Mercurial > vim
annotate src/spell.c @ 18478:94223687df0e
Added tag v8.1.2233 for changeset e93cab5d0f0f27fad7882f1f412927df055b090d
| author | Bram Moolenaar <Bram@vim.org> |
|---|---|
| date | Tue, 29 Oct 2019 04:30:05 +0100 |
| parents | 34d5cd432cac |
| children | 7e7ec935e7c8 |
| rev | line source |
|---|---|
|
10042
4aead6a9b7a9
commit https://github.com/vim/vim/commit/edf3f97ae2af024708ebb4ac614227327033ca47
Christian Brabandt <cb@256bit.org>
parents:
9953
diff
changeset
|
1 /* vi:set ts=8 sts=4 sw=4 noet: |
| 223 | 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 * |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
13 * See spellfile.c for the Vim spell file format. |
|
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
14 * |
| 300 | 15 * The spell checking mechanism uses a tree (aka trie). Each node in the tree |
| 16 * has a list of bytes that can appear (siblings). For each byte there is a | |
| 17 * pointer to the node with the byte that follows in the word (child). | |
| 324 | 18 * |
| 19 * A NUL byte is used where the word may end. The bytes are sorted, so that | |
| 20 * binary searching can be used and the NUL bytes are at the start. The | |
| 21 * number of possible bytes is stored before the list of bytes. | |
| 22 * | |
| 23 * The tree uses two arrays: "byts" stores the characters, "idxs" stores | |
| 24 * either the next index or flags. The tree starts at index 0. For example, | |
| 25 * to lookup "vi" this sequence is followed: | |
| 26 * i = 0 | |
| 27 * len = byts[i] | |
| 28 * n = where "v" appears in byts[i + 1] to byts[i + len] | |
| 29 * i = idxs[n] | |
| 30 * len = byts[i] | |
| 31 * n = where "i" appears in byts[i + 1] to byts[i + len] | |
| 32 * i = idxs[n] | |
| 33 * len = byts[i] | |
| 34 * find that byts[i + 1] is 0, idxs[i + 1] has flags for "vi". | |
| 300 | 35 * |
| 339 | 36 * There are two word trees: one with case-folded words and one with words in |
| 300 | 37 * original case. The second one is only used for keep-case words and is |
| 38 * usually small. | |
| 39 * | |
| 481 | 40 * There is one additional tree for when not all prefixes are applied when |
| 339 | 41 * generating the .spl file. This tree stores all the possible prefixes, as |
| 42 * if they were words. At each word (prefix) end the prefix nr is stored, the | |
| 43 * following word must support this prefix nr. And the condition nr is | |
| 44 * stored, used to lookup the condition that the word must match with. | |
| 45 * | |
| 300 | 46 * Thanks to Olaf Seibert for providing an example implementation of this tree |
| 47 * and the compression mechanism. | |
| 625 | 48 * LZ trie ideas: |
| 49 * http://www.irb.hr/hr/home/ristov/papers/RistovLZtrieRevision1.pdf | |
| 50 * More papers: http://www-igm.univ-mlv.fr/~laporte/publi_en.html | |
| 243 | 51 * |
| 52 * Matching involves checking the caps type: Onecap ALLCAP KeepCap. | |
| 53 * | |
| 236 | 54 * Why doesn't Vim use aspell/ispell/myspell/etc.? |
| 55 * See ":help develop-spell". | |
| 56 */ | |
| 57 | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
58 #define IN_SPELL_C |
| 223 | 59 #include "vim.h" |
| 60 | |
| 737 | 61 #if defined(FEAT_SPELL) || defined(PROTO) |
| 223 | 62 |
| 625 | 63 #ifndef UNIX /* it's in os_unix.h for Unix */ |
| 64 # include <time.h> /* for time_t */ | |
| 65 #endif | |
| 66 | |
| 307 | 67 #define REGION_ALL 0xff /* word valid in all regions */ |
| 68 | |
| 625 | 69 #define VIMSUGMAGIC "VIMsug" /* string at start of Vim .sug file */ |
| 70 #define VIMSUGMAGICL 6 | |
| 71 #define VIMSUGVERSION 1 | |
| 72 | |
| 307 | 73 /* Result values. Lower number is accepted over higher one. */ |
| 74 #define SP_BANNED -1 | |
| 236 | 75 #define SP_OK 0 |
| 307 | 76 #define SP_RARE 1 |
| 77 #define SP_LOCAL 2 | |
| 78 #define SP_BAD 3 | |
| 236 | 79 |
| 323 | 80 /* |
| 236 | 81 * Structure to store info for word matching. |
| 82 */ | |
| 83 typedef struct matchinf_S | |
| 84 { | |
| 85 langp_T *mi_lp; /* info for language and region */ | |
| 243 | 86 |
| 87 /* pointers to original text to be checked */ | |
| 236 | 88 char_u *mi_word; /* start of word being checked */ |
| 339 | 89 char_u *mi_end; /* end of matching word so far */ |
| 243 | 90 char_u *mi_fend; /* next char to be added to mi_fword */ |
| 300 | 91 char_u *mi_cend; /* char after what was used for |
| 92 mi_capflags */ | |
| 243 | 93 |
| 94 /* case-folded text */ | |
| 95 char_u mi_fword[MAXWLEN + 1]; /* mi_word case-folded */ | |
| 300 | 96 int mi_fwordlen; /* nr of valid bytes in mi_fword */ |
| 243 | 97 |
| 339 | 98 /* for when checking word after a prefix */ |
| 99 int mi_prefarridx; /* index in sl_pidxs with list of | |
| 481 | 100 affixID/condition */ |
| 339 | 101 int mi_prefcnt; /* number of entries at mi_prefarridx */ |
| 102 int mi_prefixlen; /* byte length of prefix */ | |
| 455 | 103 int mi_cprefixlen; /* byte length of prefix in original |
| 104 case */ | |
| 339 | 105 |
| 481 | 106 /* for when checking a compound word */ |
| 107 int mi_compoff; /* start of following word offset */ | |
| 483 | 108 char_u mi_compflags[MAXWLEN]; /* flags for compound words used */ |
| 109 int mi_complen; /* nr of compound words used */ | |
| 809 | 110 int mi_compextra; /* nr of COMPOUNDROOT words */ |
| 481 | 111 |
| 243 | 112 /* others */ |
| 236 | 113 int mi_result; /* result so far: SP_BAD, SP_OK, etc. */ |
| 300 | 114 int mi_capflags; /* WF_ONECAP WF_ALLCAP WF_KEEPCAP */ |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
115 win_T *mi_win; /* buffer being checked */ |
| 492 | 116 |
| 117 /* for NOBREAK */ | |
| 118 int mi_result2; /* "mi_resul" without following word */ | |
| 119 char_u *mi_end2; /* "mi_end" without following word */ | |
| 236 | 120 } matchinf_T; |
| 121 | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
122 |
|
7805
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
123 static int spell_mb_isword_class(int cl, win_T *wp); |
| 307 | 124 |
| 339 | 125 /* mode values for find_word */ |
| 481 | 126 #define FIND_FOLDWORD 0 /* find word case-folded */ |
| 127 #define FIND_KEEPWORD 1 /* find keep-case word */ | |
| 128 #define FIND_PREFIX 2 /* find word after prefix */ | |
| 129 #define FIND_COMPOUND 3 /* find case-folded compound word */ | |
| 130 #define FIND_KEEPCOMPOUND 4 /* find keep-case compound word */ | |
| 339 | 131 |
|
7805
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
132 static void find_word(matchinf_T *mip, int mode); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
133 static void find_prefix(matchinf_T *mip, int mode); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
134 static int fold_more(matchinf_T *mip); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
135 static void spell_load_cb(char_u *fname, void *cookie); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
136 static int count_syllables(slang_T *slang, char_u *word); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
137 static void clear_midword(win_T *buf); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
138 static void use_midword(slang_T *lp, win_T *buf); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
139 static int find_region(char_u *rp, char_u *region); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
140 static void spell_soundfold_sofo(slang_T *slang, char_u *inword, char_u *res); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
141 static void spell_soundfold_sal(slang_T *slang, char_u *inword, char_u *res); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
142 static void spell_soundfold_wsal(slang_T *slang, char_u *inword, char_u *res); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
143 static void dump_word(slang_T *slang, char_u *word, char_u *pat, int *dir, int round, int flags, linenr_T lnum); |
|
0b6c37dd858d
commit https://github.com/vim/vim/commit/baaa7e9ec7398a813e21285c272fa99792642077
Christian Brabandt <cb@256bit.org>
parents:
7526
diff
changeset
|
144 static linenr_T dump_prefixes(slang_T *slang, char_u *word, char_u *pat, int *dir, int round, int flags, linenr_T startlnum); |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
145 |
| 236 | 146 /* |
| 147 * Main spell-checking function. | |
| 300 | 148 * "ptr" points to a character that could be the start of a word. |
| 534 | 149 * "*attrp" is set to the highlight index for a badly spelled word. For a |
| 150 * non-word or when it's OK it remains unchanged. | |
| 236 | 151 * This must only be called when 'spelllang' is not empty. |
| 323 | 152 * |
| 385 | 153 * "capcol" is used to check for a Capitalised word after the end of a |
| 154 * sentence. If it's zero then perform the check. Return the column where to | |
| 155 * check next, or -1 when no sentence end was found. If it's NULL then don't | |
| 156 * worry. | |
| 323 | 157 * |
| 236 | 158 * Returns the length of the word in bytes, also when it's OK, so that the |
| 159 * caller can skip over the word. | |
| 160 */ | |
| 161 int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
162 spell_check( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
163 win_T *wp, /* current window */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
164 char_u *ptr, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
165 hlf_T *attrp, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
166 int *capcol, /* column to check for Capital */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
167 int docount) /* count good words */ |
| 236 | 168 { |
| 169 matchinf_T mi; /* Most things are put in "mi" so that it can | |
| 170 be passed to functions quickly. */ | |
| 344 | 171 int nrlen = 0; /* found a number first */ |
| 385 | 172 int c; |
| 483 | 173 int wrongcaplen = 0; |
| 500 | 174 int lpi; |
| 625 | 175 int count_word = docount; |
| 236 | 176 |
| 307 | 177 /* A word never starts at a space or a control character. Return quickly |
| 178 * then, skipping over the character. */ | |
| 179 if (*ptr <= ' ') | |
| 180 return 1; | |
| 690 | 181 |
| 182 /* Return here when loading language files failed. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
183 if (wp->w_s->b_langp.ga_len == 0) |
| 690 | 184 return 1; |
| 185 | |
| 483 | 186 vim_memset(&mi, 0, sizeof(matchinf_T)); |
| 236 | 187 |
| 344 | 188 /* A number is always OK. Also skip hexadecimal numbers 0xFF99 and |
| 584 | 189 * 0X99FF. But always do check spelling to find "3GPP" and "11 |
| 190 * julifeest". */ | |
| 300 | 191 if (*ptr >= '0' && *ptr <= '9') |
| 192 { | |
|
7447
ad432f8f68fb
commit https://github.com/vim/vim/commit/887c1fea4a114e7170091942d0446c8882701b5b
Christian Brabandt <cb@256bit.org>
parents:
6949
diff
changeset
|
193 if (*ptr == '0' && (ptr[1] == 'b' || ptr[1] == 'B')) |
|
ad432f8f68fb
commit https://github.com/vim/vim/commit/887c1fea4a114e7170091942d0446c8882701b5b
Christian Brabandt <cb@256bit.org>
parents:
6949
diff
changeset
|
194 mi.mi_end = skipbin(ptr + 2); |
|
ad432f8f68fb
commit https://github.com/vim/vim/commit/887c1fea4a114e7170091942d0446c8882701b5b
Christian Brabandt <cb@256bit.org>
parents:
6949
diff
changeset
|
195 else if (*ptr == '0' && (ptr[1] == 'x' || ptr[1] == 'X')) |
| 316 | 196 mi.mi_end = skiphex(ptr + 2); |
| 300 | 197 else |
| 344 | 198 mi.mi_end = skipdigits(ptr); |
| 835 | 199 nrlen = (int)(mi.mi_end - ptr); |
| 584 | 200 } |
| 346 | 201 |
| 202 /* Find the normal end of the word (until the next non-word character). */ | |
| 344 | 203 mi.mi_word = ptr; |
| 584 | 204 mi.mi_fend = ptr; |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
205 if (spell_iswordp(mi.mi_fend, wp)) |
| 344 | 206 { |
| 207 do | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
208 MB_PTR_ADV(mi.mi_fend); |
|
16162
cd5c83115ec6
patch 8.1.1086: too many curly braces
Bram Moolenaar <Bram@vim.org>
parents:
16142
diff
changeset
|
209 while (*mi.mi_fend != NUL && spell_iswordp(mi.mi_fend, wp)); |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
210 |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
211 if (capcol != NULL && *capcol == 0 && wp->w_s->b_cap_prog != NULL) |
| 385 | 212 { |
| 213 /* Check word starting with capital letter. */ | |
| 455 | 214 c = PTR2CHAR(ptr); |
| 385 | 215 if (!SPELL_ISUPPER(c)) |
| 483 | 216 wrongcaplen = (int)(mi.mi_fend - ptr); |
| 385 | 217 } |
| 218 } | |
| 219 if (capcol != NULL) | |
| 220 *capcol = -1; | |
| 344 | 221 |
| 222 /* We always use the characters up to the next non-word character, | |
| 223 * also for bad words. */ | |
| 224 mi.mi_end = mi.mi_fend; | |
| 225 | |
| 226 /* Check caps type later. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
227 mi.mi_capflags = 0; |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
228 mi.mi_cend = NULL; |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
229 mi.mi_win = wp; |
| 344 | 230 |
| 483 | 231 /* case-fold the word with one non-word character, so that we can check |
| 232 * for the word end. */ | |
| 344 | 233 if (*mi.mi_fend != NUL) |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
234 MB_PTR_ADV(mi.mi_fend); |
| 344 | 235 |
| 236 (void)spell_casefold(ptr, (int)(mi.mi_fend - ptr), mi.mi_fword, | |
| 237 MAXWLEN + 1); | |
| 835 | 238 mi.mi_fwordlen = (int)STRLEN(mi.mi_fword); |
| 344 | 239 |
| 240 /* The word is bad unless we recognize it. */ | |
| 241 mi.mi_result = SP_BAD; | |
| 492 | 242 mi.mi_result2 = SP_BAD; |
| 344 | 243 |
| 244 /* | |
| 245 * Loop over the languages specified in 'spelllang'. | |
| 625 | 246 * We check them all, because a word may be matched longer in another |
| 247 * language. | |
| 344 | 248 */ |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
249 for (lpi = 0; lpi < wp->w_s->b_langp.ga_len; ++lpi) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
250 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
251 mi.mi_lp = LANGP_ENTRY(wp->w_s->b_langp, lpi); |
| 500 | 252 |
| 253 /* If reloading fails the language is still in the list but everything | |
| 254 * has been cleared. */ | |
| 255 if (mi.mi_lp->lp_slang->sl_fidxs == NULL) | |
| 256 continue; | |
| 257 | |
| 344 | 258 /* Check for a matching word in case-folded words. */ |
| 259 find_word(&mi, FIND_FOLDWORD); | |
| 260 | |
| 261 /* Check for a matching word in keep-case words. */ | |
| 262 find_word(&mi, FIND_KEEPWORD); | |
| 263 | |
| 264 /* Check for matching prefixes. */ | |
| 485 | 265 find_prefix(&mi, FIND_FOLDWORD); |
| 492 | 266 |
| 267 /* For a NOBREAK language, may want to use a word without a following | |
| 268 * word as a backup. */ | |
| 269 if (mi.mi_lp->lp_slang->sl_nobreak && mi.mi_result == SP_BAD | |
| 270 && mi.mi_result2 != SP_BAD) | |
| 271 { | |
| 272 mi.mi_result = mi.mi_result2; | |
| 273 mi.mi_end = mi.mi_end2; | |
| 274 } | |
| 625 | 275 |
| 276 /* Count the word in the first language where it's found to be OK. */ | |
| 277 if (count_word && mi.mi_result == SP_OK) | |
| 278 { | |
| 279 count_common_word(mi.mi_lp->lp_slang, ptr, | |
| 280 (int)(mi.mi_end - ptr), 1); | |
| 281 count_word = FALSE; | |
| 282 } | |
| 344 | 283 } |
| 284 | |
| 285 if (mi.mi_result != SP_OK) | |
| 286 { | |
| 346 | 287 /* If we found a number skip over it. Allows for "42nd". Do flag |
| 288 * rare and local words, e.g., "3GPP". */ | |
| 344 | 289 if (nrlen > 0) |
| 346 | 290 { |
| 291 if (mi.mi_result == SP_BAD || mi.mi_result == SP_BANNED) | |
| 292 return nrlen; | |
| 293 } | |
| 344 | 294 |
| 295 /* When we are at a non-word character there is no error, just | |
| 296 * skip over the character (try looking for a word after it). */ | |
| 5477 | 297 else if (!spell_iswordp_nmw(ptr, wp)) |
| 243 | 298 { |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
299 if (capcol != NULL && wp->w_s->b_cap_prog != NULL) |
| 385 | 300 { |
| 301 regmatch_T regmatch; | |
| 6375 | 302 int r; |
| 385 | 303 |
| 304 /* Check for end of sentence. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
305 regmatch.regprog = wp->w_s->b_cap_prog; |
| 385 | 306 regmatch.rm_ic = FALSE; |
| 6375 | 307 r = vim_regexec(®match, ptr, 0); |
| 308 wp->w_s->b_cap_prog = regmatch.regprog; | |
| 309 if (r) | |
| 385 | 310 *capcol = (int)(regmatch.endp[0] - ptr); |
| 311 } | |
| 312 | |
| 344 | 313 if (has_mbyte) |
| 474 | 314 return (*mb_ptr2len)(ptr); |
| 344 | 315 return 1; |
| 300 | 316 } |
| 483 | 317 else if (mi.mi_end == ptr) |
| 318 /* Always include at least one character. Required for when there | |
