae_cond);
|
|
3147
|
|
3148 /* Look for an existing affix with this name and condition. */
|
|
3149 hi = hash_find(&condht, condkey);
|
|
3150 if (!HASHITEM_EMPTY(hi))
|
|
3151 /* Match with existing affix, use that one. */
|
|
3152 newnr = HI2AS(hi)->as_nr;
|
|
3153 else
|
|
3154 {
|
|
3155 EMSG(_(e_internal));
|
|
3156 retval = FAIL;
|
|
3157 }
|
|
3158 newah = ((affheader_T *)tga.ga_data) + newnr;
|
|
3159 oldae->ae_affnr = newah->ah_affnr;
|
|
3160 }
|
|
3161
|
|
3162 ga_clear(&tga);
|
|
3163 clear_affhash(&condht);
|
|
3164 }
|
|
3165 }
|
|
3166
|
|
3167 return retval;
|
|
3168 }
|
|
3169
|
|
3170 /*
|
|
3171 * Return TRUE if the affentry_T lists for "ah1" and "ah2" contain the same
|
|
3172 * items, ignoring the order.
|
|
3173 * Only compares the chop and add strings, not the condition.
|
|
3174 */
|
|
3175 static int
|
|
3176 same_affentries(ah1, ah2)
|
|
3177 affheader_T *ah1;
|
|
3178 affheader_T *ah2;
|
|
3179 {
|
|
3180 affentry_T *ae1, *ae2;
|
|
3181
|
|
3182 /* Check the length of the lists first. */
|
|
3183 ae2 = ah2->ah_first;
|
|
3184 for (ae1 = ah1->ah_first; ae1 != NULL; ae1 = ae1->ae_next)
|
|
3185 {
|
|
3186 if (ae2 == NULL)
|
|
3187 return FALSE; /* "ah1" list is longer */
|
|
3188 ae2 = ae2->ae_next;
|
|
3189 }
|
|
3190 if (ae2 != NULL)
|
|
3191 return FALSE; /* "ah2" list is longer */
|
|
3192
|
|
3193 /* Check that each entry in "ah1" appears in "ah2". */
|
|
3194 for (ae1 = ah1->ah_first; ae1 != NULL; ae1 = ae1->ae_next)
|
|
3195 {
|
|
3196 for (ae2 = ah2->ah_first; ae2 != NULL; ae2 = ae2->ae_next)
|
|
3197 {
|
|
3198 if ((ae1->ae_chop == NULL) == (ae2->ae_chop == NULL)
|
|
3199 && (ae1->ae_add == NULL) == (ae2->ae_add == NULL)
|
|
3200 && (ae1->ae_chop == NULL
|
|
3201 || STRCMP(ae1->ae_chop, ae2->ae_chop) == 0)
|
|
3202 && (ae1->ae_add == NULL
|
|
3203 || STRCMP(ae1->ae_add, ae2->ae_add) == 0))
|
|
3204 break;
|
|
3205 }
|
|
3206 if (ae2 == NULL)
|
|
3207 return FALSE;
|
|
3208 }
|
|
3209
|
|
3210 return TRUE;
|
|
3211 }
|
|
3212
|
|
3213 /*
|
|
3214 * Add a chop/add or cond hashtable entry.
|
|
3215 */
|
|
3216 static void
|
|
3217 add_affhash(ht, key, newnr)
|
|
3218 hashtab_T *ht;
|
|
3219 char_u *key;
|
|
3220 int newnr;
|
|
3221 {
|
|
3222 affhash_T *as;
|
|
3223
|
|
3224 as = (affhash_T *)alloc((unsigned)sizeof(affhash_T) + STRLEN(key));
|
|
3225 if (as != NULL)
|
|
3226 {
|
|
3227 as->as_nr = newnr;
|
|
3228 STRCPY(as->as_word, key);
|
|
3229 hash_add(ht, as->as_word);
|
|
3230 }
|
|
3231 }
|
|
3232
|
|
3233 /*
|
|
3234 * Clear the chop/add hashtable used to detect identical affixes.
|
|
3235 */
|
|
3236 static void
|
|
3237 clear_affhash(ht)
|
|
3238 hashtab_T *ht;
|
|
3239 {
|
|
3240 int todo;
|
|
3241 hashitem_T *hi;
|
|
3242
|
|
3243 todo = ht->ht_used;
|
|
3244 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
3245 {
|
|
3246 if (!HASHITEM_EMPTY(hi))
|
|
3247 {
|
|
3248 --todo;
|
|
3249 vim_free(HI2AS(hi));
|
|
3250 }
|
|
3251 }
|
|
3252 hash_clear(ht);
|
|
3253 }
|
|
3254
|
|
3255 /*
|
|
3256 * Translate list of affix names for an old word to affix numbers in a new
|
|
3257 * basic word.
|
|
3258 * This checks if the conditions match with the old word. The result is that
|
|
3259 * the new affix does not need to store the condition.
|
|
3260 */
|
|
3261 static void
|
|
3262 trans_affixes(dw, bw, oldaff, newwords)
|
|
3263 dicword_T *dw; /* old word */
|
|
3264 basicword_T *bw; /* basic word */
|
|
3265 afffile_T *oldaff; /* affixes for "oldwords" */
|
|
3266 hashtab_T *newwords; /* table with words */
|
|
3267 {
|
|
3268 char_u key[2];
|
|
3269 char_u *p;
|
|
3270 char_u *affnm;
|
255
|
3271 garray_T *gap, *agap;
|
236
|
3272 hashitem_T *aff_hi;
|
|
3273 affheader_T *ah;
|
|
3274 affentry_T *ae;
|
|
3275 regmatch_T regmatch;
|
|
3276 int i;
|
|
3277 basicword_T *nbw;
|
|
3278 int alen;
|
240
|
3279 garray_T suffixga; /* list of words with non-word suffixes */
|
|
3280 garray_T prefixga; /* list of words with non-word prefixes */
|
236
|
3281 char_u nword[MAXWLEN];
|
|
3282 int flags;
|
|
3283 int n;
|
|
3284
|
240
|
3285 ga_init2(&suffixga, (int)sizeof(basicword_T *), 5);
|
|
3286 ga_init2(&prefixga, (int)sizeof(basicword_T *), 5);
|
236
|
3287
|
|
3288 /* Loop over all the affix names of the old word. */
|
|
3289 key[1] = NUL;
|
|
3290 for (affnm = dw->dw_affnm; *affnm != NUL; ++affnm)
|
|
3291 {
|
|
3292 key[0] = *affnm;
|
|
3293 aff_hi = hash_find(&oldaff->af_pref, key);
|
|
3294 if (!HASHITEM_EMPTY(aff_hi))
|
|
3295 gap = &bw->bw_prefix; /* found a prefix */
|
|
3296 else
|
|
3297 {
|
|
3298 gap = &bw->bw_suffix; /* must be a suffix */
|
|
3299 aff_hi = hash_find(&oldaff->af_suff, key);
|
|
3300 if (HASHITEM_EMPTY(aff_hi))
|
|
3301 {
|
|
3302 smsg((char_u *)_("No affix entry '%s' for word %s"),
|
|
3303 key, dw->dw_word);
|
|
3304 continue;
|
|
3305 }
|
|
3306 }
|
|
3307
|
|
3308 /* Loop over all the affix entries for this affix name. */
|
|
3309 ah = HI2AH(aff_hi);
|
|
3310 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
|
3311 {
|
243
|
3312 /* Setup for regexp matching. Note that we don't ignore case.
|
255
|
3313 * This is weird, because the rules in an .aff file don't care
|
243
|
3314 * about case, but it's necessary for compatibility with Myspell.
|
|
3315 */
|
236
|
3316 regmatch.regprog = ae->ae_prog;
|
243
|
3317 regmatch.rm_ic = FALSE;
|
236
|
3318 if (ae->ae_prog == NULL
|
|
3319 || vim_regexec(®match, dw->dw_word, (colnr_T)0))
|
|
3320 {
|
240
|
3321 if ((ae->ae_add_nw != NULL || ae->ae_add_pw != NULL)
|
|
3322 && (gap != &bw->bw_suffix || bw->bw_addstring == NULL))
|
236
|
3323 {
|
255
|
3324 /*
|
|
3325 * Affix has a non-word character and isn't prepended to
|
236
|
3326 * leader or appended to addition. Need to use another
|
255
|
3327 * word with a leadstring and/or addstring.
|
|
3328 */
|
|
3329 if (gap == &bw->bw_suffix || ae->ae_add_nw == NULL)
|
236
|
3330 {
|
255
|
3331 /* Suffix or prefix with only non-word chars.
|
|
3332 * Build the new basic word in "nword": Remove chop
|
|
3333 * string and append/prepend addition. */
|
|
3334 if (gap == &bw->bw_suffix)
|
236
|
3335 {
|
255
|
3336 /* suffix goes at the end of the word */
|
236
|
3337 STRCPY(nword, dw->dw_word);
|
|
3338 if (ae->ae_chop != NULL)
|
|
3339 {
|
|
3340 /* Remove chop string. */
|
|
3341 p = nword + STRLEN(nword);
|
|
3342 for (i = mb_charlen(ae->ae_chop); i > 0; --i)
|
|
3343 mb_ptr_back(nword, p);
|
|
3344 *p = NUL;
|
|
3345 }
|
|
3346 STRCAT(nword, ae->ae_add);
|
255
|
3347 agap = &suffixga;
|
240
|
3348 }
|
255
|
3349 else
|
240
|
3350 {
|
255
|
3351 /* prefix goes before the word */
|
240
|
3352 STRCPY(nword, ae->ae_add);
|
|
3353 p = dw->dw_word;
|
|
3354 if (ae->ae_chop != NULL)
|
|
3355 /* Skip chop string. */
|
|
3356 for (i = mb_charlen(ae->ae_chop); i > 0; --i)
|
|
3357 mb_ptr_adv( p);
|
|
3358 STRCAT(nword, p);
|
255
|
3359 agap = &prefixga;
|
|
3360 }
|
|
3361
|
|
3362 /* Create a basicword_T from the word. */
|
|
3363 nbw = get_basicword(nword, 1);
|
|
3364 if (nbw != NULL)
|
|
3365 {
|
|
3366 nbw->bw_region = bw->bw_region;
|
|
3367 nbw->bw_flags |= bw->bw_flags
|
|
3368 & ~(BWF_ONECAP | BWF_ALLCAP | BWF_KEEPCAP);
|
|
3369
|
|
3370 if (STRCMP(bw->bw_word, nbw->bw_word) != 0)
|
|
3371 /* Basic word differs, add new word entry. */
|
|
3372 (void)add_to_wordlist(newwords, nbw);
|
240
|
3373 else
|
|
3374 {
|
|
3375 /* Basic word is the same, link "nbw" after
|
|
3376 * "bw". */
|
255
|
3377 nbw->bw_next = bw->bw_next;
|
240
|
3378 bw->bw_next = nbw;
|
|
3379 }
|
|
3380
|
255
|
3381 /* Remember this word, we need to set bw_prefix
|
|
3382 * or bw_suffix later. */
|
|
3383 if (ga_grow(agap, 1) == OK)
|
|
3384 ((basicword_T **)agap->ga_data)
|
|
3385 [agap->ga_len++] = nbw;
|
236
|
3386 }
|
|
3387 }
|
|
3388 else
|
|
3389 {
|
240
|
3390 /* Prefix with both non-word and word characters: Turn
|
|
3391 * prefix into basic word, original word becomes an
|
|
3392 * addstring. */
|
|
3393
|
|
3394 /* Fold-case the word characters in the prefix into
|
|
3395 * nword[]. */
|
|
3396 alen = 0;
|
|
3397 for (p = ae->ae_add_pw; p < ae->ae_add_nw; p += n)
|
|
3398 {
|
|
3399 #ifdef FEAT_MBYTE
|
|
3400 n = (*mb_ptr2len_check)(p);
|
|
3401 #else
|
|
3402 n = 1;
|
|
3403 #endif
|
255
|
3404 (void)spell_casefold(p, n, nword + alen,
|
240
|
3405 MAXWLEN - alen);
|
|
3406 alen += STRLEN(nword + alen);
|
|
3407 }
|
|
3408
|
|
3409 /* Allocate a new word entry. */
|
|
3410 nbw = (basicword_T *)alloc((unsigned)(
|
|
3411 sizeof(basicword_T) + alen + 1));
|
|
3412 if (nbw != NULL)
|
|
3413 {
|
|
3414 *nbw = *bw;
|
|
3415 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
3416 ga_init2(&nbw->bw_suffix, sizeof(short_u), 1);
|
|
3417
|
|
3418 mch_memmove(nbw->bw_word, nword, alen);
|
|
3419 nbw->bw_word[alen] = NUL;
|
|
3420
|
|
3421 /* Use the cap type of the prefix. */
|
|
3422 alen = ae->ae_add_nw - ae->ae_add_pw;
|
|
3423 mch_memmove(nword, ae->ae_add_pw, alen);
|
|
3424 nword[alen] = NUL;
|
|
3425 flags = captype(nword, nword + STRLEN(nword));
|
|
3426 if (flags & BWF_KEEPCAP)
|
|
3427 nbw->bw_caseword = vim_strsave(nword);
|
|
3428 else
|
|
3429 nbw->bw_caseword = NULL;
|
|
3430 nbw->bw_flags &= ~(BWF_ONECAP | BWF_ALLCAP
|
|
3431 | BWF_KEEPCAP);
|
|
3432 nbw->bw_flags |= flags;
|
|
3433
|
|
3434 /* The addstring is the prefix after the word
|
|
3435 * characters, the original word excluding "chop",
|
|
3436 * plus any addition. */
|
|
3437 STRCPY(nword, ae->ae_add_nw);
|
|
3438 p = bw->bw_word;
|
|
3439 if (ae->ae_chop != NULL)
|
|
3440 p += STRLEN(ae->ae_chop);
|
|
3441 STRCAT(nword, p);
|
|
3442 if (bw->bw_addstring != NULL)
|
|
3443 STRCAT(nword, bw->bw_addstring);
|
|
3444 nbw->bw_addstring = vim_strsave(nword);
|
|
3445
|
|
3446 if (ae->ae_add_pw > ae->ae_add)
|
|
3447 nbw->bw_leadstring = vim_strnsave(ae->ae_add,
|
|
3448 ae->ae_add_pw - ae->ae_add);
|
|
3449 else
|
|
3450 nbw->bw_leadstring = NULL;
|
|
3451
|
255
|
3452 (void)add_to_wordlist(newwords, nbw);
|
240
|
3453
|
|
3454 /* Remember this word, we need to set bw_suffix
|
|
3455 * and bw_suffix later. */
|
|
3456 if (ga_grow(&prefixga, 1) == OK)
|
|
3457 ((basicword_T **)prefixga.ga_data)
|
|
3458 [prefixga.ga_len++] = nbw;
|
|
3459 }
|
236
|
3460 }
|
|
3461 }
|
|
3462 else
|
|
3463 {
|
|
3464 /* Affix applies to this word, add the related affix
|
|
3465 * number. But only if it's not there yet. And keep the
|
|
3466 * list sorted, so that we can compare it later. */
|
|
3467 for (i = 0; i < gap->ga_len; ++i)
|
|
3468 {
|
|
3469 n = ((short_u *)gap->ga_data)[i];
|
|
3470 if (n >= ae->ae_affnr)
|
|
3471 {
|
|
3472 if (n == ae->ae_affnr)
|
|
3473 i = -1;
|
|
3474 break;
|
|
3475 }
|
|
3476 }
|
|
3477 if (i >= 0 && ga_grow(gap, 1) == OK)
|
|
3478 {
|
|
3479 if (i < gap->ga_len)
|
|
3480 mch_memmove(((short_u *)gap->ga_data) + i + 1,
|
|
3481 ((short_u *)gap->ga_data) + i,
|
|
3482 sizeof(short_u) * (gap->ga_len - i));
|
|
3483 ((short_u *)gap->ga_data)[i] = ae->ae_affnr;
|
|
3484 ++gap->ga_len;
|
|
3485 }
|
|
3486 }
|
|
3487 }
|
|
3488 }
|
|
3489 }
|
|
3490
|
|
3491 /*
|
|
3492 * For the words that we added for suffixes with non-word characters: Use
|
|
3493 * the prefix list of the main word.
|
|
3494 */
|
240
|
3495 for (i = 0; i < suffixga.ga_len; ++i)
|
236
|
3496 {
|
240
|
3497 nbw = ((basicword_T **)suffixga.ga_data)[i];
|
236
|
3498 if (ga_grow(&nbw->bw_prefix, bw->bw_prefix.ga_len) == OK)
|
|
3499 {
|
|
3500 mch_memmove(nbw->bw_prefix.ga_data, bw->bw_prefix.ga_data,
|
|
3501 bw->bw_prefix.ga_len * sizeof(short_u));
|
|
3502 nbw->bw_prefix.ga_len = bw->bw_prefix.ga_len;
|
|
3503 }
|
|
3504 }
|
|
3505
|
240
|
3506 /*
|
|
3507 * For the words that we added for prefixes with non-word characters: Use
|
|
3508 * the suffix list of the main word.
|
|
3509 */
|
|
3510 for (i = 0; i < prefixga.ga_len; ++i)
|
|
3511 {
|
|
3512 nbw = ((basicword_T **)prefixga.ga_data)[i];
|
|
3513 if (ga_grow(&nbw->bw_suffix, bw->bw_suffix.ga_len) == OK)
|
|
3514 {
|
|
3515 mch_memmove(nbw->bw_suffix.ga_data, bw->bw_suffix.ga_data,
|
|
3516 bw->bw_suffix.ga_len * sizeof(short_u));
|
|
3517 nbw->bw_suffix.ga_len = bw->bw_suffix.ga_len;
|
|
3518 }
|
|
3519 }
|
|
3520
|
|
3521 ga_clear(&suffixga);
|
|
3522 ga_clear(&prefixga);
|
236
|
3523 }
|
|
3524
|
|
3525 /*
|
|
3526 * Go over all words in "oldwords" and change the old affix names to the new
|
|
3527 * affix numbers, check the conditions, fold case, extract the basic word and
|
|
3528 * additions.
|
|
3529 */
|
|
3530 static int
|
|
3531 build_wordlist(newwords, oldwords, oldaff, regionmask)
|
|
3532 hashtab_T *newwords; /* basicword_T entries */
|
|
3533 hashtab_T *oldwords; /* dicword_T entries */
|
|
3534 afffile_T *oldaff; /* affixes for "oldwords" */
|
|
3535 int regionmask; /* value for bw_region */
|
|
3536 {
|
|
3537 int todo;
|
|
3538 hashitem_T *old_hi;
|
|
3539 dicword_T *dw;
|
|
3540 basicword_T *bw;
|
240
|
3541 char_u message[MAXLINELEN + MAXWLEN];
|
236
|
3542
|
|
3543 todo = oldwords->ht_used;
|
|
3544 for (old_hi = oldwords->ht_array; todo > 0; ++old_hi)
|
|
3545 {
|
|
3546 if (!HASHITEM_EMPTY(old_hi))
|
|
3547 {
|
|
3548 --todo;
|
|
3549 dw = HI2DW(old_hi);
|
|
3550
|
|
3551 /* This takes time, print a message now and then. */
|
240
|
3552 if ((todo & 0x3ff) == 0 || todo == (int)oldwords->ht_used - 1)
|
236
|
3553 {
|
240
|
3554 sprintf((char *)message, _("%6d todo - %s"),
|
|
3555 todo, dw->dw_word);
|
|
3556 msg_start();
|
|
3557 msg_outtrans_attr(message, 0);
|
|
3558 msg_clr_eos();
|
|
3559 msg_didout = FALSE;
|
|
3560 msg_col = 0;
|
236
|
3561 out_flush();
|
|
3562 ui_breakcheck();
|
|
3563 if (got_int)
|
|
3564 break;
|
|
3565 }
|
|
3566
|
255
|
3567 bw = get_basicword(dw->dw_word, 10);
|
236
|
3568 if (bw == NULL)
|
|
3569 break;
|
|
3570 bw->bw_region = regionmask;
|
|
3571
|
255
|
3572 (void)add_to_wordlist(newwords, bw);
|
236
|
3573
|
|
3574 /* Deal with any affix names on the old word, translate them
|
|
3575 * into affix numbers. */
|
|
3576 if (dw->dw_affnm != NULL)
|
|
3577 trans_affixes(dw, bw, oldaff, newwords);
|
|
3578 }
|
|
3579 }
|
|
3580 if (todo > 0)
|
|
3581 return FAIL;
|
|
3582 return OK;
|
|
3583 }
|
|
3584
|
|
3585 /*
|
255
|
3586 * Get a basicword_T from a word in original case.
|
|
3587 * Caller must set bw_region.
|
|
3588 * Returns NULL when something fails.