| 319 * is a mixup in "midword". */ | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
320 MB_PTR_ADV(mi.mi_end); |
| 492 | 321 else if (mi.mi_result == SP_BAD |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
322 && LANGP_ENTRY(wp->w_s->b_langp, 0)->lp_slang->sl_nobreak) |
| 492 | 323 { |
| 324 char_u *p, *fp; | |
| 325 int save_result = mi.mi_result; | |
| 326 | |
| 327 /* First language in 'spelllang' is NOBREAK. Find first position | |
| 328 * at which any word would be valid. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
329 mi.mi_lp = LANGP_ENTRY(wp->w_s->b_langp, 0); |
| 500 | 330 if (mi.mi_lp->lp_slang->sl_fidxs != NULL) |
| 331 { | |
| 332 p = mi.mi_word; | |
| 333 fp = mi.mi_fword; | |
| 334 for (;;) | |
| 335 { | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
336 MB_PTR_ADV(p); |
|
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
337 MB_PTR_ADV(fp); |
| 500 | 338 if (p >= mi.mi_end) |
| 339 break; | |
| 835 | 340 mi.mi_compoff = (int)(fp - mi.mi_fword); |
| 500 | 341 find_word(&mi, FIND_COMPOUND); |
| 342 if (mi.mi_result != SP_BAD) | |
| 343 { | |
| 344 mi.mi_end = p; | |
| 345 break; | |
| 346 } | |
| 347 } | |
| 348 mi.mi_result = save_result; | |
| 349 } | |
| 492 | 350 } |
| 243 | 351 |
| 344 | 352 if (mi.mi_result == SP_BAD || mi.mi_result == SP_BANNED) |
| 534 | 353 *attrp = HLF_SPB; |
| 344 | 354 else if (mi.mi_result == SP_RARE) |
| 534 | 355 *attrp = HLF_SPR; |
| 344 | 356 else |
| 534 | 357 *attrp = HLF_SPL; |
| 243 | 358 } |
| 359 | |
| 483 | 360 if (wrongcaplen > 0 && (mi.mi_result == SP_OK || mi.mi_result == SP_RARE)) |
| 361 { | |
| 362 /* Report SpellCap only when the word isn't badly spelled. */ | |
| 534 | 363 *attrp = HLF_SPC; |
| 483 | 364 return wrongcaplen; |
| 365 } | |
| 366 | |
| 300 | 367 return (int)(mi.mi_end - ptr); |
| 236 | 368 } |
| 369 | |
| 370 /* | |
| 300 | 371 * Check if the word at "mip->mi_word" is in the tree. |
| 339 | 372 * When "mode" is FIND_FOLDWORD check in fold-case word tree. |
| 373 * When "mode" is FIND_KEEPWORD check in keep-case word tree. | |
| 374 * When "mode" is FIND_PREFIX check for word after prefix in fold-case word | |
| 375 * tree. | |
| 300 | 376 * |
| 377 * For a match mip->mi_result is updated. | |
| 243 | 378 */ |
| 379 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
380 find_word(matchinf_T *mip, int mode) |
| 243 | 381 { |
| 324 | 382 idx_T arridx = 0; |
| 300 | 383 int endlen[MAXWLEN]; /* length at possible word endings */ |
| 324 | 384 idx_T endidx[MAXWLEN]; /* possible word endings */ |
| 300 | 385 int endidxcnt = 0; |
| 386 int len; | |
| 387 int wlen = 0; | |
| 388 int flen; | |
| 389 int c; | |
| 390 char_u *ptr; | |
| 324 | 391 idx_T lo, hi, m; |
| 300 | 392 char_u *s; |
| 339 | 393 char_u *p; |
| 307 | 394 int res = SP_BAD; |
| 300 | 395 slang_T *slang = mip->mi_lp->lp_slang; |
| 396 unsigned flags; | |
| 397 char_u *byts; | |
| 324 | 398 idx_T *idxs; |
| 481 | 399 int word_ends; |
| 485 | 400 int prefix_found; |
| 492 | 401 int nobreak_result; |
| 481 | 402 |
| 403 if (mode == FIND_KEEPWORD || mode == FIND_KEEPCOMPOUND) | |
| 236 | 404 { |
| 300 | 405 /* Check for word with matching case in keep-case tree. */ |
| 406 ptr = mip->mi_word; | |
| 407 flen = 9999; /* no case folding, always enough bytes */ | |
| 408 byts = slang->sl_kbyts; | |
| 409 idxs = slang->sl_kidxs; | |
| 481 | 410 |
| 411 if (mode == FIND_KEEPCOMPOUND) | |
| 412 /* Skip over the previously found word(s). */ | |
| 413 wlen += mip->mi_compoff; | |
| 236 | 414 } |
| 415 else | |
| 416 { | |
| 300 | 417 /* Check for case-folded in case-folded tree. */ |
| 418 ptr = mip->mi_fword; | |
| 419 flen = mip->mi_fwordlen; /* available case-folded bytes */ | |
| 420 byts = slang->sl_fbyts; | |
| 421 idxs = slang->sl_fidxs; | |
| 339 | 422 |
| 423 if (mode == FIND_PREFIX) | |
| 424 { | |
| 425 /* Skip over the prefix. */ | |
| 426 wlen = mip->mi_prefixlen; | |
| 427 flen -= mip->mi_prefixlen; | |
| 428 } | |
| 481 | 429 else if (mode == FIND_COMPOUND) |
| 430 { | |
| 431 /* Skip over the previously found word(s). */ | |
| 432 wlen = mip->mi_compoff; | |
| 433 flen -= mip->mi_compoff; | |
| 434 } | |
| 435 | |
| 243 | 436 } |
| 437 | |
| 300 | 438 if (byts == NULL) |
| 439 return; /* array is empty */ | |
| 236 | 440 |
| 441 /* | |
| 307 | 442 * Repeat advancing in the tree until: |
| 443 * - there is a byte that doesn't match, | |
| 444 * - we reach the end of the tree, | |
| 445 * - or we reach the end of the line. | |
| 236 | 446 */ |
| 300 | 447 for (;;) |
| 236 | 448 { |
| 346 | 449 if (flen <= 0 && *mip->mi_fend != NUL) |
| 339 | 450 flen = fold_more(mip); |
| 300 | 451 |
| 452 len = byts[arridx++]; | |
| 453 | |
| 454 /* If the first possible byte is a zero the word could end here. | |
| 455 * Remember this index, we first check for the longest word. */ | |
| 456 if (byts[arridx] == 0) | |
| 457 { | |
| 307 | 458 if (endidxcnt == MAXWLEN) |
| 459 { | |
| 460 /* Must be a corrupted spell file. */ | |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
461 emsg(_(e_format)); |
| 307 | 462 return; |
| 463 } | |
| 300 | 464 endlen[endidxcnt] = wlen; |
| 465 endidx[endidxcnt++] = arridx++; | |
| 466 --len; | |
| 467 | |
| 468 /* Skip over the zeros, there can be several flag/region | |
| 469 * combinations. */ | |
| 470 while (len > 0 && byts[arridx] == 0) | |
| 471 { | |
| 472 ++arridx; | |
| 473 --len; | |
| 474 } | |
| 475 if (len == 0) | |
| 476 break; /* no children, word must end here */ | |
| 477 } | |
| 478 | |
| 479 /* Stop looking at end of the line. */ | |
| 480 if (ptr[wlen] == NUL) | |
| 481 break; | |
| 482 | |
| 483 /* Perform a binary search in the list of accepted bytes. */ | |
| 484 c = ptr[wlen]; | |
| 346 | 485 if (c == TAB) /* <Tab> is handled like <Space> */ |
| 486 c = ' '; | |
| 300 | 487 lo = arridx; |
| 488 hi = arridx + len - 1; | |
| 489 while (lo < hi) | |
| 490 { | |
| 491 m = (lo + hi) / 2; | |
| 492 if (byts[m] > c) | |
| 493 hi = m - 1; | |
| 494 else if (byts[m] < c) | |
| 495 lo = m + 1; | |
| 496 else | |
| 497 { | |
| 498 lo = hi = m; | |
| 499 break; | |
| 236 | 500 } |
| 501 } | |
| 300 | 502 |
| 503 /* Stop if there is no matching byte. */ | |
| 504 if (hi < lo || byts[lo] != c) | |
| 505 break; | |
| 506 | |
| 507 /* Continue at the child (if there is one). */ | |
| 508 arridx = idxs[lo]; | |
| 509 ++wlen; | |
| 510 --flen; | |
| 346 | 511 |
| 512 /* One space in the good word may stand for several spaces in the | |
| 513 * checked word. */ | |
| 514 if (c == ' ') | |
| 515 { | |
| 516 for (;;) | |
| 517 { | |
| 518 if (flen <= 0 && *mip->mi_fend != NUL) | |
| 519 flen = fold_more(mip); | |
| 520 if (ptr[wlen] != ' ' && ptr[wlen] != TAB) | |
| 521 break; | |
| 522 ++wlen; | |
| 523 --flen; | |
| 524 } | |
| 525 } | |
| 236 | 526 } |
| 527 | |
| 300 | 528 /* |
| 529 * Verify that one of the possible endings is valid. Try the longest | |
| 530 * first. | |
| 531 */ | |
| 532 while (endidxcnt > 0) | |
| 533 { | |
| 534 --endidxcnt; | |
| 535 arridx = endidx[endidxcnt]; | |
| 536 wlen = endlen[endidxcnt]; | |
| 236 | 537 |
| 300 | 538 if ((*mb_head_off)(ptr, ptr + wlen) > 0) |
| 539 continue; /* not at first byte of character */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
540 if (spell_iswordp(ptr + wlen, mip->mi_win)) |
| 481 | 541 { |
| 492 | 542 if (slang->sl_compprog == NULL && !slang->sl_nobreak) |
| 481 | 543 continue; /* next char is a word character */ |
| 544 word_ends = FALSE; | |
| 545 } | |
| 546 else | |
| 547 word_ends = TRUE; | |
| 485 | 548 /* The prefix flag is before compound flags. Once a valid prefix flag |
| 549 * has been found we try compound flags. */ | |
| 550 prefix_found = FALSE; | |
| 300 | 551 |
| 339 | 552 if (mode != FIND_KEEPWORD && has_mbyte) |
| 300 | 553 { |
| 554 /* Compute byte length in original word, length may change | |
| 339 | 555 * when folding case. This can be slow, take a shortcut when the |
| 556 * case-folded word is equal to the keep-case word. */ | |
| 300 | 557 p = mip->mi_word; |
| 339 | 558 if (STRNCMP(ptr, p, wlen) != 0) |
| 559 { | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
560 for (s = ptr; s < ptr + wlen; MB_PTR_ADV(s)) |
|
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
561 MB_PTR_ADV(p); |
| 835 | 562 wlen = (int)(p - mip->mi_word); |
| 339 | 563 } |
| 300 | 564 } |
| 236 | 565 |
| 339 | 566 /* Check flags and region. For FIND_PREFIX check the condition and |
| 567 * prefix ID. | |
| 568 * Repeat this if there are more flags/region alternatives until there | |
| 569 * is a match. */ | |
| 570 res = SP_BAD; | |
| 571 for (len = byts[arridx - 1]; len > 0 && byts[arridx] == 0; | |
| 572 --len, ++arridx) | |
| 300 | 573 { |
| 574 flags = idxs[arridx]; | |
| 324 | 575 |
| 339 | 576 /* For the fold-case tree check that the case of the checked word |
| 577 * matches with what the word in the tree requires. | |
| 578 * For keep-case tree the case is always right. For prefixes we | |
| 579 * don't bother to check. */ | |
| 580 if (mode == FIND_FOLDWORD) | |
| 300 | 581 { |
| 582 if (mip->mi_cend != mip->mi_word + wlen) | |
| 583 { | |
| 323 | 584 /* mi_capflags was set for a different word length, need |
| 585 * to do it again. */ | |
| 300 | 586 mip->mi_cend = mip->mi_word + wlen; |
| 323 | 587 mip->mi_capflags = captype(mip->mi_word, mip->mi_cend); |
| 300 | 588 } |
| 589 | |
| 346 | 590 if (mip->mi_capflags == WF_KEEPCAP |
| 591 || !spell_valid_case(mip->mi_capflags, flags)) | |
| 339 | 592 continue; |
| 300 | 593 } |
| 236 | 594 |
| 339 | 595 /* When mode is FIND_PREFIX the word must support the prefix: |
| 596 * check the prefix ID and the condition. Do that for the list at | |
| 366 | 597 * mip->mi_prefarridx that find_prefix() filled. */ |
| 485 | 598 else if (mode == FIND_PREFIX && !prefix_found) |
| 481 | 599 { |
| 366 | 600 c = valid_word_prefix(mip->mi_prefcnt, mip->mi_prefarridx, |
| 390 | 601 flags, |
| 455 | 602 mip->mi_word + mip->mi_cprefixlen, slang, |
| 603 FALSE); | |
| 366 | 604 if (c == 0) |
| 339 | 605 continue; |
| 366 | 606 |
| 607 /* Use the WF_RARE flag for a rare prefix. */ | |
| 608 if (c & WF_RAREPFX) | |
| 609 flags |= WF_RARE; | |
| 485 | 610 prefix_found = TRUE; |
| 339 | 611 } |
| 612 | |
| 492 | 613 if (slang->sl_nobreak) |
| 614 { | |
| 615 if ((mode == FIND_COMPOUND || mode == FIND_KEEPCOMPOUND) | |
| 616 && (flags & WF_BANNED) == 0) | |
| 617 { | |
| 618 /* NOBREAK: found a valid following word. That's all we | |
| 619 * need to know, so return. */ | |
| 620 mip->mi_result = SP_OK; | |
| 621 break; | |
| 622 } | |
| 623 } | |
| 624 | |
| 625 else if ((mode == FIND_COMPOUND || mode == FIND_KEEPCOMPOUND | |
| 626 || !word_ends)) | |
| 481 | 627 { |
| 974 | 628 /* If there is no compound flag or the word is shorter than |
| 483 | 629 * COMPOUNDMIN reject it quickly. |
| 630 * Makes you wonder why someone puts a compound flag on a word | |
| 481 | 631 * that's too short... Myspell compatibility requires this |
| 632 * anyway. */ | |
| 490 | 633 if (((unsigned)flags >> 24) == 0 |
| 634 || wlen - mip->mi_compoff < slang->sl_compminlen) | |
| 483 | 635 continue; |
| 500 | 636 /* For multi-byte chars check character length against |
| 637 * COMPOUNDMIN. */ | |
| 638 if (has_mbyte | |
| 501 | 639 && slang->sl_compminlen > 0 |
| 500 | 640 && mb_charlen_len(mip->mi_word + mip->mi_compoff, |
| 641 wlen - mip->mi_compoff) < slang->sl_compminlen) | |
| 642 continue; | |
| 483 | 643 |
| 809 | 644 /* Limit the number of compound words to COMPOUNDWORDMAX if no |
| 490 | 645 * maximum for syllables is specified. */ |
| 809 | 646 if (!word_ends && mip->mi_complen + mip->mi_compextra + 2 |
| 647 > slang->sl_compmax | |
| 490 | 648 && slang->sl_compsylmax == MAXWLEN) |
| 483 | 649 continue; |
| 650 | |
| 819 | 651 /* Don't allow compounding on a side where an affix was added, |
| 652 * unless COMPOUNDPERMITFLAG was used. */ | |
| 653 if (mip->mi_complen > 0 && (flags & WF_NOCOMPBEF)) | |
| 654 continue; | |
| 655 if (!word_ends && (flags & WF_NOCOMPAFT)) | |
| 656 continue; | |
| 657 | |
| 485 | 658 /* Quickly check if compounding is possible with this flag. */ |
| 495 | 659 if (!byte_in_str(mip->mi_complen == 0 |
| 485 | 660 ? slang->sl_compstartflags |
| 661 : slang->sl_compallflags, | |
| 495 | 662 ((unsigned)flags >> 24))) |
| 481 | 663 continue; |
| 664 | |
| 1762 | 665 /* If there is a match with a CHECKCOMPOUNDPATTERN rule |
| 666 * discard the compound word. */ | |
| 667 if (match_checkcompoundpattern(ptr, wlen, &slang->sl_comppat)) | |
| 668 continue; | |
| 669 | |
| 490 | 670 if (mode == FIND_COMPOUND) |
| 671 { | |
| 672 int capflags; | |
| 673 | |
| 674 /* Need to check the caps type of the appended compound | |
| 675 * word. */ | |
| 676 if (has_mbyte && STRNCMP(ptr, mip->mi_word, | |
| 677 mip->mi_compoff) != 0) | |
| 678 { | |
| 679 /* case folding may have changed the length */ | |
| 680 p = mip->mi_word; | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
681 for (s = ptr; s < ptr + mip->mi_compoff; MB_PTR_ADV(s)) |
|
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
682 MB_PTR_ADV(p); |
| 490 | 683 } |
| 684 else | |
| 685 p = mip->mi_word + mip->mi_compoff; | |
| 686 capflags = captype(p, mip->mi_word + wlen); | |
| 687 if (capflags == WF_KEEPCAP || (capflags == WF_ALLCAP | |
| 688 && (flags & WF_FIXCAP) != 0)) | |
| 689 continue; | |
| 690 | |
| 691 if (capflags != WF_ALLCAP) | |
| 692 { | |
| 693 /* When the character before the word is a word | |
| 694 * character we do not accept a Onecap word. We do | |
| 695 * accept a no-caps word, even when the dictionary | |
| 696 * word specifies ONECAP. */ | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
697 MB_PTR_BACK(mip->mi_word, p); |
| 5477 | 698 if (spell_iswordp_nmw(p, mip->mi_win) |
| 490 | 699 ? capflags == WF_ONECAP |
| 700 : (flags & WF_ONECAP) != 0 | |
| 701 && capflags != WF_ONECAP) | |
| 702 continue; | |
| 703 } | |
| 704 } | |
| 705 | |
| 483 | 706 /* If the word ends the sequence of compound flags of the |
| 714 | 707 * words must match with one of the COMPOUNDRULE items and |
| 483 | 708 * the number of syllables must not be too large. */ |
| 709 mip->mi_compflags[mip->mi_complen] = ((unsigned)flags >> 24); | |
| 710 mip->mi_compflags[mip->mi_complen + 1] = NUL; | |
| 711 if (word_ends) | |
| 712 { | |
| 713 char_u fword[MAXWLEN]; | |
| 714 | |
| 715 if (slang->sl_compsylmax < MAXWLEN) | |
| 716 { | |
| 717 /* "fword" is only needed for checking syllables. */ | |
| 718 if (ptr == mip->mi_word) | |
| 719 (void)spell_casefold(ptr, wlen, fword, MAXWLEN); | |
| 720 else | |
| 721 vim_strncpy(fword, ptr, endlen[endidxcnt]); | |
| 722 } | |
| 723 if (!can_compound(slang, fword, mip->mi_compflags)) | |
| 724 continue; | |
| 725 } | |
| 1762 | 726 else if (slang->sl_comprules != NULL |
| 727 && !match_compoundrule(slang, mip->mi_compflags)) | |
| 728 /* The compound flags collected so far do not match any | |
| 729 * COMPOUNDRULE, discard the compounded word. */ | |
| 730 continue; | |
| 481 | 731 } |
| 732 | |
| 500 | 733 /* Check NEEDCOMPOUND: can't use word without compounding. */ |
| 734 else if (flags & WF_NEEDCOMP) | |
| 735 continue; | |
| 736 | |
| 492 | 737 nobreak_result = SP_OK; |
| 738 | |
| 481 | 739 if (!word_ends) |
| 740 { | |
| 492 | 741 int save_result = mip->mi_result; |
| 742 char_u *save_end = mip->mi_end; | |
| 501 | 743 langp_T *save_lp = mip->mi_lp; |
| 744 int lpi; | |
| 492 | 745 |
| 746 /* Check that a valid word follows. If there is one and we | |
| 747 * are compounding, it will set "mi_result", thus we are | |
| 748 * always finished here. For NOBREAK we only check that a | |
| 749 * valid word follows. | |
| 481 | 750 * Recursive! */ |
| 492 | 751 if (slang->sl_nobreak) |
| 752 mip->mi_result = SP_BAD; | |
| 481 | 753 |
| 754 /* Find following word in case-folded tree. */ | |
| 755 mip->mi_compoff = endlen[endidxcnt]; | |
| 756 if (has_mbyte && mode == FIND_KEEPWORD) | |
| 757 { | |
| 758 /* Compute byte length in case-folded word from "wlen": | |
| 759 * byte length in keep-case word. Length may change when | |
| 760 * folding case. This can be slow, take a shortcut when | |
| 761 * the case-folded word is equal to the keep-case word. */ | |
| 762 p = mip->mi_fword; | |
| 763 if (STRNCMP(ptr, p, wlen) != 0) | |
| 764 { | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