|
|
3589 */
|
|
3590 static basicword_T *
|
|
3591 get_basicword(word, asize)
|
|
3592 char_u *word;
|
|
3593 int asize; /* growsize for affix garray */
|
|
3594 {
|
|
3595 int dwlen;
|
|
3596 char_u foldword[MAXLINELEN];
|
|
3597 int flags;
|
|
3598 int clen;
|
|
3599 int leadlen;
|
|
3600 char_u *p;
|
|
3601 char_u leadstring[MAXLINELEN];
|
|
3602 int addlen;
|
|
3603 char_u addstring[MAXLINELEN];
|
|
3604 char_u *cp = NULL;
|
|
3605 int l;
|
|
3606 basicword_T *bw;
|
|
3607
|
|
3608 /* The basic words are always stored with folded case. */
|
|
3609 dwlen = STRLEN(word);
|
|
3610 (void)spell_casefold(word, dwlen, foldword, MAXLINELEN);
|
|
3611 flags = captype(word, word + dwlen);
|
|
3612
|
|
3613 /* Check for non-word characters before the word. */
|
|
3614 clen = 0;
|
|
3615 leadlen = 0;
|
|
3616 if (!spell_iswordc(foldword))
|
|
3617 {
|
|
3618 p = foldword;
|
|
3619 for (;;)
|
|
3620 {
|
|
3621 mb_ptr_adv(p);
|
|
3622 ++clen;
|
|
3623 if (*p == NUL) /* Only non-word chars (bad word!) */
|
|
3624 {
|
|
3625 if (p_verbose > 0)
|
290
|
3626 {
|
|
3627 verbose_enter();
|
255
|
3628 smsg((char_u *)_("Warning: word without word characters: \"%s\""),
|
|
3629 foldword);
|
290
|
3630 verbose_leave();
|
|
3631 }
|
255
|
3632 break;
|
|
3633 }
|
|
3634 if (spell_iswordc(p))
|
|
3635 {
|
|
3636 /* Move the leader to "leadstring" and remove it from
|
|
3637 * "foldword". */
|
|
3638 leadlen = p - foldword;
|
|
3639 mch_memmove(leadstring, foldword, leadlen);
|
|
3640 leadstring[leadlen] = NUL;
|
|
3641 mch_memmove(foldword, p, STRLEN(p) + 1);
|
|
3642 break;
|
|
3643 }
|
|
3644 }
|
|
3645 }
|
|
3646
|
|
3647 /* Check for non-word characters after word characters. */
|
|
3648 addlen = 0;
|
|
3649 for (p = foldword; spell_iswordc(p); mb_ptr_adv(p))
|
|
3650 {
|
|
3651 if (*p == NUL)
|
|
3652 break;
|
|
3653 ++clen;
|
|
3654 }
|
|
3655 if (*p != NUL)
|
|
3656 {
|
|
3657 /* Move the addition to "addstring" and truncate "foldword". */
|
|
3658 if (flags & BWF_KEEPCAP)
|
|
3659 {
|
|
3660 /* Preserve caps, need to skip the right number of
|
|
3661 * characters in the original word (case folding may
|
|
3662 * change the byte count). */
|
|
3663 l = 0;
|
|
3664 for (cp = word; l < clen; mb_ptr_adv(cp))
|
|
3665 ++l;
|
|
3666 addlen = STRLEN(cp);
|
|
3667 mch_memmove(addstring, cp, addlen + 1);
|
|
3668 }
|
|
3669 else
|
|
3670 {
|
|
3671 addlen = STRLEN(p);
|
|
3672 mch_memmove(addstring, p, addlen + 1);
|
|
3673 }
|
|
3674 *p = NUL;
|
|
3675 }
|
|
3676
|
|
3677 bw = (basicword_T *)alloc_clear((unsigned)sizeof(basicword_T)
|
|
3678 + STRLEN(foldword));
|
|
3679 if (bw == NULL)
|
|
3680 return NULL;
|
|
3681
|
|
3682 STRCPY(bw->bw_word, foldword);
|
|
3683
|
|
3684 if (leadlen > 0)
|
|
3685 bw->bw_leadstring = vim_strsave(leadstring);
|
|
3686 else
|
|
3687 bw->bw_leadstring = NULL;
|
|
3688 if (addlen > 0)
|
|
3689 bw->bw_addstring = vim_strsave(addstring);
|
|
3690 else
|
|
3691 bw->bw_addstring = NULL;
|
|
3692
|
|
3693 if (flags & BWF_KEEPCAP)
|
|
3694 {
|
|
3695 if (addlen == 0)
|
|
3696 /* use the whole word */
|
|
3697 bw->bw_caseword = vim_strsave(word + leadlen);
|
|
3698 else
|
|
3699 /* use only up to the addition */
|
|
3700 bw->bw_caseword = vim_strnsave(word + leadlen,
|
|
3701 cp - word - leadlen);
|
|
3702 }
|
|
3703
|
|
3704 bw->bw_flags = flags;
|
|
3705 ga_init2(&bw->bw_prefix, sizeof(short_u), asize);
|
|
3706 ga_init2(&bw->bw_suffix, sizeof(short_u), asize);
|
|
3707
|
|
3708 return bw;
|
|
3709 }
|
|
3710
|
|
3711 /*
|
236
|
3712 * Go through the list of words and combine the ones that are identical except
|
|
3713 * for the region.
|
|
3714 */
|
|
3715 static void
|
|
3716 combine_regions(newwords)
|
|
3717 hashtab_T *newwords;
|
|
3718 {
|
|
3719 int todo;
|
|
3720 hashitem_T *hi;
|
|
3721 basicword_T *bw, *nbw, *pbw;
|
|
3722
|
|
3723 /* Loop over all basic words in the words table. */
|
|
3724 todo = newwords->ht_used;
|
|
3725 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
3726 {
|
|
3727 if (!HASHITEM_EMPTY(hi))
|
|
3728 {
|
|
3729 --todo;
|
|
3730
|
|
3731 /* Loop over the list of words for this basic word. Compare with
|
|
3732 * each following word in the same list. */
|
|
3733 for (bw = HI2BW(hi); bw != NULL; bw = bw->bw_next)
|
|
3734 {
|
|
3735 pbw = bw;
|
|
3736 for (nbw = pbw->bw_next; nbw != NULL; nbw = pbw->bw_next)
|
|
3737 {
|
|
3738 if (bw->bw_flags == nbw->bw_flags
|
|
3739 && (bw->bw_leadstring == NULL)
|
|
3740 == (nbw->bw_leadstring == NULL)
|
|
3741 && (bw->bw_addstring == NULL)
|
|
3742 == (nbw->bw_addstring == NULL)
|
|
3743 && ((bw->bw_flags & BWF_KEEPCAP) == 0
|
255
|
3744 || bw->bw_caseword == NULL
|
|
3745 || nbw->bw_caseword == NULL
|
|
3746 || STRCMP(bw->bw_caseword,
|
|
3747 nbw->bw_caseword) == 0)
|
236
|
3748 && (bw->bw_leadstring == NULL
|
255
|
3749 || STRCMP(bw->bw_leadstring,
|
|
3750 nbw->bw_leadstring) == 0)
|
236
|
3751 && (bw->bw_addstring == NULL
|
255
|
3752 || STRCMP(bw->bw_addstring,
|
|
3753 nbw->bw_addstring) == 0)
|
236
|
3754 && same_affixes(bw, nbw)
|
|
3755 )
|
|
3756 {
|
|
3757 /* Match, combine regions and delete "nbw". */
|
|
3758 pbw->bw_next = nbw->bw_next;
|
|
3759 bw->bw_region |= nbw->bw_region;
|
|
3760 free_basicword(nbw);
|
|
3761 }
|
|
3762 else
|
|
3763 /* No match, continue with next one. */
|
|
3764 pbw = nbw;
|
|
3765 }
|
|
3766 }
|
|
3767 }
|
|
3768 }
|
|
3769 }
|
|
3770
|
|
3771 /*
|
|
3772 * Return TRUE when the prefixes and suffixes for "bw" and "nbw" are equal.
|
|
3773 */
|
|
3774 static int
|
|
3775 same_affixes(bw, nbw)
|
|
3776 basicword_T *bw;
|
|
3777 basicword_T *nbw;
|
|
3778 {
|
|
3779 return (bw->bw_prefix.ga_len == nbw->bw_prefix.ga_len
|
|
3780 && bw->bw_suffix.ga_len == nbw->bw_suffix.ga_len
|
|
3781 && (bw->bw_prefix.ga_len == 0
|
|
3782 || vim_memcmp(bw->bw_prefix.ga_data,
|
|
3783 nbw->bw_prefix.ga_data,
|
|
3784 bw->bw_prefix.ga_len * sizeof(short_u)) == 0)
|
|
3785 && (bw->bw_suffix.ga_len == 0
|
|
3786 || vim_memcmp(bw->bw_suffix.ga_data,
|
|
3787 nbw->bw_suffix.ga_data,
|
|
3788 bw->bw_suffix.ga_len * sizeof(short_u)) == 0));
|
|
3789 }
|
|
3790
|
|
3791 /*
|
243
|
3792 * For each basic word with additions turn the suffixes into other additions
|
|
3793 * and/or new basic words. For each basic word with a leadstring turn the
|
|
3794 * prefixes into other leadstrings and/or new basic words.
|
|
3795 * The result is that no affixes apply to the additions or leadstring of a
|
|
3796 * word.
|
|
3797 * This is also needed when a word with an addition has a prefix and the word
|
|
3798 * with prefix also exists. E.g., "blurp's/D" (D is prefix "de") and
|
|
3799 * "deblurp". "deblurp" would match and no prefix would be tried.
|
255
|
3800 *
|
|
3801 * Returns FAIL when out of memory.
|
236
|
3802 */
|
255
|
3803 static int
|
236
|
3804 expand_affixes(newwords, prefgap, suffgap)
|
|
3805 hashtab_T *newwords;
|
|
3806 garray_T *prefgap;
|
|
3807 garray_T *suffgap;
|
|
3808 {
|
|
3809 int todo;
|
|
3810 hashitem_T *hi;
|
|
3811 basicword_T *bw;
|
|
3812 int pi, si;
|
|
3813 affentry_T *pae, *sae;
|
|
3814 garray_T add_words;
|
|
3815 int n;
|
240
|
3816 char_u message[MAXLINELEN + MAXWLEN];
|
255
|
3817 int retval = OK;
|
236
|
3818
|
|
3819 ga_init2(&add_words, sizeof(basicword_T *), 10);
|
|
3820
|
|
3821 todo = newwords->ht_used;
|
|
3822 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
3823 {
|
|
3824 if (!HASHITEM_EMPTY(hi))
|
|
3825 {
|
|
3826 --todo;
|
240
|
3827
|
|
3828 /* This takes time, print a message now and then. */
|
|
3829 if ((todo & 0x3ff) == 0 || todo == (int)newwords->ht_used - 1)
|
|
3830 {
|
|
3831 sprintf((char *)message, _("%6d todo - %s"),
|
|
3832 todo, HI2BW(hi)->bw_word);
|
|
3833 msg_start();
|
|
3834 msg_outtrans_attr(message, 0);
|
|
3835 msg_clr_eos();
|
|
3836 msg_didout = FALSE;
|
|
3837 msg_col = 0;
|
|
3838 out_flush();
|
|
3839 ui_breakcheck();
|
|
3840 if (got_int)
|
|
3841 break;
|
|
3842 }
|
|
3843
|
236
|
3844 for (bw = HI2BW(hi); bw != NULL; bw = bw->bw_next)
|
|
3845 {
|
|
3846 /*
|
|
3847 * Need to fix affixes if there is a leader or addition and
|
|
3848 * there are prefixes or suffixes.
|
|
3849 */
|
|
3850 if ((bw->bw_leadstring != NULL || bw->bw_addstring != NULL)
|
|
3851 && (bw->bw_prefix.ga_len != 0
|
|
3852 || bw->bw_suffix.ga_len != 0))
|
|
3853 {
|
|
3854 /* Loop over all prefix numbers, but first without a
|
|
3855 * prefix. */
|
|
3856 for (pi = -1; pi < bw->bw_prefix.ga_len; ++pi)
|
|
3857 {
|
|
3858 pae = NULL;
|
|
3859 if (pi >= 0)
|
|
3860 {
|
|
3861 n = ((short_u *)bw->bw_prefix.ga_data)[pi];
|
|
3862 pae = ((affheader_T *)prefgap->ga_data + n)
|
|
3863 ->ah_first;
|
|
3864 }
|
|
3865
|
|
3866 /* Loop over all entries for prefix "pi". Do it once
|
|
3867 * when there is no prefix (pi == -1). */
|
|
3868 do
|
|
3869 {
|
243
|
3870 /* Skip prewords, they don't need to be expanded. */
|
|
3871 if (pae == NULL || !pae->ae_preword)
|
|
3872 {
|
236
|
3873 /* Loop over all suffix numbers. Do without a
|
|
3874 * suffix first when there is a prefix. */
|
|
3875 for (si = (pi == -1 ? 0 : -1);
|
|
3876 si < bw->bw_suffix.ga_len; ++si)
|
|
3877 {
|
|
3878 sae = NULL;
|
|
3879 if (si >= 0)
|
|
3880 {
|
|
3881 n = ((short_u *)bw->bw_suffix.ga_data)[si];
|
|
3882 sae = ((affheader_T *)suffgap->ga_data + n)
|
|
3883 ->ah_first;
|
|
3884 }
|
|
3885
|
|
3886 /* Loop over all entries for suffix "si". Do
|
|
3887 * it once when there is no suffix (si == -1).