765 for (s = ptr; s < ptr + wlen; MB_PTR_ADV(s)) |
|
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
766 MB_PTR_ADV(p); |
| 835 | 767 mip->mi_compoff = (int)(p - mip->mi_fword); |
| 481 | 768 } |
| 769 } | |
|
8956
d9e671c5afe6
commit https://github.com/vim/vim/commit/ba53435144f46eaaa53c63a62e748b3feee9742c
Christian Brabandt <cb@256bit.org>
parents:
8524
diff
changeset
|
770 #if 0 /* Disabled, see below */ |
| 485 | 771 c = mip->mi_compoff; |
|
8956
d9e671c5afe6
commit https://github.com/vim/vim/commit/ba53435144f46eaaa53c63a62e748b3feee9742c
Christian Brabandt <cb@256bit.org>
parents:
8524
diff
changeset
|
772 #endif |
| 483 | 773 ++mip->mi_complen; |
| 809 | 774 if (flags & WF_COMPROOT) |
| 775 ++mip->mi_compextra; | |
| 501 | 776 |
| 777 /* For NOBREAK we need to try all NOBREAK languages, at least | |
| 778 * to find the ".add" file(s). */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
779 for (lpi = 0; lpi < mip->mi_win->w_s->b_langp.ga_len; ++lpi) |
| 501 | 780 { |
| 781 if (slang->sl_nobreak) | |
| 782 { | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
783 mip->mi_lp = LANGP_ENTRY(mip->mi_win->w_s->b_langp, lpi); |
| 501 | 784 if (mip->mi_lp->lp_slang->sl_fidxs == NULL |
| 785 || !mip->mi_lp->lp_slang->sl_nobreak) | |
| 786 continue; | |
| 787 } | |
| 788 | |
| 789 find_word(mip, FIND_COMPOUND); | |
| 790 | |
| 791 /* When NOBREAK any word that matches is OK. Otherwise we | |
| 792 * need to find the longest match, thus try with keep-case | |
| 793 * and prefix too. */ | |
| 492 | 794 if (!slang->sl_nobreak || mip->mi_result == SP_BAD) |
| 795 { | |
| 501 | 796 /* Find following word in keep-case tree. */ |
| 797 mip->mi_compoff = wlen; | |
| 798 find_word(mip, FIND_KEEPCOMPOUND); | |
| 799 | |
| 819 | 800 #if 0 /* Disabled, a prefix must not appear halfway a compound word, |
| 801 unless the COMPOUNDPERMITFLAG is used and then it can't be a | |
| 802 postponed prefix. */ | |
| 501 | 803 if (!slang->sl_nobreak || mip->mi_result == SP_BAD) |
| 804 { | |
| 805 /* Check for following word with prefix. */ | |
| 806 mip->mi_compoff = c; | |
| 807 find_prefix(mip, FIND_COMPOUND); | |
| 808 } | |
| 819 | 809 #endif |
| 492 | 810 } |
| 501 | 811 |
| 812 if (!slang->sl_nobreak) | |
| 813 break; | |
| 492 | 814 } |
| 483 | 815 --mip->mi_complen; |
| 809 | 816 if (flags & WF_COMPROOT) |
| 817 --mip->mi_compextra; | |
| 501 | 818 mip->mi_lp = save_lp; |
| 485 | 819 |
| 492 | 820 if (slang->sl_nobreak) |
| 821 { | |
| 822 nobreak_result = mip->mi_result; | |
| 823 mip->mi_result = save_result; | |
| 824 mip->mi_end = save_end; | |
| 825 } | |
| 826 else | |
| 827 { | |
| 828 if (mip->mi_result == SP_OK) | |
| 829 break; | |
| 830 continue; | |
| 831 } | |
| 481 | 832 } |
| 833 | |
| 339 | 834 if (flags & WF_BANNED) |
| 835 res = SP_BANNED; | |
| 836 else if (flags & WF_REGION) | |
| 837 { | |
| 838 /* Check region. */ | |
| 390 | 839 if ((mip->mi_lp->lp_region & (flags >> 16)) != 0) |
| 300 | 840 res = SP_OK; |
| 339 | 841 else |
| 842 res = SP_LOCAL; | |
| 300 | 843 } |
| 339 | 844 else if (flags & WF_RARE) |
| 845 res = SP_RARE; | |
| 307 | 846 else |
| 339 | 847 res = SP_OK; |
| 848 | |
| 492 | 849 /* Always use the longest match and the best result. For NOBREAK |
| 850 * we separately keep the longest match without a following good | |
| 851 * word as a fall-back. */ | |
| 852 if (nobreak_result == SP_BAD) | |
| 853 { | |
| 854 if (mip->mi_result2 > res) | |
| 855 { | |
| 856 mip->mi_result2 = res; | |
| 857 mip->mi_end2 = mip->mi_word + wlen; | |
| 858 } | |
| 859 else if (mip->mi_result2 == res | |
| 860 && mip->mi_end2 < mip->mi_word + wlen) | |
| 861 mip->mi_end2 = mip->mi_word + wlen; | |
| 862 } | |
| 863 else if (mip->mi_result > res) | |
| 339 | 864 { |
| 865 mip->mi_result = res; | |
| 866 mip->mi_end = mip->mi_word + wlen; | |
| 867 } | |
| 351 | 868 else if (mip->mi_result == res && mip->mi_end < mip->mi_word + wlen) |
| 339 | 869 mip->mi_end = mip->mi_word + wlen; |
| 870 | |
| 492 | 871 if (mip->mi_result == SP_OK) |
| 339 | 872 break; |
| 300 | 873 } |
| 874 | |
| 492 | 875 if (mip->mi_result == SP_OK) |
| 300 | 876 break; |
| 877 } | |
| 236 | 878 } |
| 879 | |
| 323 | 880 /* |
| 1762 | 881 * Return TRUE if there is a match between the word ptr[wlen] and |
| 882 * CHECKCOMPOUNDPATTERN rules, assuming that we will concatenate with another | |
| 883 * word. | |
| 884 * A match means that the first part of CHECKCOMPOUNDPATTERN matches at the | |
| 885 * end of ptr[wlen] and the second part matches after it. | |
| 886 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
887 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
888 match_checkcompoundpattern( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
889 char_u *ptr, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
890 int wlen, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
891 garray_T *gap) /* &sl_comppat */ |
| 1762 | 892 { |
| 893 int i; | |
| 894 char_u *p; | |
| 895 int len; | |
| 896 | |
| 897 for (i = 0; i + 1 < gap->ga_len; i += 2) | |
| 898 { | |
| 899 p = ((char_u **)gap->ga_data)[i + 1]; | |
| 900 if (STRNCMP(ptr + wlen, p, STRLEN(p)) == 0) | |
| 901 { | |
| 902 /* Second part matches at start of following compound word, now | |
| 903 * check if first part matches at end of previous word. */ | |
| 904 p = ((char_u **)gap->ga_data)[i]; | |
| 1771 | 905 len = (int)STRLEN(p); |
| 1762 | 906 if (len <= wlen && STRNCMP(ptr + wlen - len, p, len) == 0) |
| 907 return TRUE; | |
| 908 } | |
| 909 } | |
| 910 return FALSE; | |
| 911 } | |
| 912 | |
| 913 /* | |
| 626 | 914 * Return TRUE if "flags" is a valid sequence of compound flags and "word" |
| 915 * does not have too many syllables. | |
| 482 | 916 */ |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
917 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
918 can_compound(slang_T *slang, char_u *word, char_u *flags) |
| 483 | 919 { |
| 495 | 920 char_u uflags[MAXWLEN * 2]; |
| 921 int i; | |
| 922 char_u *p; | |
| 483 | 923 |
| 924 if (slang->sl_compprog == NULL) | |
| 925 return FALSE; | |
| 495 | 926 if (enc_utf8) |
| 927 { | |
| 928 /* Need to convert the single byte flags to utf8 characters. */ | |
| 929 p = uflags; | |
| 930 for (i = 0; flags[i] != NUL; ++i) | |
|
11269
121d29004998
patch 8.0.0520: using a function pointer while the function is known
Christian Brabandt <cb@256bit.org>
parents:
11129
diff
changeset
|
931 p += utf_char2bytes(flags[i], p); |
| 495 | 932 *p = NUL; |
| 933 p = uflags; | |
| 934 } | |
| 935 else | |
| 936 p = flags; | |
| 6375 | 937 if (!vim_regexec_prog(&slang->sl_compprog, FALSE, p, 0)) |
| 483 | 938 return FALSE; |
| 939 | |
| 490 | 940 /* Count the number of syllables. This may be slow, do it last. If there |
| 941 * are too many syllables AND the number of compound words is above | |
| 809 | 942 * COMPOUNDWORDMAX then compounding is not allowed. */ |
| 483 | 943 if (slang->sl_compsylmax < MAXWLEN |
| 944 && count_syllables(slang, word) > slang->sl_compsylmax) | |
| 495 | 945 return (int)STRLEN(flags) < slang->sl_compmax; |
| 483 | 946 return TRUE; |
| 482 | 947 } |
| 948 | |
| 949 /* | |
| 1762 | 950 * Return TRUE if the compound flags in compflags[] match the start of any |
| 951 * compound rule. This is used to stop trying a compound if the flags | |
| 952 * collected so far can't possibly match any compound rule. | |
| 953 * Caller must check that slang->sl_comprules is not NULL. | |
| 954 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
955 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
956 match_compoundrule(slang_T *slang, char_u *compflags) |
| 1762 | 957 { |
| 958 char_u *p; | |
| 959 int i; | |
| 960 int c; | |
| 961 | |
| 962 /* loop over all the COMPOUNDRULE entries */ | |
| 963 for (p = slang->sl_comprules; *p != NUL; ++p) | |
| 964 { | |
| 965 /* loop over the flags in the compound word we have made, match | |
| 966 * them against the current rule entry */ | |
| 967 for (i = 0; ; ++i) | |
| 968 { | |
| 969 c = compflags[i]; | |
| 970 if (c == NUL) | |
| 971 /* found a rule that matches for the flags we have so far */ | |
| 972 return TRUE; | |
| 973 if (*p == '/' || *p == NUL) | |
| 974 break; /* end of rule, it's too short */ | |
| 975 if (*p == '[') | |
| 976 { | |
| 977 int match = FALSE; | |
| 978 | |
| 979 /* compare against all the flags in [] */ | |
| 980 ++p; | |
| 981 while (*p != ']' && *p != NUL) | |
| 982 if (*p++ == c) | |
| 983 match = TRUE; | |
| 984 if (!match) | |
| 985 break; /* none matches */ | |
| 986 } | |
| 987 else if (*p != c) | |
| 988 break; /* flag of word doesn't match flag in pattern */ | |
| 989 ++p; | |
| 990 } | |
| 991 | |
| 992 /* Skip to the next "/", where the next pattern starts. */ | |
| 993 p = vim_strchr(p, '/'); | |
| 994 if (p == NULL) | |
| 995 break; | |
| 996 } | |
| 997 | |
| 998 /* Checked all the rules and none of them match the flags, so there | |
| 999 * can't possibly be a compound starting with these flags. */ | |
| 1000 return FALSE; | |
| 1001 } | |
| 1002 | |
| 1003 /* | |
| 390 | 1004 * Return non-zero if the prefix indicated by "arridx" matches with the prefix |
| 1005 * ID in "flags" for the word "word". | |
| 366 | 1006 * The WF_RAREPFX flag is included in the return value for a rare prefix. |
| 351 | 1007 */ |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
1008 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1009 valid_word_prefix( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1010 int totprefcnt, /* nr of prefix IDs */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1011 int arridx, /* idx in sl_pidxs[] */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1012 int flags, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1013 char_u *word, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1014 slang_T *slang, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1015 int cond_req) /* only use prefixes with a condition */ |
| 351 | 1016 { |
| 1017 int prefcnt; | |
| 1018 int pidx; | |
| 6375 | 1019 regprog_T **rp; |
| 390 | 1020 int prefid; |
| 1021 | |
| 1022 prefid = (unsigned)flags >> 24; | |
| 351 | 1023 for (prefcnt = totprefcnt - 1; prefcnt >= 0; --prefcnt) |
| 1024 { | |
| 1025 pidx = slang->sl_pidxs[arridx + prefcnt]; | |
| 1026 | |
| 1027 /* Check the prefix ID. */ | |
| 1028 if (prefid != (pidx & 0xff)) | |
| 1029 continue; | |
| 1030 | |
| 390 | 1031 /* Check if the prefix doesn't combine and the word already has a |
| 1032 * suffix. */ | |
| 1033 if ((flags & WF_HAS_AFF) && (pidx & WF_PFX_NC)) | |
| 1034 continue; | |
| 1035 | |
| 351 | 1036 /* Check the condition, if there is one. The condition index is |
| 366 | 1037 * stored in the two bytes above the prefix ID byte. */ |
| 6375 | 1038 rp = &slang->sl_prefprog[((unsigned)pidx >> 8) & 0xffff]; |
| 1039 if (*rp != NULL) | |
| 1040 { | |
| 1041 if (!vim_regexec_prog(rp, FALSE, word, 0)) | |
| 351 | 1042 continue; |
| 1043 } | |
| 455 | 1044 else if (cond_req) |
| 1045 continue; | |
| 1046 | |
| 1047 /* It's a match! Return the WF_ flags. */ | |
| 366 | 1048 return pidx; |
| 1049 } | |
| 1050 return 0; | |
| 351 | 1051 } |
| 1052 | |
| 1053 /* | |
| 339 | 1054 * Check if the word at "mip->mi_word" has a matching prefix. |
| 1055 * If it does, then check the following word. | |
| 1056 * | |
| 485 | 1057 * If "mode" is "FIND_COMPOUND" then do the same after another word, find a |
| 1058 * prefix in a compound word. | |
| 1059 * | |
| 339 | 1060 * For a match mip->mi_result is updated. |
| 1061 */ | |
| 1062 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1063 find_prefix(matchinf_T *mip, int mode) |
| 339 | 1064 { |
| 1065 idx_T arridx = 0; | |
| 1066 int len; | |
| 1067 int wlen = 0; | |
| 1068 int flen; | |
| 1069 int c; | |
| 1070 char_u *ptr; | |
| 1071 idx_T lo, hi, m; | |
| 1072 slang_T *slang = mip->mi_lp->lp_slang; | |
| 1073 char_u *byts; | |
| 1074 idx_T *idxs; | |
| 1075 | |
| 375 | 1076 byts = slang->sl_pbyts; |
| 1077 if (byts == NULL) | |
| 1078 return; /* array is empty */ | |
| 1079 | |
| 339 | 1080 /* We use the case-folded word here, since prefixes are always |
| 1081 * case-folded. */ | |
| 1082 ptr = mip->mi_fword; | |
| 1083 flen = mip->mi_fwordlen; /* available case-folded bytes */ | |
| 485 | 1084 if (mode == FIND_COMPOUND) |
| 1085 { | |
| 1086 /* Skip over the previously found word(s). */ | |
| 1087 ptr += mip->mi_compoff; | |
| 1088 flen -= mip->mi_compoff; | |
| 1089 } | |
| 339 | 1090 idxs = slang->sl_pidxs; |
| 1091 | |
| 1092 /* | |
| 1093 * Repeat advancing in the tree until: | |
| 1094 * - there is a byte that doesn't match, | |
| 1095 * - we reach the end of the tree, | |
| 1096 * - or we reach the end of the line. | |
| 1097 */ | |
| 1098 for (;;) | |
| 1099 { | |
| 1100 if (flen == 0 && *mip->mi_fend != NUL) | |
| 1101 flen = fold_more(mip); | |
| 1102 | |
| 1103 len = byts[arridx++]; | |
| 1104 | |
| 1105 /* If the first possible byte is a zero the prefix could end here. | |
| 1106 * Check if the following word matches and supports the prefix. */ | |
| 1107 if (byts[arridx] == 0) | |
| 1108 { | |
| 1109 /* There can be several prefixes with different conditions. We | |
| 1110 * try them all, since we don't know which one will give the | |
| 1111 * longest match. The word is the same each time, pass the list | |
| 1112 * of possible prefixes to find_word(). */ | |
| 1113 mip->mi_prefarridx = arridx; | |
| 1114 mip->mi_prefcnt = len; | |
| 1115 while (len > 0 && byts[arridx] == 0) | |
| 1116 { | |
| 1117 ++arridx; | |
| 1118 --len; | |
| 1119 } | |
| 1120 mip->mi_prefcnt -= len; | |
| 1121 | |
| 1122 /* Find the word that comes after the prefix. */ | |
| 1123 mip->mi_prefixlen = wlen; | |
| 485 | 1124 if (mode == FIND_COMPOUND) |
| 1125 /* Skip over the previously found word(s). */ | |
| 1126 mip->mi_prefixlen += mip->mi_compoff; | |
| 1127 | |
| 455 | 1128 if (has_mbyte) |
| 1129 { | |
| 1130 /* Case-folded length may differ from original length. */ | |
| 485 | 1131 mip->mi_cprefixlen = nofold_len(mip->mi_fword, |
| 1132 mip->mi_prefixlen, mip->mi_word); | |
| 455 | 1133 } |
| 1134 else | |
| 485 | 1135 mip->mi_cprefixlen = mip->mi_prefixlen; |
| 339 | 1136 find_word(mip, FIND_PREFIX); |
| 1137 | |
| 1138 | |
| 1139 if (len == 0) | |
| 1140 break; /* no children, word must end here */ | |
| 1141 } | |
| 1142 | |
| 1143 /* Stop looking at end of the line. */ | |
| 1144 if (ptr[wlen] == NUL) | |
| 1145 break; | |
| 1146 | |
| 1147 /* Perform a binary search in the list of accepted bytes. */ | |
| 1148 c = ptr[wlen]; | |
| 1149 lo = arridx; | |
| 1150 hi = arridx + len - 1; | |
| 1151 while (lo < hi) | |
| 1152 { | |
| 1153 m = (lo + hi) / 2; | |
| 1154 if (byts[m] > c) | |
| 1155 hi = m - 1; | |
| 1156 else if (byts[m] < c) | |
| 1157 lo = m + 1; | |
| 1158 else | |
| 1159 { | |
| 1160 lo = hi = m; | |
| 1161 break; | |
| 1162 } | |
| 1163 } | |
| 1164 | |
| 1165 /* Stop if there is no matching byte. */ | |
| 1166 if (hi < lo || byts[lo] != c) | |
| 1167 break; | |
| 1168 | |
| 1169 /* Continue at the child (if there is one). */ | |
| 1170 arridx = idxs[lo]; | |
| 1171 ++wlen; | |
| 1172 --flen; | |
| 1173 } | |
| 1174 } | |
| 1175 | |
| 1176 /* | |
| 1177 * Need to fold at least one more character. Do until next non-word character | |
| 626 | 1178 * for efficiency. Include the non-word character too. |
| 339 | 1179 * Return the length of the folded chars in bytes. |
| 1180 */ | |
| 1181 static int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1182 fold_more(matchinf_T *mip) |
| 339 | 1183 { |
| 1184 int flen; | |
| 1185 char_u *p; | |
| 1186 | |
| 1187 p = mip->mi_fend; | |
| 1188 do | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
1189 MB_PTR_ADV(mip->mi_fend); |
|
16162
cd5c83115ec6
patch 8.1.1086: too many curly braces
Bram Moolenaar <Bram@vim.org>
parents:
16142
diff
changeset
|
1190 while (*mip->mi_fend != NUL && spell_iswordp(mip->mi_fend, mip->mi_win)); |
| 339 | 1191 |
| 626 | 1192 /* Include the non-word character so that we can check for the word end. */ |
| 339 | 1193 if (*mip->mi_fend != NUL) |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
1194 MB_PTR_ADV(mip->mi_fend); |
| 339 | 1195 |
| 1196 (void)spell_casefold(p, (int)(mip->mi_fend - p), | |
| 1197 mip->mi_fword + mip->mi_fwordlen, | |