|
|
3888 */
|
|
3889 do
|
|
3890 {
|
|
3891 /* Expand the word for this combination of
|
|
3892 * prefixes and affixes. */
|
255
|
3893 if (expand_one_aff(bw, &add_words,
|
|
3894 pae, sae) == FAIL)
|
|
3895 {
|
|
3896 retval = FAIL;
|
|
3897 goto theend;
|
|
3898 }
|
236
|
3899
|
|
3900 /* Advance to next suffix entry, if there
|
|
3901 * is one. */
|
|
3902 if (sae != NULL)
|
|
3903 sae = sae->ae_next;
|
|
3904 } while (sae != NULL);
|
|
3905 }
|
243
|
3906 }
|
|
3907
|
|
3908 /* Advance to next prefix entry, if there is one. */
|
|
3909 if (pae != NULL)
|
|
3910 pae = pae->ae_next;
|
236
|
3911 } while (pae != NULL);
|
|
3912 }
|
|
3913 }
|
|
3914 }
|
|
3915 }
|
|
3916 }
|
|
3917
|
|
3918 /*
|
|
3919 * Add the new words afterwards, can't change "newwords" while going over
|
|
3920 * all its items.
|
|
3921 */
|
|
3922 for (pi = 0; pi < add_words.ga_len; ++pi)
|
255
|
3923 {
|
|
3924 retval = add_to_wordlist(newwords,
|
|
3925 ((basicword_T **)add_words.ga_data)[pi]);
|
|
3926 if (retval == FAIL)
|
|
3927 break;
|
|
3928 }
|
|
3929
|
|
3930 theend:
|
236
|
3931 ga_clear(&add_words);
|
255
|
3932 return retval;
|
236
|
3933 }
|
|
3934
|
|
3935 /*
|
|
3936 * Add one word to "add_words" for basic word "bw" with additions, adding
|
|
3937 * prefix "pae" and suffix "sae". Either "pae" or "sae" can be NULL.
|
243
|
3938 * Don't do this when not necessary:
|
|
3939 * - no leadstring and adding prefix doesn't result in existing word.
|
255
|
3940 * Returns FAIL when out of memory.
|
236
|
3941 */
|
255
|
3942 static int
|
236
|
3943 expand_one_aff(bw, add_words, pae, sae)
|
|
3944 basicword_T *bw;
|
|
3945 garray_T *add_words;
|
|
3946 affentry_T *pae;
|
|
3947 affentry_T *sae;
|
|
3948 {
|
|
3949 char_u word[MAXWLEN + 1];
|
|
3950 char_u caseword[MAXWLEN + 1];
|
|
3951 int l = 0;
|
|
3952 int choplen = 0;
|
|
3953 int ll;
|
|
3954 basicword_T *nbw;
|
|
3955
|
|
3956 /* Prepend prefix to the basic word if there is a prefix and there is no
|
|
3957 * leadstring. */
|
|
3958 if (pae != NULL && bw->bw_leadstring == NULL)
|
|
3959 {
|
|
3960 if (pae->ae_add != NULL)
|
|
3961 {
|
|
3962 l = STRLEN(pae->ae_add);
|
|
3963 mch_memmove(word, pae->ae_add, l);
|
|
3964 }
|
|
3965 if (pae->ae_chop != NULL)
|
|
3966 choplen = STRLEN(pae->ae_chop);
|
|
3967 }
|
|
3968
|
|
3969 /* Copy the body of the word. */
|
|
3970 STRCPY(word + l, bw->bw_word + choplen);
|
|
3971
|
|
3972 /* Do the same for bw_caseword, if it's there. */
|
255
|
3973 if ((bw->bw_flags & BWF_KEEPCAP) && bw->bw_caseword != NULL)
|
236
|
3974 {
|
|
3975 if (l > 0)
|
|
3976 mch_memmove(caseword, pae->ae_add, l);
|
|
3977 STRCPY(caseword + l, bw->bw_caseword + choplen);
|
|
3978 }
|
|
3979
|
|
3980 /* Append suffix to the basic word if there is a suffix and there is no
|
|
3981 * addstring. */
|
|
3982 if (sae != 0 && bw->bw_addstring == NULL)
|
|
3983 {
|
|
3984 l = STRLEN(word);
|
|
3985 if (sae->ae_chop != NULL)
|
|
3986 l -= STRLEN(sae->ae_chop);
|
|
3987 if (sae->ae_add == NULL)
|
|
3988 word[l] = NUL;
|
|
3989 else
|
|
3990 STRCPY(word + l, sae->ae_add);
|
|
3991
|
|
3992 if (bw->bw_flags & BWF_KEEPCAP)
|
|
3993 {
|
|
3994 /* Do the same for the caseword. */
|
|
3995 l = STRLEN(caseword);
|
|
3996 if (sae->ae_chop != NULL)
|
|
3997 l -= STRLEN(sae->ae_chop);
|
|
3998 if (sae->ae_add == NULL)
|
|
3999 caseword[l] = NUL;
|
|
4000 else
|
|
4001 STRCPY(caseword + l, sae->ae_add);
|
|
4002 }
|
|
4003 }
|
|
4004
|
|
4005 nbw = (basicword_T *)alloc_clear((unsigned)
|
|
4006 sizeof(basicword_T) + STRLEN(word));
|
255
|
4007 if (nbw == NULL)
|
|
4008 return FAIL;
|
|
4009
|
|
4010 /* Add the new word to the list of words to be added later. */
|
|
4011 if (ga_grow(add_words, 1) == FAIL)
|
236
|
4012 {
|
255
|
4013 vim_free(nbw);
|
|
4014 return FAIL;
|
|
4015 }
|
|
4016 ((basicword_T **)add_words->ga_data)[add_words->ga_len++] = nbw;
|
|
4017
|
|
4018 /* Copy the (modified) basic word, flags and region. */
|
|
4019 STRCPY(nbw->bw_word, word);
|
|
4020 nbw->bw_flags = bw->bw_flags;
|
|
4021 nbw->bw_region = bw->bw_region;
|
|
4022
|
|
4023 /* Set the (modified) caseword. */
|
|
4024 if (bw->bw_flags & BWF_KEEPCAP)
|
|
4025 nbw->bw_caseword = vim_strsave(caseword);
|
|
4026 else
|
|
4027 nbw->bw_caseword = NULL;
|
|
4028
|
|
4029 if (bw->bw_leadstring != NULL)
|
|
4030 {
|
|
4031 if (pae != NULL)
|
236
|
4032 {
|
255
|
4033 /* Prepend prefix to leadstring. */
|
|
4034 ll = STRLEN(bw->bw_leadstring);
|
|
4035 l = choplen = 0;
|
|
4036 if (pae->ae_add != NULL)
|
|
4037 l = STRLEN(pae->ae_add);
|
|
4038 if (pae->ae_chop != NULL)
|
236
|
4039 {
|
255
|
4040 choplen = STRLEN(pae->ae_chop);
|
|
4041 if (choplen > ll) /* TODO: error? */
|
|
4042 choplen = ll;
|
236
|
4043 }
|
255
|
4044 nbw->bw_leadstring = alloc((unsigned)(ll + l - choplen + 1));
|
|
4045 if (nbw->bw_leadstring != NULL)
|
243
|
4046 {
|
255
|
4047 if (l > 0)
|
|
4048 mch_memmove(nbw->bw_leadstring, pae->ae_add, l);
|
|
4049 STRCPY(nbw->bw_leadstring + l, bw->bw_leadstring + choplen);
|
243
|
4050 }
|
|
4051 }
|
255
|
4052 else
|
|
4053 nbw->bw_leadstring = vim_strsave(bw->bw_leadstring);
|
|
4054 }
|
|
4055 else if (bw->bw_prefix.ga_len > 0)
|
|
4056 {
|
|
4057 /* There is no leadstring, copy the list of possible prefixes. */
|
|
4058 ga_init2(&nbw->bw_prefix, sizeof(short_u), 1);
|
|
4059 if (ga_grow(&nbw->bw_prefix, bw->bw_prefix.ga_len) == OK)
|
236
|
4060 {
|
255
|
4061 mch_memmove(nbw->bw_prefix.ga_data, bw->bw_prefix.ga_data,
|
|
4062 bw->bw_prefix.ga_len * sizeof(short_u));
|
|
4063 nbw->bw_prefix.ga_len = bw->bw_prefix.ga_len;
|
236
|
4064 }
|
|
4065 }
|
255
|
4066
|
|
4067 if (bw->bw_addstring != NULL)
|
|
4068 {
|
|
4069 if (sae != NULL)
|
|
4070 {
|
|
4071 /* Append suffix to addstring. */
|
|
4072 l = STRLEN(bw->bw_addstring);
|
|
4073 if (sae->ae_chop != NULL)
|
|
4074 {
|
|
4075 l -= STRLEN(sae->ae_chop);
|
|
4076 if (l < 0) /* TODO: error? */
|
|
4077 l = 0;
|
|
4078 }
|
|
4079 if (sae->ae_add == NULL)
|
|
4080 ll = 0;
|
|
4081 else
|
|
4082 ll = STRLEN(sae->ae_add);
|
|
4083 nbw->bw_addstring = alloc((unsigned)(ll + l - choplen + 1));
|
|
4084 if (nbw->bw_addstring != NULL)
|
|
4085 {
|
|
4086 STRCPY(nbw->bw_addstring, bw->bw_addstring);
|
|
4087 if (sae->ae_add == NULL)
|
|
4088 nbw->bw_addstring[l] = NUL;
|
|
4089 else
|
|
4090 STRCPY(nbw->bw_addstring + l, sae->ae_add);
|
|
4091 }
|
|
4092 }
|
|
4093 else
|
|
4094 nbw->bw_addstring = vim_strsave(bw->bw_addstring);
|
|
4095 }
|
|
4096
|
|
4097 return OK;
|
236
|
4098 }
|
|
4099
|
|
4100 /*
|
|
4101 * Add basicword_T "*bw" to wordlist "newwords".
|
|
4102 */
|
255
|
4103 static int
|
236
|
4104 add_to_wordlist(newwords, bw)
|
|
4105 hashtab_T *newwords;
|
|
4106 basicword_T *bw;
|
|
4107 {
|
|
4108 hashitem_T *hi;
|
|
4109 basicword_T *bw2;
|
255
|
4110 int retval = OK;
|
236
|
4111
|
|
4112 hi = hash_find(newwords, bw->bw_word);
|
|
4113 if (HASHITEM_EMPTY(hi))
|
|
4114 {
|
|
4115 /* New entry, add to hashlist. */
|
255
|
4116 retval = hash_add(newwords, bw->bw_word);
|
236
|
4117 bw->bw_next = NULL;
|
|
4118 }
|
|
4119 else
|
|
4120 {
|
|
4121 /* Existing entry, append to list of basic words. */
|
|
4122 bw2 = HI2BW(hi);
|
|
4123 bw->bw_next = bw2->bw_next;
|
|
4124 bw2->bw_next = bw;
|
|
4125 }
|
255
|
4126 return retval;
|
236
|
4127 }
|
|
4128
|
|
4129 /*
|
|
4130 * Write a number to file "fd", MSB first, in "len" bytes.