| 1198 MAXWLEN - mip->mi_fwordlen); | |
| 835 | 1199 flen = (int)STRLEN(mip->mi_fword + mip->mi_fwordlen); |
| 339 | 1200 mip->mi_fwordlen += flen; |
| 1201 return flen; | |
| 1202 } | |
| 1203 | |
| 1204 /* | |
| 323 | 1205 * Check case flags for a word. Return TRUE if the word has the requested |
| 1206 * case. | |
| 1207 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
1208 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1209 spell_valid_case( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1210 int wordflags, /* flags for the checked word. */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1211 int treeflags) /* flags for the word in the spell tree */ |
| 323 | 1212 { |
| 389 | 1213 return ((wordflags == WF_ALLCAP && (treeflags & WF_FIXCAP) == 0) |
| 323 | 1214 || ((treeflags & (WF_ALLCAP | WF_KEEPCAP)) == 0 |
| 474 | 1215 && ((treeflags & WF_ONECAP) == 0 |
| 1216 || (wordflags & WF_ONECAP) != 0))); | |
| 323 | 1217 } |
| 1218 | |
| 351 | 1219 /* |
| 1220 * Return TRUE if spell checking is not enabled. | |
| 1221 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
1222 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1223 no_spell_checking(win_T *wp) |
| 498 | 1224 { |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
1225 if (!wp->w_p_spell || *wp->w_s->b_p_spl == NUL |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
1226 || wp->w_s->b_langp.ga_len == 0) |
| 351 | 1227 { |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
1228 emsg(_("E756: Spell checking is not enabled")); |
| 351 | 1229 return TRUE; |
| 1230 } | |
| 1231 return FALSE; | |
| 1232 } | |
| 300 | 1233 |
| 236 | 1234 /* |
| 1235 * Move to next spell error. | |
| 500 | 1236 * "curline" is FALSE for "[s", "]s", "[S" and "]S". |
| 1237 * "curline" is TRUE to find word under/after cursor in the same line. | |
| 483 | 1238 * For Insert mode completion "dir" is BACKWARD and "curline" is TRUE: move |
| 1239 * to after badly spelled word before the cursor. | |
| 495 | 1240 * Return 0 if not found, length of the badly spelled word otherwise. |
| 236 | 1241 */ |
| 1242 int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1243 spell_move_to( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1244 win_T *wp, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1245 int dir, /* FORWARD or BACKWARD */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1246 int allwords, /* TRUE for "[s"/"]s", FALSE for "[S"/"]S" */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1247 int curline, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1248 hlf_T *attrp) /* return: attributes of bad word or NULL |
| 534 | 1249 (only when "dir" is FORWARD) */ |
| 236 | 1250 { |
| 249 | 1251 linenr_T lnum; |
| 1252 pos_T found_pos; | |
| 495 | 1253 int found_len = 0; |
| 236 | 1254 char_u *line; |
| 1255 char_u *p; | |
| 346 | 1256 char_u *endp; |
| 534 | 1257 hlf_T attr; |
| 236 | 1258 int len; |
| 5519 | 1259 #ifdef FEAT_SYN_HL |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
1260 int has_syntax = syntax_present(wp); |
| 5519 | 1261 #endif |
| 249 | 1262 int col; |
| 1263 int can_spell; | |
| 346 | 1264 char_u *buf = NULL; |
| 1265 int buflen = 0; | |
| 1266 int skip = 0; | |
| 385 | 1267 int capcol = -1; |
| 500 | 1268 int found_one = FALSE; |
| 1269 int wrapped = FALSE; | |
| 236 | 1270 |
| 498 | 1271 if (no_spell_checking(wp)) |
| 495 | 1272 return 0; |
| 236 | 1273 |
| 249 | 1274 /* |
| 1275 * Start looking for bad word at the start of the line, because we can't | |
| 817 | 1276 * start halfway a word, we don't know where it starts or ends. |
| 249 | 1277 * |
| 1278 * When searching backwards, we continue in the line to find the last | |
| 1279 * bad word (in the cursor line: before the cursor). | |
| 346 | 1280 * |
| 1281 * We concatenate the start of the next line, so that wrapped words work | |
| 1282 * (e.g. "et<line-break>cetera"). Doesn't work when searching backwards | |
| 1283 * though... | |
| 249 | 1284 */ |
| 498 | 1285 lnum = wp->w_cursor.lnum; |
|
11121
778c10516955
patch 8.0.0448: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
10950
diff
changeset
|
1286 CLEAR_POS(&found_pos); |
| 236 | 1287 |
| 1288 while (!got_int) | |
| 1289 { | |
| 498 | 1290 line = ml_get_buf(wp->w_buffer, lnum, FALSE); |
| 346 | 1291 |
| 835 | 1292 len = (int)STRLEN(line); |
| 346 | 1293 if (buflen < len + MAXWLEN + 2) |
| 1294 { | |
| 1295 vim_free(buf); | |
| 1296 buflen = len + MAXWLEN + 2; | |
| 1297 buf = alloc(buflen); | |
| 1298 if (buf == NULL) | |
| 1299 break; | |
| 1300 } | |
| 1301 | |
| 385 | 1302 /* In first line check first word for Capital. */ |
| 1303 if (lnum == 1) | |
| 1304 capcol = 0; | |
| 1305 | |
| 1306 /* For checking first word with a capital skip white space. */ | |
| 1307 if (capcol == 0) | |
|
12323
4dba3e4f3b01
patch 8.0.1041: bogus characters when indenting during visual-block append
Christian Brabandt <cb@256bit.org>
parents:
11521
diff
changeset
|
1308 capcol = getwhitecols(line); |
| 835 | 1309 else if (curline && wp == curwin) |
| 1310 { | |
| 1311 /* For spellbadword(): check if first word needs a capital. */ | |
|
12323
4dba3e4f3b01
patch 8.0.1041: bogus characters when indenting during visual-block append
Christian Brabandt <cb@256bit.org>
parents:
11521
diff
changeset
|
1312 col = getwhitecols(line); |
| 835 | 1313 if (check_need_cap(lnum, col)) |
| 1314 capcol = col; | |
| 1315 | |
| 1316 /* Need to get the line again, may have looked at the previous | |
| 1317 * one. */ | |
| 1318 line = ml_get_buf(wp->w_buffer, lnum, FALSE); | |
| 1319 } | |
| 385 | 1320 |
| 346 | 1321 /* Copy the line into "buf" and append the start of the next line if |
| 1322 * possible. */ | |
| 1323 STRCPY(buf, line); | |
| 498 | 1324 if (lnum < wp->w_buffer->b_ml.ml_line_count) |
| 883 | 1325 spell_cat_line(buf + STRLEN(buf), |
| 1326 ml_get_buf(wp->w_buffer, lnum + 1, FALSE), MAXWLEN); | |
| 346 | 1327 |
| 1328 p = buf + skip; | |
| 1329 endp = buf + len; | |
| 1330 while (p < endp) | |
| 236 | 1331 { |
| 500 | 1332 /* When searching backward don't search after the cursor. Unless |
| 1333 * we wrapped around the end of the buffer. */ | |
| 300 | 1334 if (dir == BACKWARD |
| 498 | 1335 && lnum == wp->w_cursor.lnum |
| 500 | 1336 && !wrapped |
| 498 | 1337 && (colnr_T)(p - buf) >= wp->w_cursor.col) |
| 300 | 1338 break; |
| 249 | 1339 |
| 300 | 1340 /* start of word */ |
| 534 | 1341 attr = HLF_COUNT; |
| 625 | 1342 len = spell_check(wp, p, &attr, &capcol, FALSE); |
| 249 | 1343 |
| 534 | 1344 if (attr != HLF_COUNT) |
| 300 | 1345 { |
| 1346 /* We found a bad word. Check the attribute. */ | |
| 534 | 1347 if (allwords || attr == HLF_SPB) |
| 236 | 1348 { |
| 300 | 1349 /* When searching forward only accept a bad word after |
| 1350 * the cursor. */ | |
| 1351 if (dir == BACKWARD | |
| 500 | 1352 || lnum != wp->w_cursor.lnum |
| 498 | 1353 || (lnum == wp->w_cursor.lnum |
| 500 | 1354 && (wrapped |
| 1355 || (colnr_T)(curline ? p - buf + len | |
| 346 | 1356 : p - buf) |
| 500 | 1357 > wp->w_cursor.col))) |
| 236 | 1358 { |
| 5519 | 1359 #ifdef FEAT_SYN_HL |
| 300 | 1360 if (has_syntax) |
| 249 | 1361 { |
| 835 | 1362 col = (int)(p - buf); |
| 498 | 1363 (void)syn_get_id(wp, lnum, (colnr_T)col, |
| 1504 | 1364 FALSE, &can_spell, FALSE); |
| 857 | 1365 if (!can_spell) |
| 1366 attr = HLF_COUNT; | |
| 300 | 1367 } |
| 1368 else | |
| 737 | 1369 #endif |
| 300 | 1370 can_spell = TRUE; |
| 249 | 1371 |
| 300 | 1372 if (can_spell) |
| 1373 { | |
| 857 | 1374 found_one = TRUE; |
| 300 | 1375 found_pos.lnum = lnum; |
| 835 | 1376 found_pos.col = (int)(p - buf); |
| 300 | 1377 found_pos.coladd = 0; |
| 1378 if (dir == FORWARD) | |
| 1379 { | |
| 1380 /* No need to search further. */ | |
| 498 | 1381 wp->w_cursor = found_pos; |
| 346 | 1382 vim_free(buf); |
| 498 | 1383 if (attrp != NULL) |
| 1384 *attrp = attr; | |
| 495 | 1385 return len; |
| 249 | 1386 } |
| 483 | 1387 else if (curline) |
| 1388 /* Insert mode completion: put cursor after | |
| 1389 * the bad word. */ | |
| 1390 found_pos.col += len; | |
| 495 | 1391 found_len = len; |
| 249 | 1392 } |
| 236 | 1393 } |
| 857 | 1394 else |
| 1395 found_one = TRUE; | |
| 236 | 1396 } |
| 1397 } | |
| 1398 | |
| 300 | 1399 /* advance to character after the word */ |
| 1400 p += len; | |
| 385 | 1401 capcol -= len; |
| 236 | 1402 } |
| 1403 | |
| 483 | 1404 if (dir == BACKWARD && found_pos.lnum != 0) |
| 1405 { | |
| 500 | 1406 /* Use the last match in the line (before the cursor). */ |
| 498 | 1407 wp->w_cursor = found_pos; |
| 483 | 1408 vim_free(buf); |
| 495 | 1409 return found_len; |
| 483 | 1410 } |
| 1411 | |
| 323 | 1412 if (curline) |
| 346 | 1413 break; /* only check cursor line */ |
| 323 | 1414 |
|
10950
2297aae8e127
patch 8.0.0364: ]s does not move cursor with two spell errors in one line
Christian Brabandt <cb@256bit.org>
parents:
10359
diff
changeset
|
1415 /* If we are back at the starting line and searched it again there |
|
2297aae8e127
patch 8.0.0364: ]s does not move cursor with two spell errors in one line
Christian Brabandt <cb@256bit.org>
parents:
10359
diff
changeset
|
1416 * is no match, give up. */ |
|
2297aae8e127
patch 8.0.0364: ]s does not move cursor with two spell errors in one line
Christian Brabandt <cb@256bit.org>
parents:
10359
diff
changeset
|
1417 if (lnum == wp->w_cursor.lnum && wrapped) |
|
2297aae8e127
patch 8.0.0364: ]s does not move cursor with two spell errors in one line
Christian Brabandt <cb@256bit.org>
parents:
10359
diff
changeset
|
1418 break; |
|
2297aae8e127
patch 8.0.0364: ]s does not move cursor with two spell errors in one line
Christian Brabandt <cb@256bit.org>
parents:
10359
diff
changeset
|
1419 |
| 236 | 1420 /* Advance to next line. */ |
| 249 | 1421 if (dir == BACKWARD) |
| 1422 { | |
| 500 | 1423 if (lnum > 1) |
| 1424 --lnum; | |
| 1425 else if (!p_ws) | |
| 1426 break; /* at first line and 'nowrapscan' */ | |
| 1427 else | |
| 1428 { | |
| 1429 /* Wrap around to the end of the buffer. May search the | |
| 1430 * starting line again and accept the last match. */ | |
| 1431 lnum = wp->w_buffer->b_ml.ml_line_count; | |
| 1432 wrapped = TRUE; | |
| 503 | 1433 if (!shortmess(SHM_SEARCH)) |
| 1434 give_warning((char_u *)_(top_bot_msg), TRUE); | |
| 500 | 1435 } |
| 385 | 1436 capcol = -1; |
| 249 | 1437 } |
| 1438 else | |
| 1439 { | |
| 500 | 1440 if (lnum < wp->w_buffer->b_ml.ml_line_count) |
| 1441 ++lnum; | |
| 1442 else if (!p_ws) | |
| 1443 break; /* at first line and 'nowrapscan' */ | |
| 1444 else | |
| 1445 { | |
| 1446 /* Wrap around to the start of the buffer. May search the | |
| 1447 * starting line again and accept the first match. */ | |
| 1448 lnum = 1; | |
| 1449 wrapped = TRUE; | |
| 503 | 1450 if (!shortmess(SHM_SEARCH)) |
| 1451 give_warning((char_u *)_(bot_top_msg), TRUE); | |
| 500 | 1452 } |
| 1453 | |
| 1454 /* If we are back at the starting line and there is no match then | |
| 1455 * give up. */ | |
|
10950
2297aae8e127
patch 8.0.0364: ]s does not move cursor with two spell errors in one line
Christian Brabandt <cb@256bit.org>
parents:
10359
diff
changeset
|
1456 if (lnum == wp->w_cursor.lnum && !found_one) |
| 346 | 1457 break; |
| 1458 | |
| 1459 /* Skip the characters at the start of the next line that were | |
| 1460 * included in a match crossing line boundaries. */ | |
| 534 | 1461 if (attr == HLF_COUNT) |
| 835 | 1462 skip = (int)(p - endp); |
| 346 | 1463 else |
| 1464 skip = 0; | |
| 385 | 1465 |
| 835 | 1466 /* Capcol skips over the inserted space. */ |
| 385 | 1467 --capcol; |
| 1468 | |
| 1469 /* But after empty line check first word in next line */ | |
| 1470 if (*skipwhite(line) == NUL) | |
| 1471 capcol = 0; | |
| 249 | 1472 } |
| 236 | 1473 |
| 1474 line_breakcheck(); | |
| 1475 } | |
| 1476 | |
| 346 | 1477 vim_free(buf); |
| 495 | 1478 return 0; |
| 346 | 1479 } |
| 1480 | |
| 1481 /* | |
| 1482 * For spell checking: concatenate the start of the following line "line" into | |
| 1483 * "buf", blanking-out special characters. Copy less then "maxlen" bytes. | |
| 1577 | 1484 * Keep the blanks at the start of the next line, this is used in win_line() |
| 1485 * to skip those bytes if the word was OK. | |
| 346 | 1486 */ |
| 1487 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1488 spell_cat_line(char_u *buf, char_u *line, int maxlen) |
| 346 | 1489 { |
| 1490 char_u *p; | |
| 1491 int n; | |
| 1492 | |
| 1493 p = skipwhite(line); | |
| 1494 while (vim_strchr((char_u *)"*#/\"\t", *p) != NULL) | |
| 1495 p = skipwhite(p + 1); | |
| 1496 | |
| 1497 if (*p != NUL) | |
| 1498 { | |
| 1577 | 1499 /* Only worth concatenating if there is something else than spaces to |
| 1500 * concatenate. */ | |
| 1501 n = (int)(p - line) + 1; | |
| 1502 if (n < maxlen - 1) | |
| 1503 { | |
| 1504 vim_memset(buf, ' ', n); | |
| 1505 vim_strncpy(buf + n, p, maxlen - 1 - n); | |
| 1506 } | |
| 346 | 1507 } |
| 236 | 1508 } |
| 1509 | |
| 626 | 1510 /* |
| 1511 * Structure used for the cookie argument of do_in_runtimepath(). | |
| 1512 */ | |
| 501 | 1513 typedef struct spelload_S |
| 1514 { | |
| 1515 char_u sl_lang[MAXWLEN + 1]; /* language name */ | |
| 1516 slang_T *sl_slang; /* resulting slang_T struct */ | |
| 1517 int sl_nobreak; /* NOBREAK language found */ | |
| 1518 } spelload_T; | |
| 1519 | |
| 236 | 1520 /* |
| 307 | 1521 * Load word list(s) for "lang" from Vim spell file(s). |
| 310 | 1522 * "lang" must be the language without the region: e.g., "en". |
| 236 | 1523 */ |
| 307 | 1524 static void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1525 spell_load_lang(char_u *lang) |
| 236 | 1526 { |
| 310 | 1527 char_u fname_enc[85]; |
| 236 | 1528 int r; |
| 501 | 1529 spelload_T sl; |
| 649 | 1530 int round; |
| 307 | 1531 |
| 310 | 1532 /* Copy the language name to pass it to spell_load_cb() as a cookie. |
| 307 | 1533 * It's truncated when an error is detected. */ |
| 501 | 1534 STRCPY(sl.sl_lang, lang); |
| 1535 sl.sl_slang = NULL; | |
| 1536 sl.sl_nobreak = FALSE; | |
| 307 | 1537 |
| 649 | 1538 /* We may retry when no spell file is found for the language, an |
| 1539 * autocommand may load it then. */ | |
| 1540 for (round = 1; round <= 2; ++round) | |
| 1541 { | |
| 1542 /* | |
| 1543 * Find the first spell file for "lang" in 'runtimepath' and load it. | |
| 1544 */ | |
| 1545 vim_snprintf((char *)fname_enc, sizeof(fname_enc) - 5, | |
| 2660 | 1546 #ifdef VMS |
| 1547 "spell/%s_%s.spl", | |
| 1548 #else | |
| 1549 "spell/%s.%s.spl", | |
| 1550 #endif | |
| 1551 lang, spell_enc()); | |
|
8524
2f57bbe870ea
commit https://github.com/vim/vim/commit/7f8989dd8a627af2185df381195351a913f3777f
Christian Brabandt <cb@256bit.org>
parents:
7835
diff
changeset
|
1552 r = do_in_runtimepath(fname_enc, 0, spell_load_cb, &sl); |
| 649 | 1553 |
| 1554 if (r == FAIL && *sl.sl_lang != NUL) | |
| 1555 { | |
| 1556 /* Try loading the ASCII version. */ | |
| 1557 vim_snprintf((char *)fname_enc, sizeof(fname_enc) - 5, | |
| 2660 | 1558 #ifdef VMS |
| 1559 "spell/%s_ascii.spl", | |
| 1560 #else | |
| 1561 "spell/%s.ascii.spl", | |
| 1562 #endif | |
| 1563 lang); | |
|
8524
2f57bbe870ea
commit https://github.com/vim/vim/commit/7f8989dd8a627af2185df381195351a913f3777f
Christian Brabandt <cb@256bit.org>
parents:
7835
diff
changeset
|
1564 r = do_in_runtimepath(fname_enc, 0, spell_load_cb, &sl); |
| 649 | 1565 |
| 1566 if (r == FAIL && *sl.sl_lang != NUL && round == 1 | |
| 1567 && apply_autocmds(EVENT_SPELLFILEMISSING, lang, | |
| 1568 curbuf->b_fname, FALSE, curbuf)) | |
| 1569 continue; | |
| 1570 break; | |
|
13380
69517d67421f
patch 8.0.1564: too many #ifdefs
Christian Brabandt <cb@256bit.org>
parents:
13308
diff
changeset
|
1571 } |
| 714 | 1572 break; |
| 307 | 1573 } |
| 1574 | |
| 1575 if (r == FAIL) | |
| 649 | 1576 { |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
1577 smsg( |
| 2660 | 1578 #ifdef VMS |
| 1579 _("Warning: Cannot find word list \"%s_%s.spl\" or \"%s_ascii.spl\""), | |
| 1580 #else | |
| 1581 _("Warning: Cannot find word list \"%s.%s.spl\" or \"%s.ascii.spl\""), | |
| 1582 #endif | |
| 483 | 1583 lang, spell_enc(), lang); |
| 649 | 1584 } |
| 501 | 1585 else if (sl.sl_slang != NULL) |
| 1586 { | |
| 626 | 1587 /* At least one file was loaded, now load ALL the additions. */ |
| 310 | 1588 STRCPY(fname_enc + STRLEN(fname_enc) - 3, "add.spl"); |
|
8524
2f57bbe870ea
commit https://github.com/vim/vim/commit/7f8989dd8a627af2185df381195351a913f3777f
Christian Brabandt <cb@256bit.org>
parents:
7835
diff
changeset
|
1589 do_in_runtimepath(fname_enc, DIP_ALL, spell_load_cb, &sl); |
| 310 | 1590 } |
| 1591 } | |
| 1592 | |
| 1593 /* | |
| 1594 * Return the encoding used for spell checking: Use 'encoding', except that we | |
| 1595 * use "latin1" for "latin9". And limit to 60 characters (just in case). | |
| 1596 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1597 char_u * |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1598 spell_enc(void) |
| 310 | 1599 { |
| 1600 | |
| 1601 if (STRLEN(p_enc) < 60 && STRCMP(p_enc, "iso-8859-15") != 0) | |
| 1602 return p_enc; | |
| 1603 return (char_u *)"latin1"; | |
| 236 | 1604 } |
| 1605 | |
| 1606 /* | |
| 385 | 1607 * Get the name of the .spl file for the internal wordlist into |
| 1608 * "fname[MAXPATHL]". | |
| 1609 */ | |
| 1610 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1611 int_wordlist_spl(char_u *fname) |
| 385 | 1612 { |
| 2660 | 1613 vim_snprintf((char *)fname, MAXPATHL, SPL_FNAME_TMPL, |
| 385 | 1614 int_wordlist, spell_enc()); |
| 1615 } | |
| 1616 | |
| 1617 /* | |
| 625 | 1618 * Allocate a new slang_T for language "lang". "lang" can be NULL. |
| 236 | 1619 * Caller must fill "sl_next". |
| 1620 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1621 slang_T * |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1622 slang_alloc(char_u *lang) |
| 236 | 1623 { |
| 1624 slang_T *lp; | |
| 1625 | |
|
16825
ce04ebdf26b8
patch 8.1.1414: alloc() returning "char_u *" causes a lot of type casts
Bram Moolenaar <Bram@vim.org>
parents:
16768
diff
changeset
|
1626 lp = ALLOC_CLEAR_ONE(slang_T); |
| 236 | 1627 if (lp != NULL) |
| 1628 { | |
| 625 | 1629 if (lang != NULL) |
| 1630 lp->sl_name = vim_strsave(lang); | |
| 323 | 1631 ga_init2(&lp->sl_rep, sizeof(fromto_T), 10); |
| 625 | 1632 ga_init2(&lp->sl_repsal, sizeof(fromto_T), 10); |
| 483 | 1633 lp->sl_compmax = MAXWLEN; |
| 1634 lp->sl_compsylmax = MAXWLEN; | |
| 625 | 1635 hash_init(&lp->sl_wordcount); |
| 1636 } | |
| 1637 | |
| 236 | 1638 return lp; |
| 1639 } | |
| 1640 | |
| 1641 /* | |
| 1642 * Free the contents of an slang_T and the structure itself. | |
| 1643 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1644 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1645 slang_free(slang_T *lp) |
| 236 | 1646 { |
| 1647 vim_free(lp->sl_name); | |
| 310 | 1648 vim_free(lp->sl_fname); |
| 1649 slang_clear(lp); | |
| 1650 vim_free(lp); | |
| 1651 } | |
| 1652 | |
| 1653 /* | |
| 1654 * Clear an slang_T so that the file can be reloaded. | |
| 1655 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1656 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1657 slang_clear(slang_T *lp) |
| 310 | 1658 { |
| 339 | 1659 garray_T *gap; |
| 1660 fromto_T *ftp; | |
| 344 | 1661 salitem_T *smp; |
| 339 | 1662 int i; |
| 625 | 1663 int round; |
| 323 | 1664 |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1665 VIM_CLEAR(lp->sl_fbyts); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1666 VIM_CLEAR(lp->sl_kbyts); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1667 VIM_CLEAR(lp->sl_pbyts); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1668 |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1669 VIM_CLEAR(lp->sl_fidxs); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1670 VIM_CLEAR(lp->sl_kidxs); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1671 VIM_CLEAR(lp->sl_pidxs); |
| 323 | 1672 |
| 625 | 1673 for (round = 1; round <= 2; ++round) |
| 1674 { | |
| 1675 gap = round == 1 ? &lp->sl_rep : &lp->sl_repsal; | |
| 1676 while (gap->ga_len > 0) | |
| 1677 { | |
| 1678 ftp = &((fromto_T *)gap->ga_data)[--gap->ga_len]; | |
| 1679 vim_free(ftp->ft_from); | |
| 1680 vim_free(ftp->ft_to); | |
| 1681 } | |
| 1682 ga_clear(gap); | |
| 1683 } | |
| 344 | 1684 |
| 1685 gap = &lp->sl_sal; | |
| 375 | 1686 if (lp->sl_sofo) |
| 376 | 1687 { |
| 1688 /* "ga_len" is set to 1 without adding an item for latin1 */ | |
| 1689 if (gap->ga_data != NULL) | |
| 1690 /* SOFOFROM and SOFOTO items: free lists of wide characters. */ | |
| 1691 for (i = 0; i < gap->ga_len; ++i) | |
| 1692 vim_free(((int **)gap->ga_data)[i]); | |
| 1693 } | |
| 375 | 1694 else |
| 1695 /* SAL items: free salitem_T items */ | |
| 1696 while (gap->ga_len > 0) | |
| 1697 { | |
| 1698 smp = &((salitem_T *)gap->ga_data)[--gap->ga_len]; | |
| 1699 vim_free(smp->sm_lead); | |
| 1700 /* Don't free sm_oneof and sm_rules, they point into sm_lead. */ | |
| 1701 vim_free(smp->sm_to); | |
| 1702 vim_free(smp->sm_lead_w); | |
| 1703 vim_free(smp->sm_oneof_w); | |
| 1704 vim_free(smp->sm_to_w); | |
| 1705 } | |
| 344 | 1706 ga_clear(gap); |
| 323 | 1707 |
| 339 | 1708 for (i = 0; i < lp->sl_prefixcnt; ++i) |
|
4805
66803af09906
updated for version 7.3.1149
Bram Moolenaar <bram@vim.org>
parents:
4352
diff
changeset
|
1709 vim_regfree(lp->sl_prefprog[i]); |
| 376 | 1710 lp->sl_prefixcnt = 0; |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1711 VIM_CLEAR(lp->sl_prefprog); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1712 |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1713 VIM_CLEAR(lp->sl_info); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1714 |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1715 VIM_CLEAR(lp->sl_midword); |
| 339 | 1716 |
|
4805
66803af09906
updated for version 7.3.1149
Bram Moolenaar <bram@vim.org>
parents:
4352
diff
changeset
|
1717 vim_regfree(lp->sl_compprog); |
| 483 | 1718 lp->sl_compprog = NULL; |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1719 VIM_CLEAR(lp->sl_comprules); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1720 VIM_CLEAR(lp->sl_compstartflags); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1721 VIM_CLEAR(lp->sl_compallflags); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1722 |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1723 VIM_CLEAR(lp->sl_syllable); |
| 483 | 1724 ga_clear(&lp->sl_syl_items); |
| 481 | 1725 |
| 809 | 1726 ga_clear_strings(&lp->sl_comppat); |
| 1727 | |
| 625 | 1728 hash_clear_all(&lp->sl_wordcount, WC_KEY_OFF); |
| 1729 hash_init(&lp->sl_wordcount); | |
| 1730 | |
| 1731 hash_clear_all(&lp->sl_map_hash, 0); | |
| 1732 | |
| 1733 /* Clear info from .sug file. */ | |
| 1734 slang_clear_sug(lp); | |
| 483 | 1735 |
| 1736 lp->sl_compmax = MAXWLEN; | |
| 501 | 1737 lp->sl_compminlen = 0; |
| 483 | 1738 lp->sl_compsylmax = MAXWLEN; |
| 1739 lp->sl_regions[0] = NUL; | |
| 236 | 1740 } |
| 1741 | |
| 1742 /* | |
| 625 | 1743 * Clear the info from the .sug file in "lp". |
| 1744 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1745 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1746 slang_clear_sug(slang_T *lp) |
| 625 | 1747 { |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1748 VIM_CLEAR(lp->sl_sbyts); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
1749 VIM_CLEAR(lp->sl_sidxs); |
| 625 | 1750 close_spellbuf(lp->sl_sugbuf); |
| 1751 lp->sl_sugbuf = NULL; | |
| 1752 lp->sl_sugloaded = FALSE; | |
| 1753 lp->sl_sugtime = 0; | |
| 1754 } | |
| 1755 | |
| 1756 /* | |
| 307 | 1757 * Load one spell file and store the info into a slang_T. |
| 236 | 1758 * Invoked through do_in_runtimepath(). |
| 1759 */ | |
| 1760 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1761 spell_load_cb(char_u *fname, void *cookie) |
| 501 | 1762 { |
| 1763 spelload_T *slp = (spelload_T *)cookie; | |
| 1764 slang_T *slang; | |
| 1765 | |
| 1766 slang = spell_load_file(fname, slp->sl_lang, NULL, FALSE); | |
| 1767 if (slang != NULL) | |
| 1768 { | |
| 1769 /* When a previously loaded file has NOBREAK also use it for the | |
| 1770 * ".add" files. */ | |
| 1771 if (slp->sl_nobreak && slang->sl_add) | |
| 1772 slang->sl_nobreak = TRUE; | |
| 1773 else if (slang->sl_nobreak) | |
| 1774 slp->sl_nobreak = TRUE; | |
| 1775 | |
| 1776 slp->sl_slang = slang; | |
| 1777 } | |
| 310 | 1778 } |
| 1779 | |
| 625 | 1780 |
| 1781 /* | |
| 1782 * Add a word to the hashtable of common words. | |
| 1783 * If it's already there then the counter is increased. | |
| 1784 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1785 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1786 count_common_word( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1787 slang_T *lp, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1788 char_u *word, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1789 int len, /* word length, -1 for upto NUL */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1790 int count) /* 1 to count once, 10 to init */ |
| 625 | 1791 { |
| 1792 hash_T hash; | |
| 1793 hashitem_T *hi; | |
| 1794 wordcount_T *wc; | |
| 1795 char_u buf[MAXWLEN]; | |
| 1796 char_u *p; | |
| 1797 | |
| 1798 if (len == -1) | |
| 1799 p = word; | |
|
17653
cc68aca87c17
patch 8.1.1824: crash when correctly spelled word is very long
Bram Moolenaar <Bram@vim.org>
parents:
17434
diff
changeset
|
1800 else if (len >= MAXWLEN) |
|
cc68aca87c17
patch 8.1.1824: crash when correctly spelled word is very long
Bram Moolenaar <Bram@vim.org>
parents:
17434
diff
changeset
|
1801 return; |
| 625 | 1802 else |
| 1803 { | |
| 1804 vim_strncpy(buf, word, len); | |
| 1805 p = buf; | |
| 1806 } | |
| 1807 | |
| 1808 hash = hash_hash(p); | |
| 1809 hi = hash_lookup(&lp->sl_wordcount, p, hash); | |
| 1810 if (HASHITEM_EMPTY(hi)) | |
| 1811 { | |
|
16825
ce04ebdf26b8
patch 8.1.1414: alloc() returning "char_u *" causes a lot of type casts
Bram Moolenaar <Bram@vim.org>
parents:
16768
diff
changeset
|
1812 wc = alloc(sizeof(wordcount_T) + STRLEN(p)); |
| 625 | 1813 if (wc == NULL) |
| 1814 return; | |
| 1815 STRCPY(wc->wc_word, p); | |
| 1816 wc->wc_count = count; | |
| 1817 hash_add_item(&lp->sl_wordcount, hi, wc->wc_word, hash); | |
| 1818 } | |
| 1819 else | |
| 1820 { | |
| 1821 wc = HI2WC(hi); | |
| 1822 if ((wc->wc_count += count) < (unsigned)count) /* check for overflow */ | |
| 1823 wc->wc_count = MAXWORDCOUNT; | |
| 1824 } | |
| 1825 } | |
| 1826 | |
| 1827 /* | |
| 498 | 1828 * Return TRUE if byte "n" appears in "str". |
| 495 | 1829 * Like strchr() but independent of locale. |
| 1830 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1831 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1832 byte_in_str(char_u *str, int n) |
| 495 | 1833 { |
| 1834 char_u *p; | |
| 1835 | |
| 1836 for (p = str; *p != NUL; ++p) | |
| 498 | 1837 if (*p == n) |
| 495 | 1838 return TRUE; |
| 1839 return FALSE; | |
| 1840 } | |
| 1841 | |
| 483 | 1842 #define SY_MAXLEN 30 |
| 1843 typedef struct syl_item_S | |
| 1844 { | |
| 1845 char_u sy_chars[SY_MAXLEN]; /* the sequence of chars */ | |
| 1846 int sy_len; | |
| 1847 } syl_item_T; | |
| 1848 | |
| 1849 /* | |
| 1850 * Truncate "slang->sl_syllable" at the first slash and put the following items | |
| 1851 * in "slang->sl_syl_items". | |
| 1852 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
1853 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1854 init_syl_tab(slang_T *slang) |
| 483 | 1855 { |
| 1856 char_u *p; | |
| 1857 char_u *s; | |
| 1858 int l; | |
| 1859 syl_item_T *syl; | |
| 1860 | |
| 1861 ga_init2(&slang->sl_syl_items, sizeof(syl_item_T), 4); | |
| 1862 p = vim_strchr(slang->sl_syllable, '/'); | |
| 1863 while (p != NULL) | |
| 1864 { | |
| 1865 *p++ = NUL; | |
| 495 | 1866 if (*p == NUL) /* trailing slash */ |
| 483 | 1867 break; |
| 1868 s = p; | |
| 1869 p = vim_strchr(p, '/'); | |
| 1870 if (p == NULL) | |
| 835 | 1871 l = (int)STRLEN(s); |
| 483 | 1872 else |
| 835 | 1873 l = (int)(p - s); |
| 483 | 1874 if (l >= SY_MAXLEN) |
| 1875 return SP_FORMERROR; | |
| 1876 if (ga_grow(&slang->sl_syl_items, 1) == FAIL) | |
| 495 | 1877 return SP_OTHERERROR; |
| 483 | 1878 syl = ((syl_item_T *)slang->sl_syl_items.ga_data) |
| 1879 + slang->sl_syl_items.ga_len++; | |
| 1880 vim_strncpy(syl->sy_chars, s, l); | |
| 1881 syl->sy_len = l; | |
| 1882 } | |
| 1883 return OK; | |
| 1884 } | |
| 1885 | |
| 1886 /* | |
| 1887 * Count the number of syllables in "word". | |
| 1888 * When "word" contains spaces the syllables after the last space are counted. | |
| 1889 * Returns zero if syllables are not defines. | |
| 1890 */ | |
| 1891 static int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1892 count_syllables(slang_T *slang, char_u *word) |
| 483 | 1893 { |
| 1894 int cnt = 0; | |
| 1895 int skip = FALSE; | |
| 1896 char_u *p; | |
| 1897 int len; | |
| 1898 int i; | |
| 1899 syl_item_T *syl; | |
| 1900 int c; | |
| 1901 | |
| 1902 if (slang->sl_syllable == NULL) | |
| 1903 return 0; | |
| 1904 | |
| 1905 for (p = word; *p != NUL; p += len) | |
| 1906 { | |
| 1907 /* When running into a space reset counter. */ | |
| 1908 if (*p == ' ') | |
| 1909 { | |
| 1910 len = 1; | |
| 1911 cnt = 0; | |
| 1912 continue; | |
| 1913 } | |
| 1914 | |
| 1915 /* Find longest match of syllable items. */ | |
| 1916 len = 0; | |
| 1917 for (i = 0; i < slang->sl_syl_items.ga_len; ++i) | |
| 1918 { | |
| 1919 syl = ((syl_item_T *)slang->sl_syl_items.ga_data) + i; | |
| 1920 if (syl->sy_len > len | |
| 1921 && STRNCMP(p, syl->sy_chars, syl->sy_len) == 0) | |
| 1922 len = syl->sy_len; | |
| 1923 } | |
| 1924 if (len != 0) /* found a match, count syllable */ | |
| 1925 { | |
| 1926 ++cnt; | |
| 1927 skip = FALSE; | |
| 1928 } | |
| 1929 else | |
| 1930 { | |
| 1931 /* No recognized syllable item, at least a syllable char then? */ | |
| 1932 c = mb_ptr2char(p); | |
| 1933 len = (*mb_ptr2len)(p); | |
| 1934 if (vim_strchr(slang->sl_syllable, c) == NULL) | |
| 1935 skip = FALSE; /* No, search for next syllable */ | |
| 1936 else if (!skip) | |
| 1937 { | |
| 1938 ++cnt; /* Yes, count it */ | |
| 1939 skip = TRUE; /* don't count following syllable chars */ | |
| 1940 } | |
| 1941 } | |
| 1942 } | |
| 1943 return cnt; | |
| 376 | 1944 } |
| 1945 | |
| 381 | 1946 /* |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
1947 * Parse 'spelllang' and set w_s->b_langp accordingly. |
| 351 | 1948 * Returns NULL if it's OK, an error message otherwise. |
| 236 | 1949 */ |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
1950 char * |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
1951 did_set_spelllang(win_T *wp) |
| 236 | 1952 { |
| 1953 garray_T ga; | |
| 351 | 1954 char_u *splp; |
| 236 | 1955 char_u *region; |
| 416 | 1956 char_u region_cp[3]; |
| 355 | 1957 int filename; |
| 236 | 1958 int region_mask; |
| 503 | 1959 slang_T *slang; |
| 236 | 1960 int c; |
| 351 | 1961 char_u lang[MAXWLEN + 1]; |
| 323 | 1962 char_u spf_name[MAXPATHL]; |
| 351 | 1963 int len; |
| 1964 char_u *p; | |
| 381 | 1965 int round; |
| 385 | 1966 char_u *spf; |
| 389 | 1967 char_u *use_region = NULL; |
| 1968 int dont_use_region = FALSE; | |
| 501 | 1969 int nobreak = FALSE; |
| 503 | 1970 int i, j; |
| 1971 langp_T *lp, *lp2; | |
| 1185 | 1972 static int recursive = FALSE; |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
1973 char *ret_msg = NULL; |
| 1185 | 1974 char_u *spl_copy; |
|
9487
69ed2c9d34a6
commit https://github.com/vim/vim/commit/7c0a2f367f2507669560b1a66423155c70d2e75b
Christian Brabandt <cb@256bit.org>
parents:
9387
diff
changeset
|
1975 bufref_T bufref; |
|
69ed2c9d34a6
commit https://github.com/vim/vim/commit/7c0a2f367f2507669560b1a66423155c70d2e75b
Christian Brabandt <cb@256bit.org>
parents:
9387
diff
changeset
|
1976 |
|
69ed2c9d34a6
commit https://github.com/vim/vim/commit/7c0a2f367f2507669560b1a66423155c70d2e75b
Christian Brabandt <cb@256bit.org>
parents:
9387
diff
changeset
|
1977 set_bufref(&bufref, wp->w_buffer); |
| 1185 | 1978 |
| 1979 /* We don't want to do this recursively. May happen when a language is | |
| 1980 * not available and the SpellFileMissing autocommand opens a new buffer | |
| 1981 * in which 'spell' is set. */ | |
| 1982 if (recursive) | |
| 1983 return NULL; | |
| 1984 recursive = TRUE; | |
| 236 | 1985 |
| 1986 ga_init2(&ga, sizeof(langp_T), 2); | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
1987 clear_midword(wp); |
| 236 | 1988 |
| 4352 | 1989 /* Make a copy of 'spelllang', the SpellFileMissing autocommands may change |
| 1185 | 1990 * it under our fingers. */ |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
1991 spl_copy = vim_strsave(wp->w_s->b_p_spl); |
| 1185 | 1992 if (spl_copy == NULL) |
| 1993 goto theend; | |