|
|
4131 */
|
255
|
4132 void
|
236
|
4133 put_bytes(fd, nr, len)
|
|
4134 FILE *fd;
|
|
4135 long_u nr;
|
|
4136 int len;
|
|
4137 {
|
|
4138 int i;
|
|
4139
|
|
4140 for (i = len - 1; i >= 0; --i)
|
|
4141 putc((int)(nr >> (i * 8)), fd);
|
|
4142 }
|
|
4143
|
|
4144 /*
|
243
|
4145 * Write affix info. <affitemcnt> <affitem> ...
|
236
|
4146 */
|
|
4147 static void
|
|
4148 write_affix(fd, ah)
|
|
4149 FILE *fd;
|
|
4150 affheader_T *ah;
|
|
4151 {
|
|
4152 int i = 0;
|
|
4153 affentry_T *ae;
|
|
4154 char_u *p;
|
|
4155 int round;
|
243
|
4156 int flags;
|
236
|
4157
|
|
4158 /* Count the number of entries. */
|
|
4159 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
|
4160 ++i;
|
|
4161 put_bytes(fd, (long_u)i, 2); /* <affitemcnt> */
|
|
4162
|
243
|
4163 /* <affitem>: <affflags> <affchoplen> <affchop> <affaddlen> <affadd> */
|
236
|
4164 for (ae = ah->ah_first; ae != NULL; ae = ae->ae_next)
|
243
|
4165 {
|
|
4166 flags = ah->ah_combine ? AFF_COMBINE : 0;
|
|
4167 if (ae->ae_preword)
|
|
4168 flags |= AFF_PREWORD;
|
|
4169 fputc(flags, fd); /* <affflags> */
|
|
4170
|
236
|
4171 for (round = 1; round <= 2; ++round)
|
|
4172 {
|
|
4173 p = round == 1 ? ae->ae_chop : ae->ae_add;
|
|
4174 if (p == NULL)
|
|
4175 putc(0, fd); /* <affchoplen> / <affaddlen> */
|
|
4176 else
|
|
4177 {
|
|
4178 putc(STRLEN(p), fd); /* <affchoplen> / <affaddlen> */
|
|
4179 /* <affchop> / <affadd> */
|
|
4180 fwrite(p, STRLEN(p), (size_t)1, fd);
|
|
4181 }
|
|
4182 }
|
243
|
4183 }
|
236
|
4184 }
|
|
4185
|
|
4186 /*
|
|
4187 * Write list of affix NRs: <affixcnt> <affixNR> ...
|
|
4188 */
|
|
4189 static void
|
|
4190 write_affixlist(fd, aff, bytes)
|
|
4191 FILE *fd;
|
|
4192 garray_T *aff;
|
|
4193 int bytes;
|
|
4194 {
|
|
4195 int i;
|
|
4196
|
|
4197 if (aff->ga_len > 0)
|
|
4198 {
|
|
4199 putc(aff->ga_len, fd); /* <affixcnt> */
|
|
4200 for (i = 0; i < aff->ga_len; ++i)
|
|
4201 put_bytes(fd, (long_u )((short_u *)aff->ga_data)[i], bytes);
|
|
4202 }
|
|
4203 }
|
|
4204
|
|
4205 /*
|
|
4206 * Vim spell file format: <HEADER> <PREFIXLIST> <SUFFIXLIST>
|
|
4207 * <SUGGEST> <WORDLIST>
|
|
4208 *
|
|
4209 * <HEADER>: <fileID> <regioncnt> <regionname> ...
|
255
|
4210 * <charflagslen> <charflags> <fcharslen> <fchars>
|
236
|
4211 *
|
255
|
4212 * <fileID> 10 bytes "VIMspell04"
|
236
|
4213 * <regioncnt> 1 byte number of regions following (8 supported)
|
|
4214 * <regionname> 2 bytes Region name: ca, au, etc.
|
|
4215 * First <regionname> is region 1.
|
|
4216 *
|
255
|
4217 * <charflagslen> 1 byte Number of bytes in <charflags> (should be 128).
|
|
4218 * <charflags> N bytes List of flags (first one is for character 128):
|
|
4219 * 0x01 word character
|
|
4220 * 0x01 upper-case character
|
|
4221 * <fcharslen> 2 bytes Number of bytes in <fchars>.
|
|
4222 * <fchars> N bytes Folded characters, first one is for character 128.
|
236
|
4223 *
|
255
|
4224 *
|
|
4225 * <PREFIXLIST>: <affcount> <affix> ...
|
|
4226 * <SUFFIXLIST>: <affcount> <affix> ...
|
236
|
4227 * list of possible affixes: prefixes and suffixes.
|
|
4228 *
|
|
4229 * <affcount> 2 bytes Number of affixes (MSB comes first).
|
|
4230 * When more than 256 an affixNR is 2 bytes.
|
|
4231 * This is separate for prefixes and suffixes!
|
|
4232 * First affixNR is 0.
|
|
4233 *
|
243
|
4234 * <affix>: <affitemcnt> <affitem> ...
|
236
|
4235 *
|
|
4236 * <affitemcnt> 2 bytes Number of affixes with this affixNR (MSB first).
|
|
4237 *
|
243
|
4238 * <affitem>: <affflags> <affchoplen> <affchop> <affaddlen> <affadd>
|
236
|
4239 *
|
243
|
4240 * <affflags> 1 byte 0x01: prefix combines with suffix, AFF_COMBINE
|
|
4241 * 0x02: prefix includes word, AFF_PREWORD
|
|
4242 * 0x04-0x80: unset
|
236
|
4243 * <affchoplen> 1 byte Length of <affchop> in bytes.
|
|
4244 * <affchop> N bytes To be removed from basic word.
|
|
4245 * <affaddlen> 1 byte Length of <affadd> in bytes.
|
|
4246 * <affadd> N bytes To be added to basic word.
|
|
4247 *
|
|
4248 *
|
|
4249 * <SUGGEST> : <suggestlen> <more> ...
|
|
4250 *
|
|
4251 * <suggestlen> 4 bytes Length of <SUGGEST> in bytes, excluding
|
|
4252 * <suggestlen>. MSB first.
|
|
4253 * <more> To be defined.
|
|
4254 *
|
|
4255 *
|
|
4256 * <WORDLIST>: <wordcount> <worditem> ...
|
|
4257 *
|
|
4258 * <wordcount> 4 bytes Number of <worditem> following. MSB first.
|
|
4259 *
|
|
4260 * <worditem>: <nr> <string> <flags> [<flags2>]
|
|
4261 * [<caselen> <caseword>]
|
|
4262 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
4263 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
4264 * [<region>]
|
|
4265 * [<addcnt> <add> ...]
|
|
4266 *
|
|
4267 * <nr> i 1 byte Number of bytes copied from previous word.
|
|
4268 * <string> N bytes Additional bytes for word, up to byte smaller than
|
|
4269 * 0x20 (space).
|
|
4270 * Must only contain case-folded word characters.
|
243
|
4271 * <flags> 1 byte 0x01: word is valid without addition, BWF_VALID
|
|
4272 * 0x02: has region byte, BWF_REGION
|
|
4273 * 0x04: first letter must be upper-case, BWF_ONECAP
|
236
|
4274 * 0x08: has suffixes, <affixcnt> and <affixNR> follow
|
243
|
4275 * BWF_SUFFIX
|
|
4276 * 0x10: more flags, <flags2> follows next, BWF_SECOND
|
236
|
4277 * 0x20-0x80: can't be used, unset
|
243
|
4278 * <flags2> 1 byte 0x01: has additions, <addcnt> and <add> follow,
|
|
4279 * BWF_ADDS
|
236
|
4280 * 0x02: has prefixes, <affixcnt> and <affixNR> follow
|
243
|
4281 * BWF_PREFIX
|
|
4282 * 0x04: all letters must be upper-case, BWF_ALLCAP
|
|
4283 * 0x08: case must match, BWF_KEEPCAP
|
249
|
4284 * 0x10: has more than 255 additions, <addcnt> is two
|
|
4285 * bytes, BWF_ADDS_M
|
236
|
4286 * 0x10-0x80: unset
|
|
4287 * <caselen> 1 byte Length of <caseword>.
|
|
4288 * <caseword> N bytes Word with matching case.
|
|
4289 * <affixcnt> 1 byte Number of affix NRs following.
|
|
4290 * <affixNR> 1 or 2 byte Number of possible affix for this word.
|
|
4291 * When using 2 bytes MSB comes first.
|
|
4292 * <region> 1 byte Bitmask for regions in which word is valid. When
|
|
4293 * omitted it's valid in all regions.
|
|
4294 * Lowest bit is for region 1.
|
249
|
4295 * <addcnt> 1 or 2 byte Number of <add> items following.
|
236
|
4296 *
|
249
|
4297 * <add>: <addflags> <addlen> [<leadlen>] [<copylen>] [<addstring>] [<region>]
|
236
|
4298 *
|
243
|
4299 * <addflags> 1 byte 0x01: unset
|
|
4300 * 0x02: has region byte, ADD_REGION
|
|
4301 * 0x04: first letter must be upper-case, ADD_ONECAP
|
249
|
4302 * 0x08: unset
|
|
4303 * 0x10: has a <leadlen>, ADD_LEADLEN
|
|
4304 * 0x20: has a <copylen>, ADD_COPYLEN
|
243
|
4305 * 0x40: all letters must be upper-case, ADD_ALLCAP
|
|
4306 * 0x80: fixed case, <addstring> is the whole word
|
|
4307 * with matching case, ADD_KEEPCAP.
|
236
|
4308 * <addlen> 1 byte Length of <addstring> in bytes.
|
|
4309 * <leadlen> 1 byte Number of bytes at start of <addstring> that must
|
|
4310 * come before the start of the basic word.
|
249
|
4311 * <copylen> 1 byte Number of bytes copied from previous <addstring>.
|
236
|
4312 * <addstring> N bytes Word characters, before/in/after the word.
|
|
4313 *
|
|
4314 * All text characters are in 'encoding': <affchop>, <affadd>, <string>,
|
|
4315 * <caseword>> and <addstring>.
|
|
4316 * All other fields are ASCII: <regionname>
|
|
4317 * <string> is always case-folded.
|
|
4318 */
|
|
4319
|
|
4320 /*
|
|
4321 * Write the Vim spell file "fname".
|
|
4322 */
|
|
4323 static void
|
260
|
4324 write_vim_spell(fname, prefga, suffga, newwords, regcount, regchars, ascii)
|
236
|
4325 char_u *fname;
|
|
4326 garray_T *prefga; /* prefixes, affheader_T entries */
|
|
4327 garray_T *suffga; /* suffixes, affheader_T entries */
|
|
4328 hashtab_T *newwords; /* basic words, basicword_T entries */
|
|
4329 int regcount; /* number of regions */
|
|
4330 char_u *regchars; /* region names */
|
260
|
4331 int ascii; /* TRUE for ascii spell file */
|
236
|
4332 {
|
249
|
4333 winfo_T wif;
|
236
|
4334 garray_T *gap;
|
|
4335 hashitem_T *hi;
|
|
4336 char_u **wtab;
|
|
4337 int todo;
|
|
4338 int flags, aflags;
|
249
|
4339 basicword_T *bw, *bwf, *bw2 = NULL;
|
236
|
4340 int i;
|
|
4341 int round;
|
|
4342 garray_T bwga;
|
|
4343
|
249
|
4344 vim_memset(&wif, 0, sizeof(winfo_T));
|
|
4345
|
|
4346 wif.wif_fd = fopen((char *)fname, "w");
|
|
4347 if (wif.wif_fd == NULL)
|
236
|
4348 {
|
|
4349 EMSG2(_(e_notopen), fname);
|
|
4350 return;
|
|
4351 }
|
|
4352
|
255
|
4353 /* <HEADER>: <fileID> <regioncnt> <regionname> ...