| 1994 | |
| 5477 | 1995 wp->w_s->b_cjk = 0; |
| 1996 | |
| 1997 /* Loop over comma separated language names. */ | |
| 1185 | 1998 for (splp = spl_copy; *splp != NUL; ) |
| 351 | 1999 { |
|
16277
5ef25fa57f71
patch 8.1.1143: may pass weird strings to file name expansion
Bram Moolenaar <Bram@vim.org>
parents:
16239
diff
changeset
|
2000 // Get one language name. |
| 351 | 2001 copy_option_part(&splp, lang, MAXWLEN, ","); |
| 240 | 2002 region = NULL; |
| 835 | 2003 len = (int)STRLEN(lang); |
| 355 | 2004 |
|
16277
5ef25fa57f71
patch 8.1.1143: may pass weird strings to file name expansion
Bram Moolenaar <Bram@vim.org>
parents:
16239
diff
changeset
|
2005 if (!valid_spellang(lang)) |
|
5ef25fa57f71
patch 8.1.1143: may pass weird strings to file name expansion
Bram Moolenaar <Bram@vim.org>
parents:
16239
diff
changeset
|
2006 continue; |
|
5ef25fa57f71
patch 8.1.1143: may pass weird strings to file name expansion
Bram Moolenaar <Bram@vim.org>
parents:
16239
diff
changeset
|
2007 |
| 5477 | 2008 if (STRCMP(lang, "cjk") == 0) |
| 2009 { | |
| 2010 wp->w_s->b_cjk = 1; | |
| 2011 continue; | |
| 2012 } | |
| 2013 | |
| 355 | 2014 /* If the name ends in ".spl" use it as the name of the spell file. |
| 2015 * If there is a region name let "region" point to it and remove it | |
| 2016 * from the name. */ | |
| 2017 if (len > 4 && fnamecmp(lang + len - 4, ".spl") == 0) | |
| 236 | 2018 { |
| 355 | 2019 filename = TRUE; |
| 2020 | |
| 416 | 2021 /* Locate a region and remove it from the file name. */ |
| 2022 p = vim_strchr(gettail(lang), '_'); | |
| 2023 if (p != NULL && ASCII_ISALPHA(p[1]) && ASCII_ISALPHA(p[2]) | |
| 2024 && !ASCII_ISALPHA(p[3])) | |
| 2025 { | |
| 2026 vim_strncpy(region_cp, p + 1, 2); | |
| 2027 mch_memmove(p, p + 3, len - (p - lang) - 2); | |
| 2028 region = region_cp; | |
| 2029 } | |
| 2030 else | |
| 2031 dont_use_region = TRUE; | |
| 2032 | |
| 355 | 2033 /* Check if we loaded this language before. */ |
| 503 | 2034 for (slang = first_lang; slang != NULL; slang = slang->sl_next) |
|
16738
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2035 if (fullpathcmp(lang, slang->sl_fname, FALSE, TRUE) == FPC_SAME) |
| 355 | 2036 break; |
| 236 | 2037 } |
| 355 | 2038 else |
| 2039 { | |
| 2040 filename = FALSE; | |
| 2041 if (len > 3 && lang[len - 3] == '_') | |
| 2042 { | |
| 2043 region = lang + len - 2; | |
| 2044 len -= 3; | |
| 2045 lang[len] = NUL; | |
| 389 | 2046 } |
| 2047 else | |
| 2048 dont_use_region = TRUE; | |
| 355 | 2049 |
| 2050 /* Check if we loaded this language before. */ | |
| 503 | 2051 for (slang = first_lang; slang != NULL; slang = slang->sl_next) |
| 2052 if (STRICMP(lang, slang->sl_name) == 0) | |
| 355 | 2053 break; |
| 2054 } | |
| 236 | 2055 |
| 416 | 2056 if (region != NULL) |
| 2057 { | |
| 2058 /* If the region differs from what was used before then don't | |
| 2059 * use it for 'spellfile'. */ | |
| 2060 if (use_region != NULL && STRCMP(region, use_region) != 0) | |
| 2061 dont_use_region = TRUE; | |
| 2062 use_region = region; | |
| 2063 } | |
| 2064 | |
| 351 | 2065 /* If not found try loading the language now. */ |
| 503 | 2066 if (slang == NULL) |
| 355 | 2067 { |
| 2068 if (filename) | |
| 2069 (void)spell_load_file(lang, lang, NULL, FALSE); | |
| 2070 else | |
| 1185 | 2071 { |
| 355 | 2072 spell_load_lang(lang); |
| 1185 | 2073 /* SpellFileMissing autocommands may do anything, including |
| 2074 * destroying the buffer we are using... */ | |
|
9487
69ed2c9d34a6
commit https://github.com/vim/vim/commit/7c0a2f367f2507669560b1a66423155c70d2e75b
Christian Brabandt <cb@256bit.org>
parents:
9387
diff
changeset
|
2075 if (!bufref_valid(&bufref)) |
| 1185 | 2076 { |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
2077 ret_msg = N_("E797: SpellFileMissing autocommand deleted buffer"); |
| 1185 | 2078 goto theend; |
| 2079 } | |
| 2080 } | |
| 355 | 2081 } |
| 236 | 2082 |
| 307 | 2083 /* |
| 351 | 2084 * Loop over the languages, there can be several files for "lang". |
| 307 | 2085 */ |
| 503 | 2086 for (slang = first_lang; slang != NULL; slang = slang->sl_next) |
|
16738
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2087 if (filename ? fullpathcmp(lang, slang->sl_fname, FALSE, TRUE) |
|
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2088 == FPC_SAME |
| 503 | 2089 : STRICMP(lang, slang->sl_name) == 0) |
| 236 | 2090 { |
| 316 | 2091 region_mask = REGION_ALL; |
| 355 | 2092 if (!filename && region != NULL) |
| 236 | 2093 { |
| 307 | 2094 /* find region in sl_regions */ |
| 503 | 2095 c = find_region(slang->sl_regions, region); |
| 307 | 2096 if (c == REGION_ALL) |
| 2097 { | |
| 503 | 2098 if (slang->sl_add) |
| 389 | 2099 { |
| 503 | 2100 if (*slang->sl_regions != NUL) |
| 389 | 2101 /* This addition file is for other regions. */ |
| 2102 region_mask = 0; | |
| 2103 } | |
| 2104 else | |
| 2105 /* This is probably an error. Give a warning and | |
| 2106 * accept the words anyway. */ | |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
2107 smsg(_("Warning: region %s not supported"), |
| 351 | 2108 region); |
| 307 | 2109 } |
| 2110 else | |
| 2111 region_mask = 1 << c; | |
| 236 | 2112 } |
| 307 | 2113 |
| 389 | 2114 if (region_mask != 0) |
| 307 | 2115 { |
| 389 | 2116 if (ga_grow(&ga, 1) == FAIL) |
| 2117 { | |
| 2118 ga_clear(&ga); | |
| 1185 | 2119 ret_msg = e_outofmem; |
| 2120 goto theend; | |
| 389 | 2121 } |
| 503 | 2122 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = slang; |
| 389 | 2123 LANGP_ENTRY(ga, ga.ga_len)->lp_region = region_mask; |
| 2124 ++ga.ga_len; | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2125 use_midword(slang, wp); |
| 503 | 2126 if (slang->sl_nobreak) |
| 501 | 2127 nobreak = TRUE; |
| 307 | 2128 } |
| 385 | 2129 } |
| 2130 } | |
| 2131 | |
| 2132 /* round 0: load int_wordlist, if possible. | |
| 2133 * round 1: load first name in 'spellfile'. | |
| 2134 * round 2: load second name in 'spellfile. | |
| 2135 * etc. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2136 spf = curwin->w_s->b_p_spf; |
| 385 | 2137 for (round = 0; round == 0 || *spf != NUL; ++round) |
| 2138 { | |
| 2139 if (round == 0) | |
| 2140 { | |
| 2141 /* Internal wordlist, if there is one. */ | |
| 2142 if (int_wordlist == NULL) | |
| 381 | 2143 continue; |
| 385 | 2144 int_wordlist_spl(spf_name); |
| 381 | 2145 } |
| 2146 else | |
| 2147 { | |
| 385 | 2148 /* One entry in 'spellfile'. */ |
| 2149 copy_option_part(&spf, spf_name, MAXPATHL - 5, ","); | |
| 2150 STRCAT(spf_name, ".spl"); | |
| 2151 | |
| 2152 /* If it was already found above then skip it. */ | |
| 2153 for (c = 0; c < ga.ga_len; ++c) | |
| 500 | 2154 { |
| 2155 p = LANGP_ENTRY(ga, c)->lp_slang->sl_fname; | |
|
16738
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2156 if (p != NULL && fullpathcmp(spf_name, p, FALSE, TRUE) |
|
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2157 == FPC_SAME) |
| 385 | 2158 break; |
| 500 | 2159 } |
| 385 | 2160 if (c < ga.ga_len) |
| 381 | 2161 continue; |
| 2162 } | |
| 2163 | |
| 351 | 2164 /* Check if it was loaded already. */ |
| 503 | 2165 for (slang = first_lang; slang != NULL; slang = slang->sl_next) |
|
16738
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2166 if (fullpathcmp(spf_name, slang->sl_fname, FALSE, TRUE) |
|
b52ea9c5f1db
patch 8.1.1371: cannot recover from a swap file
Bram Moolenaar <Bram@vim.org>
parents:
16606
diff
changeset
|
2167 == FPC_SAME) |
| 323 | 2168 break; |
| 503 | 2169 if (slang == NULL) |
| 323 | 2170 { |
| 351 | 2171 /* Not loaded, try loading it now. The language name includes the |
| 385 | 2172 * region name, the region is ignored otherwise. for int_wordlist |
| 2173 * use an arbitrary name. */ | |
| 2174 if (round == 0) | |
| 2175 STRCPY(lang, "internal wordlist"); | |
| 2176 else | |
| 2177 { | |
| 2178 vim_strncpy(lang, gettail(spf_name), MAXWLEN); | |
| 381 | 2179 p = vim_strchr(lang, '.'); |
| 2180 if (p != NULL) | |
| 2181 *p = NUL; /* truncate at ".encoding.add" */ | |
| 2182 } | |
| 503 | 2183 slang = spell_load_file(spf_name, lang, NULL, TRUE); |
| 501 | 2184 |
| 2185 /* If one of the languages has NOBREAK we assume the addition | |
| 2186 * files also have this. */ | |
| 503 | 2187 if (slang != NULL && nobreak) |
| 2188 slang->sl_nobreak = TRUE; | |
| 2189 } | |
| 2190 if (slang != NULL && ga_grow(&ga, 1) == OK) | |
| 323 | 2191 { |
| 389 | 2192 region_mask = REGION_ALL; |
| 2193 if (use_region != NULL && !dont_use_region) | |
| 2194 { | |
| 2195 /* find region in sl_regions */ | |
| 503 | 2196 c = find_region(slang->sl_regions, use_region); |
| 389 | 2197 if (c != REGION_ALL) |
| 2198 region_mask = 1 << c; | |
| 503 | 2199 else if (*slang->sl_regions != NUL) |
| 389 | 2200 /* This spell file is for other regions. */ |
| 2201 region_mask = 0; | |
| 2202 } | |
| 2203 | |
| 2204 if (region_mask != 0) | |
| 2205 { | |
| 503 | 2206 LANGP_ENTRY(ga, ga.ga_len)->lp_slang = slang; |
| 2207 LANGP_ENTRY(ga, ga.ga_len)->lp_sallang = NULL; | |
| 2208 LANGP_ENTRY(ga, ga.ga_len)->lp_replang = NULL; | |
| 389 | 2209 LANGP_ENTRY(ga, ga.ga_len)->lp_region = region_mask; |
| 2210 ++ga.ga_len; | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2211 use_midword(slang, wp); |
| 389 | 2212 } |
| 323 | 2213 } |
| 2214 } | |
| 2215 | |
| 236 | 2216 /* Everything is fine, store the new b_langp value. */ |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2217 ga_clear(&wp->w_s->b_langp); |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2218 wp->w_s->b_langp = ga; |
| 236 | 2219 |
| 503 | 2220 /* For each language figure out what language to use for sound folding and |
| 2221 * REP items. If the language doesn't support it itself use another one | |
| 2222 * with the same name. E.g. for "en-math" use "en". */ | |
| 2223 for (i = 0; i < ga.ga_len; ++i) | |
| 2224 { | |
| 2225 lp = LANGP_ENTRY(ga, i); | |
| 2226 | |
| 2227 /* sound folding */ | |
| 2228 if (lp->lp_slang->sl_sal.ga_len > 0) | |
| 2229 /* language does sound folding itself */ | |
| 2230 lp->lp_sallang = lp->lp_slang; | |
| 2231 else | |
| 2232 /* find first similar language that does sound folding */ | |
| 2233 for (j = 0; j < ga.ga_len; ++j) | |
| 2234 { | |
| 2235 lp2 = LANGP_ENTRY(ga, j); | |
| 2236 if (lp2->lp_slang->sl_sal.ga_len > 0 | |
| 2237 && STRNCMP(lp->lp_slang->sl_name, | |
| 2238 lp2->lp_slang->sl_name, 2) == 0) | |
| 2239 { | |
| 2240 lp->lp_sallang = lp2->lp_slang; | |
| 2241 break; | |
| 2242 } | |
| 2243 } | |
| 2244 | |
| 2245 /* REP items */ | |
| 2246 if (lp->lp_slang->sl_rep.ga_len > 0) | |
| 2247 /* language has REP items itself */ | |
| 2248 lp->lp_replang = lp->lp_slang; | |
| 2249 else | |
| 625 | 2250 /* find first similar language that has REP items */ |
| 503 | 2251 for (j = 0; j < ga.ga_len; ++j) |
| 2252 { | |
| 2253 lp2 = LANGP_ENTRY(ga, j); | |
| 2254 if (lp2->lp_slang->sl_rep.ga_len > 0 | |
| 2255 && STRNCMP(lp->lp_slang->sl_name, | |
| 2256 lp2->lp_slang->sl_name, 2) == 0) | |
| 2257 { | |
| 2258 lp->lp_replang = lp2->lp_slang; | |
| 2259 break; | |
| 2260 } | |
| 2261 } | |
| 2262 } | |
| 2263 | |
| 1185 | 2264 theend: |
| 2265 vim_free(spl_copy); | |
| 2266 recursive = FALSE; | |
| 5891 | 2267 redraw_win_later(wp, NOT_VALID); |
| 1185 | 2268 return ret_msg; |
| 236 | 2269 } |
| 2270 | |
| 2271 /* | |
| 376 | 2272 * Clear the midword characters for buffer "buf". |
| 2273 */ | |
| 2274 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2275 clear_midword(win_T *wp) |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2276 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2277 vim_memset(wp->w_s->b_spell_ismw, 0, 256); |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
2278 VIM_CLEAR(wp->w_s->b_spell_ismw_mb); |
| 376 | 2279 } |
| 2280 | |
| 2281 /* | |
| 2282 * Use the "sl_midword" field of language "lp" for buffer "buf". | |
| 2283 * They add up to any currently used midword characters. | |
| 2284 */ | |
| 2285 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2286 use_midword(slang_T *lp, win_T *wp) |
| 376 | 2287 { |
| 2288 char_u *p; | |
| 2289 | |
| 389 | 2290 if (lp->sl_midword == NULL) /* there aren't any */ |
| 2291 return; | |
| 2292 | |
| 376 | 2293 for (p = lp->sl_midword; *p != NUL; ) |
| 2294 if (has_mbyte) | |
| 2295 { | |
| 2296 int c, l, n; | |
| 2297 char_u *bp; | |
| 2298 | |
| 2299 c = mb_ptr2char(p); | |
| 474 | 2300 l = (*mb_ptr2len)(p); |
| 2301 if (c < 256 && l <= 2) | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2302 wp->w_s->b_spell_ismw[c] = TRUE; |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2303 else if (wp->w_s->b_spell_ismw_mb == NULL) |
| 376 | 2304 /* First multi-byte char in "b_spell_ismw_mb". */ |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2305 wp->w_s->b_spell_ismw_mb = vim_strnsave(p, l); |
| 376 | 2306 else |
| 2307 { | |
| 2308 /* Append multi-byte chars to "b_spell_ismw_mb". */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2309 n = (int)STRLEN(wp->w_s->b_spell_ismw_mb); |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2310 bp = vim_strnsave(wp->w_s->b_spell_ismw_mb, n + l); |
| 376 | 2311 if (bp != NULL) |
| 2312 { | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2313 vim_free(wp->w_s->b_spell_ismw_mb); |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2314 wp->w_s->b_spell_ismw_mb = bp; |
| 376 | 2315 vim_strncpy(bp + n, p, l); |
| 2316 } | |
| 2317 } | |
| 2318 p += l; | |
| 2319 } | |
| 2320 else | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2321 wp->w_s->b_spell_ismw[*p++] = TRUE; |
| 376 | 2322 } |
| 2323 | |
| 2324 /* | |
| 236 | 2325 * Find the region "region[2]" in "rp" (points to "sl_regions"). |
|
15034
6e4e0d43b20b
patch 8.1.0528: various typos in comments
Bram Moolenaar <Bram@vim.org>
parents:
14862
diff
changeset
|
2326 * Each region is simply stored as the two characters of its name. |
| 381 | 2327 * Returns the index if found (first is 0), REGION_ALL if not found. |
| 236 | 2328 */ |
| 2329 static int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2330 find_region(char_u *rp, char_u *region) |
| 236 | 2331 { |
| 2332 int i; | |
| 2333 | |
| 2334 for (i = 0; ; i += 2) | |
| 2335 { | |
| 2336 if (rp[i] == NUL) | |
| 2337 return REGION_ALL; | |
| 2338 if (rp[i] == region[0] && rp[i + 1] == region[1]) | |
| 2339 break; | |
| 2340 } | |
| 2341 return i / 2; | |
| 2342 } | |
| 2343 | |
| 2344 /* | |
| 323 | 2345 * Return case type of word: |
| 236 | 2346 * w word 0 |
| 300 | 2347 * Word WF_ONECAP |
| 2348 * W WORD WF_ALLCAP | |
| 2349 * WoRd wOrd WF_KEEPCAP | |
| 236 | 2350 */ |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
2351 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2352 captype( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2353 char_u *word, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2354 char_u *end) /* When NULL use up to NUL byte. */ |
| 236 | 2355 { |
| 2356 char_u *p; | |
| 2357 int c; | |
| 2358 int firstcap; | |
| 2359 int allcap; | |
| 2360 int past_second = FALSE; /* past second word char */ | |
| 2361 | |
| 2362 /* find first letter */ | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
2363 for (p = word; !spell_iswordp_nmw(p, curwin); MB_PTR_ADV(p)) |
| 323 | 2364 if (end == NULL ? *p == NUL : p >= end) |
| 236 | 2365 return 0; /* only non-word characters, illegal word */ |
| 310 | 2366 if (has_mbyte) |
| 2367 c = mb_ptr2char_adv(&p); | |
| 2368 else | |
| 2369 c = *p++; | |
| 324 | 2370 firstcap = allcap = SPELL_ISUPPER(c); |
| 236 | 2371 |
| 2372 /* | |
| 2373 * Need to check all letters to find a word with mixed upper/lower. | |
| 2374 * But a word with an upper char only at start is a ONECAP. | |
| 2375 */ | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
2376 for ( ; end == NULL ? *p != NUL : p < end; MB_PTR_ADV(p)) |
| 5477 | 2377 if (spell_iswordp_nmw(p, curwin)) |
| 236 | 2378 { |
| 455 | 2379 c = PTR2CHAR(p); |
| 324 | 2380 if (!SPELL_ISUPPER(c)) |
| 236 | 2381 { |
| 2382 /* UUl -> KEEPCAP */ | |
| 2383 if (past_second && allcap) | |
| 300 | 2384 return WF_KEEPCAP; |
| 236 | 2385 allcap = FALSE; |
| 2386 } | |
| 2387 else if (!allcap) | |
| 2388 /* UlU -> KEEPCAP */ | |
| 300 | 2389 return WF_KEEPCAP; |
| 236 | 2390 past_second = TRUE; |
| 2391 } | |
| 2392 | |
| 2393 if (allcap) | |
| 300 | 2394 return WF_ALLCAP; |
| 236 | 2395 if (firstcap) |
| 300 | 2396 return WF_ONECAP; |
| 236 | 2397 return 0; |
| 2398 } | |
| 2399 | |
| 474 | 2400 /* |
| 5519 | 2401 * Delete the internal wordlist and its .spl file. |
| 2402 */ | |
| 2403 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2404 spell_delete_wordlist(void) |
| 5519 | 2405 { |
| 2406 char_u fname[MAXPATHL]; | |
| 2407 | |
| 2408 if (int_wordlist != NULL) | |
| 2409 { | |
| 2410 mch_remove(int_wordlist); | |
| 2411 int_wordlist_spl(fname); | |
| 2412 mch_remove(fname); | |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
2413 VIM_CLEAR(int_wordlist); |
| 5519 | 2414 } |
| 2415 } | |
| 2416 | |
| 355 | 2417 /* |
| 2418 * Free all languages. | |
| 2419 */ | |
| 2420 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2421 spell_free_all(void) |
| 355 | 2422 { |
| 503 | 2423 slang_T *slang; |
| 355 | 2424 buf_T *buf; |
| 2425 | |
| 2580 | 2426 /* Go through all buffers and handle 'spelllang'. <VN> */ |
|
9649
fd9727ae3c49
commit https://github.com/vim/vim/commit/2932359000b2f918d5fade79ea4d124d5943cd07
Christian Brabandt <cb@256bit.org>
parents:
9583
diff
changeset
|
2427 FOR_ALL_BUFFERS(buf) |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2428 ga_clear(&buf->b_s.b_langp); |
| 355 | 2429 |
| 2430 while (first_lang != NULL) | |
| 2431 { | |
| 503 | 2432 slang = first_lang; |
| 2433 first_lang = slang->sl_next; | |
| 2434 slang_free(slang); | |
| 355 | 2435 } |
| 366 | 2436 |
| 5519 | 2437 spell_delete_wordlist(); |
| 381 | 2438 |
|
13244
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
2439 VIM_CLEAR(repl_to); |
|
ac42c4b11dbc
patch 8.0.1496: clearing a pointer takes two lines
Christian Brabandt <cb@256bit.org>
parents:
12674
diff
changeset
|
2440 VIM_CLEAR(repl_from); |
| 355 | 2441 } |
|
15605
62b3805506b3
patch 8.1.0810: too many #ifdefs
Bram Moolenaar <Bram@vim.org>
parents:
15543
diff
changeset
|
2442 |
| 236 | 2443 /* |
| 2444 * Clear all spelling tables and reload them. | |
| 307 | 2445 * Used after 'encoding' is set and when ":mkspell" was used. |
| 236 | 2446 */ |
| 2447 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2448 spell_reload(void) |
| 236 | 2449 { |
| 316 | 2450 win_T *wp; |
| 236 | 2451 |
| 358 | 2452 /* Initialize the table for spell_iswordp(). */ |
| 236 | 2453 init_spell_chartab(); |
| 2454 | |
| 2455 /* Unload all allocated memory. */ | |
| 355 | 2456 spell_free_all(); |
| 236 | 2457 |
| 2458 /* Go through all buffers and handle 'spelllang'. */ | |
|
9649
fd9727ae3c49
commit https://github.com/vim/vim/commit/2932359000b2f918d5fade79ea4d124d5943cd07
Christian Brabandt <cb@256bit.org>
parents:
9583
diff
changeset
|
2459 FOR_ALL_WINDOWS(wp) |
| 236 | 2460 { |
| 316 | 2461 /* Only load the wordlists when 'spelllang' is set and there is a |
| 2462 * window for this buffer in which 'spell' is set. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2463 if (*wp->w_s->b_p_spl != NUL) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2464 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2465 if (wp->w_p_spell) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2466 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2467 (void)did_set_spelllang(wp); |
| 316 | 2468 break; |
| 2469 } | |
| 2470 } | |
| 236 | 2471 } |
| 2472 } | |
| 2473 | |
| 310 | 2474 /* |
| 625 | 2475 * Open a spell buffer. This is a nameless buffer that is not in the buffer |
| 2476 * list and only contains text lines. Can use a swapfile to reduce memory | |
| 2477 * use. | |
| 2478 * Most other fields are invalid! Esp. watch out for string options being | |
| 2479 * NULL and there is no undo info. | |
| 2480 * Returns NULL when out of memory. | |
| 2481 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
2482 buf_T * |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2483 open_spellbuf(void) |
| 625 | 2484 { |
| 2485 buf_T *buf; | |
| 2486 | |
|
16825
ce04ebdf26b8
patch 8.1.1414: alloc() returning "char_u *" causes a lot of type casts
Bram Moolenaar <Bram@vim.org>
parents:
16768
diff
changeset
|
2487 buf = ALLOC_CLEAR_ONE(buf_T); |
| 625 | 2488 if (buf != NULL) |
| 2489 { | |
| 2490 buf->b_spell = TRUE; | |
| 2491 buf->b_p_swf = TRUE; /* may create a swap file */ | |
|
5204
7aca84c0cd37
updated for version 7.4a.028
Bram Moolenaar <bram@vim.org>
parents:
4805
diff
changeset
|
2492 #ifdef FEAT_CRYPT |
|
7aca84c0cd37
updated for version 7.4a.028
Bram Moolenaar <bram@vim.org>
parents:
4805
diff
changeset
|
2493 buf->b_p_key = empty_option; |
|
7aca84c0cd37
updated for version 7.4a.028
Bram Moolenaar <bram@vim.org>
parents:
4805
diff
changeset
|
2494 #endif |
| 625 | 2495 ml_open(buf); |
| 2496 ml_open_file(buf); /* create swap file now */ | |
| 2497 } | |
| 2498 return buf; | |
| 2499 } | |
| 2500 | |
| 2501 /* | |
| 2502 * Close the buffer used for spell info. | |
| 2503 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
2504 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2505 close_spellbuf(buf_T *buf) |
| 625 | 2506 { |
| 2507 if (buf != NULL) | |
| 2508 { | |
| 2509 ml_close(buf, TRUE); | |
| 2510 vim_free(buf); | |
| 2511 } | |
| 2512 } | |
| 2513 | |
| 307 | 2514 /* |
| 2515 * Init the chartab used for spelling for ASCII. | |
| 2516 * EBCDIC is not supported! | |
| 2517 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
2518 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2519 clear_spell_chartab(spelltab_T *sp) |
| 307 | 2520 { |
| 324 | 2521 int i; |
| 307 | 2522 |
| 2523 /* Init everything to FALSE. */ | |
| 2524 vim_memset(sp->st_isw, FALSE, sizeof(sp->st_isw)); | |
| 2525 vim_memset(sp->st_isu, FALSE, sizeof(sp->st_isu)); | |
| 2526 for (i = 0; i < 256; ++i) | |
| 324 | 2527 { |
| 307 | 2528 sp->st_fold[i] = i; |
| 324 | 2529 sp->st_upper[i] = i; |
| 2530 } | |
| 307 | 2531 |
| 2532 /* We include digits. A word shouldn't start with a digit, but handling | |
| 2533 * that is done separately. */ | |
| 2534 for (i = '0'; i <= '9'; ++i) | |
| 2535 sp->st_isw[i] = TRUE; | |
| 2536 for (i = 'A'; i <= 'Z'; ++i) | |
| 2537 { | |
| 2538 sp->st_isw[i] = TRUE; | |
| 2539 sp->st_isu[i] = TRUE; | |
| 2540 sp->st_fold[i] = i + 0x20; | |
| 2541 } | |
| 2542 for (i = 'a'; i <= 'z'; ++i) | |
| 324 | 2543 { |
| 307 | 2544 sp->st_isw[i] = TRUE; |
| 324 | 2545 sp->st_upper[i] = i - 0x20; |
| 2546 } | |
| 307 | 2547 } |
| 2548 | |
| 2549 /* | |
| 2550 * Init the chartab used for spelling. Only depends on 'encoding'. | |
| 2551 * Called once while starting up and when 'encoding' changes. | |
| 2552 * The default is to use isalpha(), but the spell file should define the word | |
| 2553 * characters to make it possible that 'encoding' differs from the current | |
| 390 | 2554 * locale. For utf-8 we don't use isalpha() but our own functions. |
| 307 | 2555 */ |
| 2556 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2557 init_spell_chartab(void) |
| 307 | 2558 { |
| 2559 int i; | |
| 2560 | |
| 2561 did_set_spelltab = FALSE; | |
| 2562 clear_spell_chartab(&spelltab); | |
| 2563 if (enc_dbcs) | |
| 2564 { | |
| 2565 /* DBCS: assume double-wide characters are word characters. */ | |
| 2566 for (i = 128; i <= 255; ++i) | |
| 2567 if (MB_BYTE2LEN(i) == 2) | |
| 2568 spelltab.st_isw[i] = TRUE; | |
| 2569 } | |
| 324 | 2570 else if (enc_utf8) |
| 2571 { | |
| 2572 for (i = 128; i < 256; ++i) | |
| 2573 { | |
|
2140
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2574 int f = utf_fold(i); |
|
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2575 int u = utf_toupper(i); |
|
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2576 |
| 324 | 2577 spelltab.st_isu[i] = utf_isupper(i); |
| 2578 spelltab.st_isw[i] = spelltab.st_isu[i] || utf_islower(i); | |
|
2140
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2579 /* The folded/upper-cased value is different between latin1 and |
|
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2580 * utf8 for 0xb5, causing E763 for no good reason. Use the latin1 |
|
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2581 * value for utf-8 to avoid this. */ |
|
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2582 spelltab.st_fold[i] = (f < 256) ? f : i; |
|
12aba62fa7c6
updated for version 7.2.422
Bram Moolenaar <bram@zimbu.org>
parents:
2046
diff
changeset
|
2583 spelltab.st_upper[i] = (u < 256) ? u : i; |
| 324 | 2584 } |
| 2585 } | |
| 307 | 2586 else |
| 2587 { | |
| 324 | 2588 /* Rough guess: use locale-dependent library functions. */ |
| 307 | 2589 for (i = 128; i < 256; ++i) |
| 2590 { | |
| 2591 if (MB_ISUPPER(i)) | |
| 2592 { | |
| 324 | 2593 spelltab.st_isw[i] = TRUE; |
| 307 | 2594 spelltab.st_isu[i] = TRUE; |
| 2595 spelltab.st_fold[i] = MB_TOLOWER(i); | |
| 2596 } | |
| 324 | 2597 else if (MB_ISLOWER(i)) |
| 2598 { | |
| 2599 spelltab.st_isw[i] = TRUE; | |
| 2600 spelltab.st_upper[i] = MB_TOUPPER(i); | |
| 2601 } | |
| 307 | 2602 } |
| 2603 } | |
| 2604 } | |
| 2605 | |
| 2606 | |
| 2607 /* | |
| 358 | 2608 * Return TRUE if "p" points to a word character. |
| 366 | 2609 * As a special case we see "midword" characters as word character when it is |
| 358 | 2610 * followed by a word character. This finds they'there but not 'they there'. |
| 366 | 2611 * Thus this only works properly when past the first character of the word. |
| 358 | 2612 */ |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
2613 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2614 spell_iswordp( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2615 char_u *p, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2616 win_T *wp) /* buffer used */ |
| 358 | 2617 { |
| 2618 char_u *s; | |
| 366 | 2619 int l; |
| 2620 int c; | |
| 2621 | |
| 2622 if (has_mbyte) | |
| 2623 { | |
|
18251
c8a53c0daeed
patch 8.1.2120: some MB_ macros are more complicated than necessary
Bram Moolenaar <Bram@vim.org>
parents:
18172
diff
changeset
|
2624 l = mb_ptr2len(p); |
| 358 | 2625 s = p; |
| 366 | 2626 if (l == 1) |
| 2627 { | |
| 2628 /* be quick for ASCII */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2629 if (wp->w_s->b_spell_ismw[*p]) |
| 366 | 2630 s = p + 1; /* skip a mid-word character */ |
| 2631 } | |
| 2632 else | |
| 2633 { | |
| 2634 c = mb_ptr2char(p); | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2635 if (c < 256 ? wp->w_s->b_spell_ismw[c] |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2636 : (wp->w_s->b_spell_ismw_mb != NULL |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2637 && vim_strchr(wp->w_s->b_spell_ismw_mb, c) != NULL)) |
| 366 | 2638 s = p + l; |
| 2639 } | |
| 2640 | |
| 390 | 2641 c = mb_ptr2char(s); |
| 2642 if (c > 255) | |
| 5477 | 2643 return spell_mb_isword_class(mb_get_class(s), wp); |
| 390 | 2644 return spelltab.st_isw[c]; |
| 366 | 2645 } |
| 2646 | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2647 return spelltab.st_isw[wp->w_s->b_spell_ismw[*p] ? p[1] : p[0]]; |
| 376 | 2648 } |
| 2649 | |
| 2650 /* | |
| 2651 * Return TRUE if "p" points to a word character. | |
| 2652 * Unlike spell_iswordp() this doesn't check for "midword" characters. | |
| 2653 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
2654 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2655 spell_iswordp_nmw(char_u *p, win_T *wp) |
| 376 | 2656 { |
| 390 | 2657 int c; |
| 2658 | |
| 2659 if (has_mbyte) | |
| 2660 { | |
| 2661 c = mb_ptr2char(p); | |
| 2662 if (c > 255) | |
| 5477 | 2663 return spell_mb_isword_class(mb_get_class(p), wp); |
| 390 | 2664 return spelltab.st_isw[c]; |
| 2665 } | |
| 376 | 2666 return spelltab.st_isw[*p]; |
| 358 | 2667 } |
| 2668 | |
| 372 | 2669 /* |
| 1580 | 2670 * Return TRUE if word class indicates a word character. |
| 2671 * Only for characters above 255. | |
| 2672 * Unicode subscript and superscript are not considered word characters. | |
| 5477 | 2673 * See also dbcs_class() and utf_class() in mbyte.c. |
| 2674 */ | |
| 2675 static int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2676 spell_mb_isword_class(int cl, win_T *wp) |
| 5477 | 2677 { |
| 2678 if (wp->w_s->b_cjk) | |
| 2679 /* East Asian characters are not considered word characters. */ | |
| 2680 return cl == 2 || cl == 0x2800; | |
|
17434
26e8d42987ca
patch 8.1.1715: emoji characters are seen as word characters for spelling
Bram Moolenaar <Bram@vim.org>
parents:
16825
diff
changeset
|
2681 return cl >= 2 && cl != 0x2070 && cl != 0x2080 && cl != 3; |
| 1580 | 2682 } |
| 2683 | |
| 2684 /* | |
| 372 | 2685 * Return TRUE if "p" points to a word character. |
| 2686 * Wide version of spell_iswordp(). | |
| 2687 */ | |
| 2688 static int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2689 spell_iswordp_w(int *p, win_T *wp) |
| 372 | 2690 { |
| 2691 int *s; | |
| 2692 | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2693 if (*p < 256 ? wp->w_s->b_spell_ismw[*p] |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2694 : (wp->w_s->b_spell_ismw_mb != NULL |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2695 && vim_strchr(wp->w_s->b_spell_ismw_mb, *p) != NULL)) |
| 372 | 2696 s = p + 1; |
| 2697 else | |
| 2698 s = p; | |
| 2699 | |
| 390 | 2700 if (*s > 255) |
| 372 | 2701 { |
| 2702 if (enc_utf8) | |
| 5477 | 2703 return spell_mb_isword_class(utf_class(*s), wp); |
| 372 | 2704 if (enc_dbcs) |
| 5477 | 2705 return spell_mb_isword_class( |
| 2706 dbcs_class((unsigned)*s >> 8, *s & 0xff), wp); | |
| 372 | 2707 return 0; |
| 2708 } | |
| 2709 return spelltab.st_isw[*s]; | |
| 2710 } | |
| 2711 | |
| 358 | 2712 /* |
| 324 | 2713 * Case-fold "str[len]" into "buf[buflen]". The result is NUL terminated. |
| 2714 * Uses the character definitions from the .spl file. | |
| 307 | 2715 * When using a multi-byte 'encoding' the length may change! |
| 2716 * Returns FAIL when something wrong. | |
| 2717 */ | |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
2718 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2719 spell_casefold( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2720 char_u *str, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2721 int len, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2722 char_u *buf, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2723 int buflen) |
| 307 | 2724 { |
| 2725 int i; | |
| 2726 | |
| 2727 if (len >= buflen) | |
| 2728 { | |
| 2729 buf[0] = NUL; | |
| 2730 return FAIL; /* result will not fit */ | |
| 2731 } | |
| 2732 | |
| 2733 if (has_mbyte) | |
| 2734 { | |
| 324 | 2735 int outi = 0; |
| 2736 char_u *p; | |
| 307 | 2737 int c; |
| 2738 | |
| 2739 /* Fold one character at a time. */ | |
| 324 | 2740 for (p = str; p < str + len; ) |
| 307 | 2741 { |
| 2742 if (outi + MB_MAXBYTES > buflen) | |
| 2743 { | |
| 2744 buf[outi] = NUL; | |
| 2745 return FAIL; | |
| 2746 } | |
| 474 | 2747 c = mb_cptr2char_adv(&p); |
| 324 | 2748 outi += mb_char2bytes(SPELL_TOFOLD(c), buf + outi); |
| 307 | 2749 } |
| 2750 buf[outi] = NUL; | |
| 2751 } | |
| 2752 else | |
| 2753 { | |
| 2754 /* Be quick for non-multibyte encodings. */ | |
| 2755 for (i = 0; i < len; ++i) | |
| 324 | 2756 buf[i] = spelltab.st_fold[str[i]]; |
| 307 | 2757 buf[i] = NUL; |
| 2758 } | |
| 2759 | |
| 2760 return OK; | |
| 2761 } | |
| 2762 | |
| 344 | 2763 /* |
| 475 | 2764 * Check if the word at line "lnum" column "col" is required to start with a |
| 2765 * capital. This uses 'spellcapcheck' of the current buffer. | |
| 2766 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
2767 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2768 check_need_cap(linenr_T lnum, colnr_T col) |
| 475 | 2769 { |
| 2770 int need_cap = FALSE; | |
| 2771 char_u *line; | |
| 2772 char_u *line_copy = NULL; | |
| 2773 char_u *p; | |
| 2774 colnr_T endcol; | |
| 2775 regmatch_T regmatch; | |
| 2776 | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2777 if (curwin->w_s->b_cap_prog == NULL) |
| 475 | 2778 return FALSE; |
| 2779 | |
| 2780 line = ml_get_curline(); | |
| 2781 endcol = 0; | |
|
12323
4dba3e4f3b01
patch 8.0.1041: bogus characters when indenting during visual-block append
Christian Brabandt <cb@256bit.org>
parents:
11521
diff
changeset
|
2782 if (getwhitecols(line) >= (int)col) |
| 475 | 2783 { |
| 2784 /* At start of line, check if previous line is empty or sentence | |
| 2785 * ends there. */ | |
| 2786 if (lnum == 1) | |
| 2787 need_cap = TRUE; | |
| 2788 else | |
| 2789 { | |
| 2790 line = ml_get(lnum - 1); | |
| 2791 if (*skipwhite(line) == NUL) | |