|
|
4354 * <charflagslen> <charflags> <fcharslen> <fchars> */
|
|
4355 fwrite(VIMSPELLMAGIC, VIMSPELLMAGICL, (size_t)1, wif.wif_fd); /* <fileID> */
|
236
|
4356
|
|
4357 /* write the region names if there is more than one */
|
|
4358 if (regcount > 1)
|
|
4359 {
|
255
|
4360 putc(regcount, wif.wif_fd); /* <regioncnt> <regionname> ... */
|
249
|
4361 fwrite(regchars, (size_t)(regcount * 2), (size_t)1, wif.wif_fd);
|
|
4362 wif.wif_regionmask = (1 << regcount) - 1;
|
236
|
4363 }
|
|
4364 else
|
|
4365 {
|
249
|
4366 putc(0, wif.wif_fd);
|
|
4367 wif.wif_regionmask = 0;
|
236
|
4368 }
|
|
4369
|
255
|
4370 /* Write the table with character flags and table for case folding.
|
260
|
4371 * <charflagslen> <charflags> <fcharlen> <fchars>
|
|
4372 * Skip this for ASCII, the table may conflict with the one used for
|
|
4373 * 'encoding'. */
|
|
4374 if (ascii)
|
|
4375 {
|
|
4376 putc(0, wif.wif_fd);
|
|
4377 putc(0, wif.wif_fd);
|
|
4378 putc(0, wif.wif_fd);
|
|
4379 }
|
|
4380 else
|
|
4381 write_spell_chartab(wif.wif_fd);
|
255
|
4382
|
|
4383 /* <PREFIXLIST>: <affcount> <affix> ...
|
|
4384 * <SUFFIXLIST>: <affcount> <affix> ... */
|
236
|
4385 for (round = 1; round <= 2; ++round)
|
|
4386 {
|
|
4387 gap = round == 1 ? prefga : suffga;
|
249
|
4388 put_bytes(wif.wif_fd, (long_u)gap->ga_len, 2); /* <affcount> */
|
236
|
4389
|
|
4390 for (i = 0; i < gap->ga_len; ++i)
|
249
|
4391 write_affix(wif.wif_fd, (affheader_T *)gap->ga_data + i);
|
236
|
4392 }
|
|
4393
|
|
4394 /* Number of bytes used for affix NR depends on affix count. */
|
249
|
4395 wif.wif_prefm = (prefga->ga_len > 256) ? 2 : 1;
|
|
4396 wif.wif_suffm = (suffga->ga_len > 256) ? 2 : 1;
|
236
|
4397
|
255
|
4398 /* <SUGGEST> : <suggestlen> <more> ...
|
|
4399 * TODO. Only write a zero length for now. */
|
|
4400 put_bytes(wif.wif_fd, 0L, 4); /* <suggestlen> */
|
236
|
4401
|
|
4402 /*
|
255
|
4403 * <WORDLIST>: <wordcount> <worditem> ...
|
236
|
4404 */
|
255
|
4405
|
236
|
4406 /* number of basic words in 4 bytes */
|
249
|
4407 put_bytes(wif.wif_fd, newwords->ht_used, 4); /* <wordcount> */
|
236
|
4408
|
|
4409 /*
|
249
|
4410 * Sort the word list, so that we can copy as many bytes as possible from
|
|
4411 * the previous word.
|
236
|
4412 */
|
|
4413 wtab = (char_u **)alloc((unsigned)(sizeof(char_u *) * newwords->ht_used));
|
|
4414 if (wtab != NULL)
|
|
4415 {
|
|
4416 /* Make a table with pointers to each word. */
|
|
4417 todo = newwords->ht_used;
|
|
4418 for (hi = newwords->ht_array; todo > 0; ++hi)
|
|
4419 if (!HASHITEM_EMPTY(hi))
|
|
4420 wtab[--todo] = hi->hi_key;
|
|
4421
|
|
4422 /* Sort. */
|
|
4423 sort_strings(wtab, (int)newwords->ht_used);
|
|
4424
|
|
4425 /* Now write each basic word to the spell file. */
|
|
4426 ga_init2(&bwga, sizeof(basicword_T *), 10);
|
240
|
4427 for (todo = 0; (long_u)todo < newwords->ht_used; ++todo)
|
236
|
4428 {
|
|
4429 bwf = KEY2BW(wtab[todo]);
|
|
4430
|
|
4431 /*
|
|
4432 * Reorder the list of basicword_T words: make a list for words
|
|
4433 * with the same case-folded word. Put them together for same
|
|
4434 * caps (ONECAP, ALLCAP and various KEEPCAP words) and same
|
|
4435 * affixes. Each list will then be put as a basic word with
|
|
4436 * additions.
|
|
4437 * This won't take much space, since the basic word is the same
|
|
4438 * every time, only its length is written.
|
|
4439 */
|
|
4440 bwga.ga_len = 0;
|
|
4441 for (bw = bwf; bw != NULL; bw = bw->bw_next)
|
|
4442 {
|
|
4443 flags = bw->bw_flags & (BWF_ONECAP | BWF_KEEPCAP | BWF_ALLCAP);
|
|
4444
|
|
4445 /* Go through the lists we found so far. Break when the case
|
|
4446 * matches. */
|
|
4447 for (i = 0; i < bwga.ga_len; ++i)
|
|
4448 {
|
|
4449 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4450 aflags = bw2->bw_flags & (BWF_ONECAP | BWF_KEEPCAP
|
|
4451 | BWF_ALLCAP);
|
|
4452 if (flags == aflags
|
|
4453 && ((flags & BWF_KEEPCAP) == 0
|
255
|
4454 || bw->bw_caseword == NULL
|
|
4455 || bw2->bw_caseword == NULL
|
|
4456 || STRCMP(bw->bw_caseword,
|
|
4457 bw2->bw_caseword) == 0)
|
236
|
4458 && same_affixes(bw, bw2))
|
|
4459 break;
|
|
4460 }
|
|
4461 if (i == bwga.ga_len)
|
|
4462 {
|
|
4463 /* No word with similar caps, make a new list. */
|
|
4464 if (ga_grow(&bwga, 1) == FAIL)
|
|
4465 break;
|
|
4466 ((basicword_T **)bwga.ga_data)[i] = bw;
|
|
4467 bw->bw_cnext = NULL;
|
|
4468 ++bwga.ga_len;
|
|
4469 }
|
|
4470 else
|
|
4471 {
|
|
4472 /* Add to list of words with similar caps. */
|
|
4473 bw->bw_cnext = bw2->bw_cnext;
|
|
4474 bw2->bw_cnext = bw;
|
|
4475 }
|
|
4476 }
|
|
4477
|
|
4478 /* Prefer the word with no caps to use as the first basic word.
|
|
4479 * At least one without KEEPCAP. */
|
|
4480 bw = NULL;
|
|
4481 for (i = 0; i < bwga.ga_len; ++i)
|
|
4482 {
|
|
4483 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4484 if (bw == NULL
|
|
4485 || (bw2->bw_flags & (BWF_ONECAP | BWF_KEEPCAP
|
|
4486 | BWF_ALLCAP)) == 0
|
|
4487 || (bw->bw_flags & BWF_KEEPCAP))
|
|
4488 bw = bw2;
|
|
4489 }
|
|
4490
|
|
4491 /* Write first basic word. If it's KEEPCAP then we need a word
|
|
4492 * without VALID flag first (makes it easier to read the list back
|
|
4493 * in). */
|
|
4494 if (bw->bw_flags & BWF_KEEPCAP)
|
249
|
4495 write_bword(&wif, bw, TRUE);
|
|
4496 write_bword(&wif, bw, FALSE);
|
236
|
4497
|
|
4498 /* Write other basic words, with different caps. */
|
|
4499 for (i = 0; i < bwga.ga_len; ++i)
|
|
4500 {
|
|
4501 bw2 = ((basicword_T **)bwga.ga_data)[i];
|
|
4502 if (bw2 != bw)
|
249
|
4503 write_bword(&wif, bw2, FALSE);
|
236
|
4504 }
|
|
4505 }
|
|
4506
|
|
4507 ga_clear(&bwga);
|
255
|
4508 vim_free(wtab);
|
236
|
4509 }
|
|
4510
|
249
|
4511 fclose(wif.wif_fd);
|
|
4512
|
|
4513 /* Print a few statistics. */
|
|
4514 if (wif.wif_addmaxw == NULL)
|
|
4515 wif.wif_addmaxw = (char_u *)"";
|
|
4516 smsg((char_u *)_("Maximum number of adds on a word: %ld (%s)"),
|
|
4517 wif.wif_addmax, wif.wif_addmaxw);
|
|
4518 smsg((char_u *)_("Average number of adds on a word: %f"),
|
|
4519 (float)wif.wif_acount / (float)wif.wif_wcount);
|
|
4520 }
|
|
4521
|
|
4522 /*
|
|
4523 * Compare two basic words for their <addstring>.
|
|
4524 */
|
|
4525 static int
|
|
4526 #ifdef __BORLANDC__
|
|
4527 _RTLENTRYF
|
|
4528 #endif
|
|
4529 bw_compare __ARGS((const void *s1, const void *s2));
|
|
4530
|
|
4531 static int
|
|
4532 #ifdef __BORLANDC__
|
|
4533 _RTLENTRYF
|
|
4534 #endif
|
|
4535 bw_compare(s1, s2)
|
|
4536 const void *s1;
|
|
4537 const void *s2;
|
|
4538 {
|
|
4539 basicword_T *bw1 = *(basicword_T **)s1;
|
|
4540 basicword_T *bw2 = *(basicword_T **)s2;
|
|
4541 int i = 0;
|
|
4542
|
|
4543 /* compare the leadstrings */
|
|
4544 if (bw1->bw_leadstring == NULL)
|
|
4545 {
|
|
4546 if (bw2->bw_leadstring != NULL)
|
|
4547 return 1;
|
|
4548 }
|
|
4549 else if (bw2->bw_leadstring == NULL)
|
|
4550 return -1;
|
|
4551 else
|
|
4552 i = STRCMP(bw1->bw_leadstring, bw2->bw_leadstring);
|
|
4553
|
|
4554 if (i == 0)
|
|
4555 {
|
|
4556 /* leadstrings are identical, compare the addstrings */
|
|
4557 if (bw1->bw_addstring == NULL)
|
|
4558 {
|
|
4559 if (bw2->bw_addstring != NULL)
|
|
4560 return 1;
|
|
4561 }
|
|
4562 else if (bw2->bw_addstring == NULL)
|
|
4563 return -1;
|
|
4564 else
|
|
4565 i = STRCMP(bw1->bw_addstring, bw2->bw_addstring);
|
|
4566 }
|
|
4567 return i;
|
236
|
4568 }
|
|
4569
|
|
4570 /*
|
|
4571 * Write basic word, followed by any additions.
|
|
4572 *
|
|
4573 * <worditem>: <nr> <string> <flags> [<flags2>]
|
|
4574 * [<caselen> <caseword>]
|
|
4575 * [<affixcnt> <affixNR> ...] (prefixes)
|
|
4576 * [<affixcnt> <affixNR> ...] (suffixes)
|
|
4577 * [<region>]
|
|
4578 * [<addcnt> <add> ...]
|
|
4579 */
|
|
4580 static void
|
249
|
4581 write_bword(wif, bwf, lowcap)
|
|
4582 winfo_T *wif; /* info for writing */
|
236
|
4583 basicword_T *bwf;
|
|
4584 int lowcap; /* write KEEPKAP word as not-valid */
|
|
4585 {
|
249
|
4586 FILE *fd = wif->wif_fd;
|
236
|
4587 int flags;
|
|
4588 int aflags;
|
|
4589 int len;
|
|
4590 int leadlen, addlen;
|
249
|
4591 int copylen;
|
236
|
4592 int clen;
|
|
4593 int adds = 0;
|
|
4594 int i;
|
249
|
4595 int idx;
|
236
|
4596 basicword_T *bw, *bw2;
|
249
|
4597 basicword_T **wtab;
|
|
4598 int count;
|
|
4599 int l;
|
236
|
4600
|
|
4601 /* Check how many bytes can be copied from the previous word. */
|
|
4602 len = STRLEN(bwf->bw_word);
|
249
|
4603 if (wif->wif_prevbw == NULL)
|
236
|
4604 clen = 0;
|
|
4605 else
|
|
4606 for (clen = 0; clen < len
|
249
|
4607 && wif->wif_prevbw->bw_word[clen] == bwf->bw_word[clen]; ++clen)
|
236
|
4608 ;
|
|
4609 putc(clen, fd); /* <nr> */
|
249
|
4610 wif->wif_prevbw = bwf;
|
236
|
4611 /* <string> */
|
|
4612 if (len > clen)
|
|
4613 fwrite(bwf->bw_word + clen, (size_t)(len - clen), (size_t)1, fd);
|
|
4614
|
|
4615 /* Try to find a word without additions to use first. */
|
|
4616 bw = bwf;
|
|
4617 for (bw2 = bwf; bw2 != NULL; bw2 = bw2->bw_cnext)
|
|
4618 {
|
|
4619 if (bw2->bw_addstring != NULL || bw2->bw_leadstring != NULL)
|
|
4620 ++adds;
|
|
4621 else
|
|
4622 bw = bw2;
|
|
4623 }
|
|
4624
|
|
4625 /* Flags: If there is no leadstring and no addstring the basic word is
|
|
4626 * valid, may have prefixes, suffixes and region. */
|
|
4627 flags = bw->bw_flags;
|
|
4628 if (bw->bw_addstring == NULL && bw->bw_leadstring == NULL)
|
|
4629 {
|
|
4630 flags |= BWF_VALID;
|
|
4631
|
|
4632 /* Flags: add the region byte if the word isn't valid in all
|
|
4633 * regions. */
|
249
|
4634 if (wif->wif_regionmask != 0 && (bw->bw_region & wif->wif_regionmask)
|
|
4635 != wif->wif_regionmask)
|
236
|
4636 flags |= BWF_REGION;
|
|
4637 }
|
243
|
4638 /* Add the prefix/suffix list if there are prefixes/suffixes. */
|
|
4639 if (bw->bw_leadstring == NULL && bw->bw_prefix.ga_len > 0)
|
|
4640 flags |= BWF_PREFIX;
|
|
4641 if (bw->bw_addstring == NULL && bw->bw_suffix.ga_len > 0)
|
|
4642 flags |= BWF_SUFFIX;
|
236
|
4643
|
|
4644 /* Flags: may have additions. */
|
|
4645 if (adds > 0)
|
249
|
4646 {
|
236
|
4647 flags |= BWF_ADDS;
|
249
|
4648 if (adds >= 256)
|
|
4649 flags |= BWF_ADDS_M;
|
|
4650 }
|
236
|
4651
|
|
4652 /* The dummy word before a KEEPCAP word doesn't have any flags, they are
|
|
4653 * in the actual word that follows. */
|
|
4654 if (lowcap)
|
|
4655 flags = 0;
|
|
4656
|
|
4657 /* Flags: when the upper byte is not used we only write one flags
|
|
4658 * byte, if it's used then set an extra flag in the first byte and
|
|
4659 * also write the second byte. */
|
|
4660 if ((flags & 0xff00) == 0)
|
|
4661 putc(flags, fd); /* <flags> */
|
|
4662 else
|
|
4663 {
|
|
4664 putc(flags | BWF_SECOND, fd); /* <flags> */
|
|
4665 putc((int)((unsigned)flags >> 8), fd); /* <flags2> */
|
|
4666 }
|
|
4667
|
|
4668 /* First dummy word doesn't need anything but flags. */
|
|
4669 if (lowcap)
|
|
4670 return;
|
|
4671
|
255
|
4672 if ((flags & BWF_KEEPCAP) && bw->bw_caseword != NULL)
|
236
|
4673 {
|
|
4674 len = STRLEN(bw->bw_caseword);
|
|
4675 putc(len, fd); /* <caselen> */
|
|
4676 for (i = 0; i < len; ++i)
|
|
4677 putc(bw->bw_caseword[i], fd); /* <caseword> */
|
|
4678 }
|
|
4679
|
|
4680 /* write prefix and suffix lists: <affixcnt> <affixNR> ... */
|
|
4681 if (flags & BWF_PREFIX)
|
249
|
4682 write_affixlist(fd, &bw->bw_prefix, wif->wif_prefm);
|
236
|
4683 if (flags & BWF_SUFFIX)
|
249
|
4684 write_affixlist(fd, &bw->bw_suffix, wif->wif_suffm);
|
236
|
4685
|
|
4686 if (flags & BWF_REGION)
|
|
4687 putc(bw->bw_region, fd); /* <region> */
|
|
4688
|
249
|
4689 ++wif->wif_wcount;
|
|
4690
|
236
|
4691 /*
|
|
4692 * Additions.
|
|
4693 */
|
|
4694 if (adds > 0)
|
|
4695 {
|
249
|
4696 if (adds >= 256)
|
|
4697 put_bytes(fd, (long_u)adds, 2); /* 2 byte <addcnt> */
|
|
4698 else
|
|
4699 putc(adds, fd); /* 1 byte <addcnt> */
|
|
4700
|
|
4701 /* statistics */
|
|
4702 wif->wif_acount += adds;
|
|
4703 if (wif->wif_addmax < adds)
|
|
4704 {
|
|
4705 wif->wif_addmax = adds;
|
|
4706 wif->wif_addmaxw = bw->bw_word;
|
|
4707 }
|
|
4708
|
|
4709 /*
|
|
4710 * Sort the list of additions, so that we can copy as many bytes as
|
|
4711 * possible from the previous addstring.
|
|
4712 */
|
|
4713
|
|
4714 /* Make a table with pointers to each basic word that has additions. */
|
|
4715 wtab = (basicword_T **)alloc((unsigned)(sizeof(basicword_T *) * adds));
|
|
4716 if (wtab == NULL)
|
|
4717 return;
|
|
4718 count = 0;
|
236
|
4719 for (bw = bwf; bw != NULL; bw = bw->bw_cnext)
|
|
4720 if (bw->bw_leadstring != NULL || bw->bw_addstring != NULL)
|
249
|
4721 wtab[count++] = bw;
|
|
4722
|
|
4723 /* Sort. */
|
|
4724 qsort((void *)wtab, (size_t)count, sizeof(basicword_T *), bw_compare);
|
|
4725
|
|
4726 /* Now write each basic word to the spell file. Copy bytes from the
|
|
4727 * previous leadstring/addstring if possible. */
|
|
4728 bw2 = NULL;
|
|
4729 for (idx = 0; idx < count; ++idx)
|
|
4730 {
|
|
4731 bw = wtab[idx];
|
|
4732
|
|
4733 /* <add>: <addflags> <addlen> [<leadlen>] [<copylen>]
|
|
4734 * [<addstring>] [<region>] */
|
|
4735 copylen = 0;
|
|
4736 if (bw->bw_leadstring == NULL)
|
|
4737 leadlen = 0;
|
|
4738 else
|
236
|
4739 {
|
249
|
4740 leadlen = STRLEN(bw->bw_leadstring);
|
|
4741 if (bw2 != NULL && bw2->bw_leadstring != NULL)
|
|
4742 for ( ; copylen < leadlen; ++copylen)
|
|
4743 if (bw->bw_leadstring[copylen]
|
|
4744 != bw2->bw_leadstring[copylen])
|
|
4745 break;
|
|
4746 }
|
|
4747 if (bw->bw_addstring == NULL)
|
|
4748 addlen = 0;
|
|
4749 else
|
|
4750 {
|
|
4751 addlen = STRLEN(bw->bw_addstring);
|
|
4752 if (bw2 != NULL && copylen == leadlen
|
|
4753 && bw2->bw_addstring != NULL)
|
|
4754 {
|
|
4755 for (i = 0; i < addlen; ++i)
|
|
4756 if (bw->bw_addstring[i] != bw2->bw_addstring[i])
|
|
4757 break;
|
|
4758 copylen += i;
|
|
4759 }
|
|
4760 }
|
|
4761
|
|
4762 aflags = 0;
|
|
4763 /* Only copy bytes when it's more than one, the length itself
|
|
4764 * takes an extra byte. */
|
|
4765 if (copylen > 1)
|
|
4766 aflags |= ADD_COPYLEN;
|
|
4767 else
|
|
4768 copylen = 0;
|
|
4769
|
|
4770 if (bw->bw_flags & BWF_ONECAP)
|
|
4771 aflags |= ADD_ONECAP;
|
|
4772 if (bw->bw_flags & BWF_ALLCAP)
|
|
4773 aflags |= ADD_ALLCAP;
|
|
4774 if (bw->bw_flags & BWF_KEEPCAP)
|
|
4775 aflags |= ADD_KEEPCAP;
|
|
4776 if (wif->wif_regionmask != 0 && (bw->bw_region
|
|
4777 & wif->wif_regionmask) != wif->wif_regionmask)
|
|
4778 aflags |= ADD_REGION;
|
|
4779 if (leadlen > 0)
|
|
4780 aflags |= ADD_LEADLEN;
|
|
4781 putc(aflags, fd); /* <addflags> */
|
|
4782
|
|
4783 putc(leadlen + addlen, fd); /* <addlen> */
|
|
4784 if (aflags & ADD_LEADLEN)
|
|
4785 putc(leadlen, fd); /* <leadlen> */
|
|
4786 if (aflags & ADD_COPYLEN)
|
|
4787 putc(copylen, fd); /* <copylen> */
|
|
4788
|
|
4789 /* <addstring> */
|
|
4790 if (leadlen > copylen && bw->bw_leadstring != NULL)
|
|
4791 fwrite(bw->bw_leadstring + copylen,
|
|
4792 (size_t)(leadlen - copylen), (size_t)1, fd);
|
|
4793 if (leadlen + addlen > copylen && bw->bw_addstring != NULL)
|
|
4794 {
|
|
4795 if (copylen >= leadlen)
|
|
4796 l = copylen - leadlen;
|
236
|
4797 else
|
249
|
4798 l = 0;
|
|
4799 fwrite(bw->bw_addstring + l,
|
|
4800 (size_t)(addlen - l), (size_t)1, fd);
|
236
|
4801 }
|
249
|
4802
|
|
4803 if (aflags & ADD_REGION)
|
|
4804 putc(bw->bw_region, fd); /* <region> */
|
|
4805
|
|
4806 bw2 = bw;
|
|
4807 }
|
255
|
4808
|
249
|
4809 vim_free(wtab);
|
236
|
4810 }
|
|
4811 }
|
|
4812
|
|
4813
|
|
4814 /*
|
|
4815 * ":mkspell outfile infile ..."
|
|
4816 */
|
|
4817 void
|
|
4818 ex_mkspell(eap)
|
|
4819 exarg_T *eap;
|
|
4820 {
|
|
4821 int fcount;
|
|
4822 char_u **fnames;
|
|
4823 char_u fname[MAXPATHL];
|
|
4824 char_u wfname[MAXPATHL];
|
|
4825 afffile_T *(afile[8]);
|
|
4826 hashtab_T dfile[8];
|
|
4827 int i;
|
|
4828 int len;
|
|
4829 char_u region_name[16];
|
|
4830 struct stat st;
|
|
4831 int round;
|
|
4832 vimconv_T conv;
|
240
|
4833 int ascii = FALSE;
|
|
4834 char_u *arg = eap->arg;
|
255
|
4835 int error = FALSE;
|
240
|
4836
|
|
4837 if (STRNCMP(arg, "-ascii", 6) == 0)
|
|
4838 {
|
|
4839 ascii = TRUE;
|
|
4840 arg = skipwhite(arg + 6);
|
|
4841 }
|
|
4842
|
|
4843 /* Expand all the remaining arguments (e.g., $VIMRUNTIME). */
|
|
4844 if (get_arglist_exp(arg, &fcount, &fnames) == FAIL)
|
236
|
4845 return;
|
|
4846 if (fcount < 2)
|
|
4847 EMSG(_(e_invarg)); /* need at least output and input names */
|
|
4848 else if (fcount > 9)
|
|
4849 EMSG(_("E754: Only up to 8 regions supported"));
|
|
4850 else
|
|
4851 {
|
|
4852 /* Check for overwriting before doing things that may take a lot of
|
|
4853 * time. */
|
272
|
4854 vim_snprintf((char *)wfname, sizeof(wfname), "%s.%s.spl", fnames[0],
|
240
|
4855 ascii ? (char_u *)"ascii" : p_enc);
|
236
|
4856 if (!eap->forceit && mch_stat((char *)wfname, &st) >= 0)
|
|
4857 {
|
|
4858 EMSG(_(e_exists));
|
|
4859 goto theend;
|
|
4860 }
|
|
4861 if (mch_isdir(fnames[0]))
|
|
4862 {
|
|
4863 EMSG2(_(e_isadir2), fnames[0]);
|
|
4864 goto theend;
|
|
4865 }
|
|
4866
|
|
4867 /*
|
|
4868 * Init the aff and dic pointers.
|
|
4869 * Get the region names if there are more than 2 arguments.
|
|
4870 */
|
|
4871 for (i = 1; i < fcount; ++i)
|
|
4872 {
|
|
4873 afile[i - 1] = NULL;
|
|
4874 hash_init(&dfile[i - 1]);
|
|
4875 if (fcount > 2)
|
|
4876 {
|
|
4877 len = STRLEN(fnames[i]);
|
|
4878 if (STRLEN(gettail(fnames[i])) < 5 || fnames[i][len - 3] != '_')
|
|
4879 {
|
|
4880 EMSG2(_("E755: Invalid region in %s"), fnames[i]);
|
|
4881 goto theend;
|
|
4882 }
|
|
4883 else
|
|
4884 {
|
|
4885 region_name[(i - 1) * 2] = TOLOWER_ASC(fnames[i][len - 2]);
|
|
4886 region_name[(i - 1) * 2 + 1] =
|
|
4887 TOLOWER_ASC(fnames[i][len - 1]);
|
|
4888 }
|
|
4889 }
|
|
4890 }
|
|
4891
|
255
|
4892 /* Clear the char type tables, don't want to use any of the currently
|
|
4893 * used spell properties. */
|
|
4894 init_spell_chartab();
|
|
4895
|
236
|
4896 /*
|
|
4897 * Read all the .aff and .dic files.
|
|
4898 * Text is converted to 'encoding'.
|
|
4899 */
|
|
4900 for (i = 1; i < fcount; ++i)
|
|
4901 {
|
|
4902 /* Read the .aff file. Will init "conv" based on the "SET" line. */
|
|
4903 conv.vc_type = CONV_NONE;
|
272
|
4904 vim_snprintf((char *)fname, sizeof(fname), "%s.aff", fnames[i]);
|
240
|
4905 if ((afile[i - 1] = spell_read_aff(fname, &conv, ascii)) == NULL)
|
236
|
4906 break;
|
|
4907
|
|
4908 /* Read the .dic file. */
|
272
|
4909 vim_snprintf((char *)fname, sizeof(fname), "%s.dic", fnames[i]);
|
240
|
4910 if (spell_read_dic(&dfile[i - 1], fname, &conv, ascii) == FAIL)
|
236
|
4911 break;
|
|
4912
|
|
4913 /* Free any conversion stuff. */
|
|
4914 convert_setup(&conv, NULL, NULL);
|
|
4915 }
|
|
4916
|
|
4917 /* Process the data when all the files could be read. */
|
|
4918 if (i == fcount)
|
|
4919 {
|
|
4920 garray_T prefga;
|
|
4921 garray_T suffga;
|
|
4922 garray_T *gap;
|
|
4923 hashtab_T newwords;
|
|
4924
|
|
4925 /*
|
|
4926 * Combine all the affixes into one new affix list. This is done
|
|
4927 * for prefixes and suffixes separately.
|
|
4928 * We need to do this for each region, try to re-use the same
|
|
4929 * affixes.
|
|
4930 * Since we number the new affix entries, a growarray will do. In
|
|
4931 * the affheader_T the ah_key is unused.
|
|
4932 */
|
|
4933 MSG(_("Combining affixes..."));
|
|
4934 out_flush();
|
|
4935 for (round = 1; round <= 2; ++round)
|
|
4936 {
|
|
4937 gap = round == 1 ? &prefga : &suffga;
|
|
4938 ga_init2(gap, sizeof(affheader_T), 50);
|
|
4939 for (i = 1; i < fcount; ++i)
|
|
4940 get_new_aff(round == 1 ? &afile[i - 1]->af_pref
|
240
|
4941 : &afile[i - 1]->af_suff,
|
|
4942 gap, round == 1);
|
236
|
4943 }
|
|
4944
|
|
4945 /*
|
|
4946 * Go over all words and:
|
|
4947 * - change the old affix names to the new affix numbers
|
|
4948 * - check the conditions
|
|
4949 * - fold case
|
|
4950 * - extract the basic word and additions.
|
|
4951 * Do this for each region.
|
|
4952 */
|
|
4953 MSG(_("Building word list..."));
|
|
4954 out_flush();
|
|
4955 hash_init(&newwords);
|
|
4956
|
|
4957 for (i = 1; i < fcount; ++i)
|
|
4958 build_wordlist(&newwords, &dfile[i - 1], afile[i - 1],
|
|
4959 1 << (i - 1));
|
|
4960
|
|
4961 if (fcount > 2)
|
|
4962 {
|
|
4963 /* Combine words for the different regions into one. */
|
|
4964 MSG(_("Combining regions..."));
|
|
4965 out_flush();
|
|
4966 combine_regions(&newwords);
|
|
4967 }
|
|
4968
|
|
4969 /*
|
|
4970 * Affixes on a word with additions are clumsy, would require
|
|
4971 * inefficient searching. Turn the affixes into additions and/or
|
|
4972 * the expanded word.
|
|
4973 */
|
|
4974 MSG(_("Processing words..."));
|
|
4975 out_flush();
|
255
|
4976 error = expand_affixes(&newwords, &prefga, &suffga) == FAIL;
|
|
4977
|
|
4978 if (!error)
|
|
4979 {
|
|
4980 /* Write the info in the spell file. */
|
|
4981 smsg((char_u *)_("Writing spell file %s..."), wfname);
|
|
4982 out_flush();
|
|
4983 write_vim_spell(wfname, &prefga, &suffga, &newwords,
|
260
|
4984 fcount - 1, region_name, ascii);
|
255
|
4985 MSG(_("Done!"));
|
|
4986 out_flush();
|
|
4987 }
|
236
|
4988
|
|
4989 /* Free the allocated stuff. */
|
|
4990 free_wordtable(&newwords);
|
|
4991 for (round = 1; round <= 2; ++round)
|
|
4992 {
|
|
4993 gap = round == 1 ? &prefga: &suffga;
|
|
4994 for (i = 0; i < gap->ga_len; ++i)
|
|
4995 free_affixentries(((affheader_T *)gap->ga_data + i)
|
|
4996 ->ah_first);
|
|
4997 ga_clear(gap);
|
|
4998 }
|
|
4999 }
|
|
5000
|
|
5001 /* Free the .aff and .dic file structures. */
|
|
5002 for (i = 1; i < fcount; ++i)
|
|
5003 {
|
|
5004 if (afile[i - 1] != NULL)
|
|
5005 spell_free_aff(afile[i - 1]);
|
|
5006 spell_free_dic(&dfile[i - 1]);
|
|
5007 }
|
|
5008 }
|
|
5009
|
|
5010 theend:
|
|
5011 FreeWild(fcount, fnames);
|
|
5012 }
|
|
5013
|
|
5014 static void
|
|
5015 free_wordtable(ht)
|
|
5016 hashtab_T *ht;
|
|
5017 {
|
|
5018 int todo;
|
|
5019 basicword_T *bw, *nbw;
|
|
5020 hashitem_T *hi;
|
|
5021
|
|
5022 todo = ht->ht_used;
|
|
5023 for (hi = ht->ht_array; todo > 0; ++hi)
|
|
5024 {
|
|
5025 if (!HASHITEM_EMPTY(hi))
|
|
5026 {
|
|
5027 --todo;
|
|
5028 for (bw = HI2BW(hi); bw != NULL; bw = nbw)
|
|
5029 {
|
|
5030 nbw = bw->bw_next;
|
|
5031 free_basicword(bw);
|
|
5032 }
|
|
5033 }
|
|
5034 }
|
|
5035 }
|
|
5036
|
|
5037 /*
|
|
5038 * Free a basicword_T and what it contains.
|
|
5039 */
|
|
5040 static void
|
|
5041 free_basicword(bw)
|
|
5042 basicword_T *bw;
|
|
5043 {
|
|
5044 ga_clear(&bw->bw_prefix);
|
|
5045 ga_clear(&bw->bw_suffix);
|
|
5046 vim_free(bw->bw_caseword);
|
|
5047 vim_free(bw->bw_leadstring);
|
|
5048 vim_free(bw->bw_addstring);
|
|
5049 vim_free(bw);
|
|
5050 }
|
|
5051
|
|
5052 /*
|
240
|
5053 * Free a list of affentry_T and what they contain.
|
236
|
5054 */
|
|
5055 static void
|
|
5056 free_affixentries(first)
|
|
5057 affentry_T *first;
|
|
5058 {
|
|
5059 affentry_T *ap, *an;
|
|
5060
|
|
5061 for (ap = first; ap != NULL; ap = an)
|
|
5062 {
|
|
5063 an = ap->ae_next;
|
240
|
5064 free_affix_entry(ap);
|
236
|
5065 }
|
|
5066 }
|
|
5067
|
240
|
5068 /*
|
|
5069 * Free one affentry_T and what it contains.
|
|
5070 */
|
|
5071 static void
|
|
5072 free_affix_entry(ap)
|
|
5073 affentry_T *ap;
|
|
5074 {
|
|
5075 vim_free(ap->ae_chop);
|
|
5076 vim_free(ap->ae_add);
|
|
5077 vim_free(ap->ae_cond);
|
|
5078 vim_free(ap->ae_prog);
|
|
5079 vim_free(ap);
|
|
5080 }
|
|
5081
|
236
|
5082 #endif /* FEAT_MBYTE */
|
|
5083
|
|
5084 #endif /* FEAT_SYN_HL */
|