| 2792 need_cap = TRUE; | |
| 2793 else | |
| 2794 { | |
| 2795 /* Append a space in place of the line break. */ | |
| 2796 line_copy = concat_str(line, (char_u *)" "); | |
| 2797 line = line_copy; | |
| 835 | 2798 endcol = (colnr_T)STRLEN(line); |
| 475 | 2799 } |
| 2800 } | |
| 2801 } | |
| 2802 else | |
| 2803 endcol = col; | |
| 2804 | |
| 2805 if (endcol > 0) | |
| 2806 { | |
| 2807 /* Check if sentence ends before the bad word. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
2808 regmatch.regprog = curwin->w_s->b_cap_prog; |
| 475 | 2809 regmatch.rm_ic = FALSE; |
| 2810 p = line + endcol; | |
| 2811 for (;;) | |
| 2812 { | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
2813 MB_PTR_BACK(line, p); |
| 5477 | 2814 if (p == line || spell_iswordp_nmw(p, curwin)) |
| 475 | 2815 break; |
| 2816 if (vim_regexec(®match, p, 0) | |
| 2817 && regmatch.endp[0] == line + endcol) | |
| 2818 { | |
| 2819 need_cap = TRUE; | |
| 2820 break; | |
| 2821 } | |
| 2822 } | |
| 6375 | 2823 curwin->w_s->b_cap_prog = regmatch.regprog; |
| 475 | 2824 } |
| 2825 | |
| 2826 vim_free(line_copy); | |
| 2827 | |
| 2828 return need_cap; | |
| 2829 } | |
| 2830 | |
| 2831 | |
| 2832 /* | |
| 372 | 2833 * ":spellrepall" |
| 2834 */ | |
| 2835 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2836 ex_spellrepall(exarg_T *eap UNUSED) |
| 372 | 2837 { |
| 2838 pos_T pos = curwin->w_cursor; | |
| 2839 char_u *frompat; | |
| 2840 int addlen; | |
| 2841 char_u *line; | |
| 2842 char_u *p; | |
| 2843 int save_ws = p_ws; | |
| 483 | 2844 linenr_T prev_lnum = 0; |
| 372 | 2845 |
| 2846 if (repl_from == NULL || repl_to == NULL) | |
| 2847 { | |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
2848 emsg(_("E752: No previous spell replacement")); |
| 372 | 2849 return; |
| 2850 } | |
| 835 | 2851 addlen = (int)(STRLEN(repl_to) - STRLEN(repl_from)); |
| 2852 | |
|
16764
ef00b6bc186b
patch 8.1.1384: using "int" for alloc() often results in compiler warnings
Bram Moolenaar <Bram@vim.org>
parents:
16738
diff
changeset
|
2853 frompat = alloc(STRLEN(repl_from) + 7); |
| 372 | 2854 if (frompat == NULL) |
| 2855 return; | |
| 2856 sprintf((char *)frompat, "\\V\\<%s\\>", repl_from); | |
| 2857 p_ws = FALSE; | |
| 2858 | |
| 483 | 2859 sub_nsubs = 0; |
| 2860 sub_nlines = 0; | |
| 372 | 2861 curwin->w_cursor.lnum = 0; |
| 2862 while (!got_int) | |
| 2863 { | |
|
18358
34d5cd432cac
patch 8.1.2173: searchit() has too many arguments
Bram Moolenaar <Bram@vim.org>
parents:
18305
diff
changeset
|
2864 if (do_search(NULL, '/', frompat, 1L, SEARCH_KEEP, NULL) == 0 |
| 372 | 2865 || u_save_cursor() == FAIL) |
| 2866 break; | |
| 2867 | |
| 2868 /* Only replace when the right word isn't there yet. This happens | |
| 2869 * when changing "etc" to "etc.". */ | |
| 2870 line = ml_get_curline(); | |
| 2871 if (addlen <= 0 || STRNCMP(line + curwin->w_cursor.col, | |
| 2872 repl_to, STRLEN(repl_to)) != 0) | |
| 2873 { | |
|
16764
ef00b6bc186b
patch 8.1.1384: using "int" for alloc() often results in compiler warnings
Bram Moolenaar <Bram@vim.org>
parents:
16738
diff
changeset
|
2874 p = alloc(STRLEN(line) + addlen + 1); |
| 372 | 2875 if (p == NULL) |
| 2876 break; | |
| 2877 mch_memmove(p, line, curwin->w_cursor.col); | |
| 2878 STRCPY(p + curwin->w_cursor.col, repl_to); | |
| 2879 STRCAT(p, line + curwin->w_cursor.col + STRLEN(repl_from)); | |
| 2880 ml_replace(curwin->w_cursor.lnum, p, FALSE); | |
| 2881 changed_bytes(curwin->w_cursor.lnum, curwin->w_cursor.col); | |
| 483 | 2882 |
| 2883 if (curwin->w_cursor.lnum != prev_lnum) | |
| 2884 { | |
| 2885 ++sub_nlines; | |
| 2886 prev_lnum = curwin->w_cursor.lnum; | |
| 2887 } | |
| 2888 ++sub_nsubs; | |
| 372 | 2889 } |
| 835 | 2890 curwin->w_cursor.col += (colnr_T)STRLEN(repl_to); |
| 372 | 2891 } |
| 2892 | |
| 2893 p_ws = save_ws; | |
| 2894 curwin->w_cursor = pos; | |
| 2895 vim_free(frompat); | |
| 2896 | |
| 483 | 2897 if (sub_nsubs == 0) |
|
15470
55ccc2d353bd
patch 8.1.0743: giving error messages is not flexible
Bram Moolenaar <Bram@vim.org>
parents:
15034
diff
changeset
|
2898 semsg(_("E753: Not found: %s"), repl_from); |
| 483 | 2899 else |
| 2900 do_sub_msg(FALSE); | |
| 372 | 2901 } |
| 2902 | |
| 2903 /* | |
| 324 | 2904 * Make a copy of "word", with the first letter upper or lower cased, to |
| 2905 * "wcopy[MAXWLEN]". "word" must not be empty. | |
| 2906 * The result is NUL terminated. | |
| 323 | 2907 */ |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
2908 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2909 onecap_copy( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2910 char_u *word, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2911 char_u *wcopy, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2912 int upper) /* TRUE: first letter made upper case */ |
| 323 | 2913 { |
| 2914 char_u *p; | |
| 2915 int c; | |
| 2916 int l; | |
| 2917 | |
| 2918 p = word; | |
| 2919 if (has_mbyte) | |
| 474 | 2920 c = mb_cptr2char_adv(&p); |
| 323 | 2921 else |
| 2922 c = *p++; | |
| 2923 if (upper) | |
| 324 | 2924 c = SPELL_TOUPPER(c); |
| 323 | 2925 else |
| 324 | 2926 c = SPELL_TOFOLD(c); |
| 323 | 2927 if (has_mbyte) |
| 2928 l = mb_char2bytes(c, wcopy); | |
| 2929 else | |
| 2930 { | |
| 2931 l = 1; | |
| 2932 wcopy[0] = c; | |
| 2933 } | |
| 376 | 2934 vim_strncpy(wcopy + l, p, MAXWLEN - l - 1); |
| 323 | 2935 } |
| 2936 | |
| 2937 /* | |
| 324 | 2938 * Make a copy of "word" with all the letters upper cased into |
| 2939 * "wcopy[MAXWLEN]". The result is NUL terminated. | |
| 323 | 2940 */ |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
2941 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2942 allcap_copy(char_u *word, char_u *wcopy) |
| 323 | 2943 { |
| 2944 char_u *s; | |
| 2945 char_u *d; | |
| 2946 int c; | |
| 2947 | |
| 2948 d = wcopy; | |
| 2949 for (s = word; *s != NUL; ) | |
| 2950 { | |
| 2951 if (has_mbyte) | |
| 474 | 2952 c = mb_cptr2char_adv(&s); |
| 323 | 2953 else |
| 2954 c = *s++; | |
| 492 | 2955 |
| 3042 | 2956 /* We only change 0xdf to SS when we are certain latin1 is used. It |
| 492 | 2957 * would cause weird errors in other 8-bit encodings. */ |
| 2958 if (enc_latin1like && c == 0xdf) | |
| 2959 { | |
| 2960 c = 'S'; | |
| 2961 if (d - wcopy >= MAXWLEN - 1) | |
| 2962 break; | |
| 2963 *d++ = c; | |
| 2964 } | |
| 2965 else | |
| 2966 c = SPELL_TOUPPER(c); | |
| 323 | 2967 |
| 2968 if (has_mbyte) | |
| 2969 { | |
| 2970 if (d - wcopy >= MAXWLEN - MB_MAXBYTES) | |
| 2971 break; | |
| 2972 d += mb_char2bytes(c, d); | |
| 2973 } | |
| 2974 else | |
| 2975 { | |
| 2976 if (d - wcopy >= MAXWLEN - 1) | |
| 2977 break; | |
| 2978 *d++ = c; | |
| 2979 } | |
| 2980 } | |
| 2981 *d = NUL; | |
| 2982 } | |
| 2983 | |
| 2984 /* | |
| 455 | 2985 * Case-folding may change the number of bytes: Count nr of chars in |
| 2986 * fword[flen] and return the byte length of that many chars in "word". | |
| 2987 */ | |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
2988 int |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
2989 nofold_len(char_u *fword, int flen, char_u *word) |
| 455 | 2990 { |
| 2991 char_u *p; | |
| 2992 int i = 0; | |
| 2993 | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
2994 for (p = fword; p < fword + flen; MB_PTR_ADV(p)) |
| 455 | 2995 ++i; |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
2996 for (p = word; i > 0; MB_PTR_ADV(p)) |
| 455 | 2997 --i; |
| 2998 return (int)(p - word); | |
| 2999 } | |
| 3000 | |
| 323 | 3001 /* |
| 324 | 3002 * Copy "fword" to "cword", fixing case according to "flags". |
| 323 | 3003 */ |
|
18172
6e53d83e021d
patch 8.1.2081: the spell.c file is too big
Bram Moolenaar <Bram@vim.org>
parents:
17940
diff
changeset
|
3004 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3005 make_case_word(char_u *fword, char_u *cword, int flags) |
| 323 | 3006 { |
| 3007 if (flags & WF_ALLCAP) | |
| 3008 /* Make it all upper-case */ | |
| 3009 allcap_copy(fword, cword); | |
| 3010 else if (flags & WF_ONECAP) | |
| 3011 /* Make the first letter upper-case */ | |
| 324 | 3012 onecap_copy(fword, cword, TRUE); |
| 323 | 3013 else |
| 3014 /* Use goodword as-is. */ | |
| 3015 STRCPY(cword, fword); | |
| 3016 } | |
| 3017 | |
| 372 | 3018 #if defined(FEAT_EVAL) || defined(PROTO) |
| 3019 /* | |
| 3020 * Soundfold a string, for soundfold(). | |
| 3021 * Result is in allocated memory, NULL for an error. | |
| 3022 */ | |
| 3023 char_u * | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3024 eval_soundfold(char_u *word) |
| 372 | 3025 { |
| 3026 langp_T *lp; | |
| 3027 char_u sound[MAXWLEN]; | |
| 500 | 3028 int lpi; |
| 372 | 3029 |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3030 if (curwin->w_p_spell && *curwin->w_s->b_p_spl != NUL) |
| 372 | 3031 /* Use the sound-folding of the first language that supports it. */ |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3032 for (lpi = 0; lpi < curwin->w_s->b_langp.ga_len; ++lpi) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3033 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3034 lp = LANGP_ENTRY(curwin->w_s->b_langp, lpi); |
| 372 | 3035 if (lp->lp_slang->sl_sal.ga_len > 0) |
| 3036 { | |
| 3037 /* soundfold the word */ | |
| 375 | 3038 spell_soundfold(lp->lp_slang, word, FALSE, sound); |
| 372 | 3039 return vim_strsave(sound); |
| 3040 } | |
| 500 | 3041 } |
| 372 | 3042 |
| 3043 /* No language with sound folding, return word as-is. */ | |
| 3044 return vim_strsave(word); | |
| 3045 } | |
| 3046 #endif | |
| 3047 | |
| 323 | 3048 /* |
| 3049 * Turn "inword" into its sound-a-like equivalent in "res[MAXWLEN]". | |
| 485 | 3050 * |
| 3051 * There are many ways to turn a word into a sound-a-like representation. The | |
| 3052 * oldest is Soundex (1918!). A nice overview can be found in "Approximate | |
| 3053 * swedish name matching - survey and test of different algorithms" by Klas | |
| 3054 * Erikson. | |
| 3055 * | |
| 3056 * We support two methods: | |
| 3057 * 1. SOFOFROM/SOFOTO do a simple character mapping. | |
| 3058 * 2. SAL items define a more advanced sound-folding (and much slower). | |
| 323 | 3059 */ |
|
9583
b0c7061d6439
commit https://github.com/vim/vim/commit/9ccfebddc3ff2a3c2853cf706fd4c26f639bf381
Christian Brabandt <cb@256bit.org>
parents:
9570
diff
changeset
|
3060 void |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3061 spell_soundfold( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3062 slang_T *slang, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3063 char_u *inword, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3064 int folded, /* "inword" is already case-folded */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3065 char_u *res) |
| 375 | 3066 { |
| 3067 char_u fword[MAXWLEN]; | |
| 3068 char_u *word; | |
| 3069 | |
| 3070 if (slang->sl_sofo) | |
| 3071 /* SOFOFROM and SOFOTO used */ | |
| 3072 spell_soundfold_sofo(slang, inword, res); | |
| 3073 else | |
| 3074 { | |
| 3075 /* SAL items used. Requires the word to be case-folded. */ | |
| 3076 if (folded) | |
| 3077 word = inword; | |
| 3078 else | |
| 3079 { | |
| 835 | 3080 (void)spell_casefold(inword, (int)STRLEN(inword), fword, MAXWLEN); |
| 375 | 3081 word = fword; |
| 3082 } | |
| 3083 | |
| 3084 if (has_mbyte) | |
| 3085 spell_soundfold_wsal(slang, word, res); | |
| 3086 else | |
| 3087 spell_soundfold_sal(slang, word, res); | |
| 3088 } | |
| 3089 } | |
| 3090 | |
| 3091 /* | |
| 3092 * Perform sound folding of "inword" into "res" according to SOFOFROM and | |
| 3093 * SOFOTO lines. | |
| 3094 */ | |
| 3095 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3096 spell_soundfold_sofo(slang_T *slang, char_u *inword, char_u *res) |
| 375 | 3097 { |
| 3098 char_u *s; | |
| 3099 int ri = 0; | |
| 3100 int c; | |
| 3101 | |
| 3102 if (has_mbyte) | |
| 3103 { | |
| 3104 int prevc = 0; | |
| 3105 int *ip; | |
| 3106 | |
| 3107 /* The sl_sal_first[] table contains the translation for chars up to | |
| 3108 * 255, sl_sal the rest. */ | |
| 3109 for (s = inword; *s != NUL; ) | |
| 3110 { | |
| 474 | 3111 c = mb_cptr2char_adv(&s); |
|
11129
f4ea50924c6d
patch 8.0.0452: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11127
diff
changeset
|
3112 if (enc_utf8 ? utf_class(c) == 0 : VIM_ISWHITE(c)) |
| 375 | 3113 c = ' '; |
| 3114 else if (c < 256) | |
| 3115 c = slang->sl_sal_first[c]; | |
| 3116 else | |
| 3117 { | |
| 3118 ip = ((int **)slang->sl_sal.ga_data)[c & 0xff]; | |
| 3119 if (ip == NULL) /* empty list, can't match */ | |
| 3120 c = NUL; | |
| 3121 else | |
| 3122 for (;;) /* find "c" in the list */ | |
| 3123 { | |
| 3124 if (*ip == 0) /* not found */ | |
| 3125 { | |
| 3126 c = NUL; | |
| 3127 break; | |
| 3128 } | |
| 3129 if (*ip == c) /* match! */ | |
| 3130 { | |
| 3131 c = ip[1]; | |
| 3132 break; | |
| 3133 } | |
| 3134 ip += 2; | |
| 3135 } | |
| 3136 } | |
| 3137 | |
| 3138 if (c != NUL && c != prevc) | |
| 3139 { | |
| 3140 ri += mb_char2bytes(c, res + ri); | |
| 3141 if (ri + MB_MAXBYTES > MAXWLEN) | |
| 3142 break; | |
| 3143 prevc = c; | |
| 3144 } | |
| 3145 } | |
| 3146 } | |
| 3147 else | |
| 3148 { | |
| 3149 /* The sl_sal_first[] table contains the translation. */ | |
| 3150 for (s = inword; (c = *s) != NUL; ++s) | |
| 3151 { | |
|
11129
f4ea50924c6d
patch 8.0.0452: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11127
diff
changeset
|
3152 if (VIM_ISWHITE(c)) |
| 375 | 3153 c = ' '; |
| 3154 else | |
| 3155 c = slang->sl_sal_first[c]; | |
| 3156 if (c != NUL && (ri == 0 || res[ri - 1] != c)) | |
| 3157 res[ri++] = c; | |
| 3158 } | |
| 3159 } | |
| 3160 | |
| 3161 res[ri] = NUL; | |
| 3162 } | |
| 3163 | |
| 3164 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3165 spell_soundfold_sal(slang_T *slang, char_u *inword, char_u *res) |
| 323 | 3166 { |
| 344 | 3167 salitem_T *smp; |
| 323 | 3168 char_u word[MAXWLEN]; |
| 375 | 3169 char_u *s = inword; |
| 323 | 3170 char_u *t; |
| 344 | 3171 char_u *pf; |
| 323 | 3172 int i, j, z; |
| 344 | 3173 int reslen; |
| 323 | 3174 int n, k = 0; |
| 3175 int z0; | |
| 3176 int k0; | |
| 3177 int n0; | |
| 3178 int c; | |
| 3179 int pri; | |
| 3180 int p0 = -333; | |
| 3181 int c0; | |
| 3182 | |
| 324 | 3183 /* Remove accents, if wanted. We actually remove all non-word characters. |
| 375 | 3184 * But keep white space. We need a copy, the word may be changed here. */ |
| 323 | 3185 if (slang->sl_rem_accents) |
| 3186 { | |
| 3187 t = word; | |
| 375 | 3188 while (*s != NUL) |
| 323 | 3189 { |
|
11129
f4ea50924c6d
patch 8.0.0452: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11127
diff
changeset
|
3190 if (VIM_ISWHITE(*s)) |
| 344 | 3191 { |
| 3192 *t++ = ' '; | |
| 3193 s = skipwhite(s); | |
| 3194 } | |
| 323 | 3195 else |
| 3196 { | |
| 5477 | 3197 if (spell_iswordp_nmw(s, curwin)) |
| 323 | 3198 *t++ = *s; |
| 3199 ++s; | |
| 3200 } | |
| 3201 } | |
| 3202 *t = NUL; | |
| 3203 } | |
| 3204 else | |
| 2768 | 3205 vim_strncpy(word, s, MAXWLEN - 1); |
| 323 | 3206 |
| 344 | 3207 smp = (salitem_T *)slang->sl_sal.ga_data; |
| 323 | 3208 |
| 3209 /* | |
| 3210 * This comes from Aspell phonet.cpp. Converted from C++ to C. | |
| 324 | 3211 * Changed to keep spaces. |
| 323 | 3212 */ |
| 344 | 3213 i = reslen = z = 0; |
| 323 | 3214 while ((c = word[i]) != NUL) |
| 3215 { | |
| 344 | 3216 /* Start with the first rule that has the character in the word. */ |
| 323 | 3217 n = slang->sl_sal_first[c]; |
| 3218 z0 = 0; | |
| 3219 | |
| 3220 if (n >= 0) | |
| 3221 { | |
| 3222 /* check all rules for the same letter */ | |
| 344 | 3223 for (; (s = smp[n].sm_lead)[0] == c; ++n) |
| 323 | 3224 { |
| 344 | 3225 /* Quickly skip entries that don't match the word. Most |
| 3226 * entries are less then three chars, optimize for that. */ | |
| 3227 k = smp[n].sm_leadlen; | |
| 3228 if (k > 1) | |
| 323 | 3229 { |
| 344 | 3230 if (word[i + 1] != s[1]) |
| 3231 continue; | |
| 3232 if (k > 2) | |
| 3233 { | |
| 3234 for (j = 2; j < k; ++j) | |
| 3235 if (word[i + j] != s[j]) | |
| 3236 break; | |
| 3237 if (j < k) | |
| 3238 continue; | |
| 3239 } | |
| 323 | 3240 } |
| 3241 | |
| 375 | 3242 if ((pf = smp[n].sm_oneof) != NULL) |
| 323 | 3243 { |
| 375 | 3244 /* Check for match with one of the chars in "sm_oneof". */ |
| 344 | 3245 while (*pf != NUL && *pf != word[i + k]) |
| 3246 ++pf; | |
| 3247 if (*pf == NUL) | |
| 3248 continue; | |
| 3249 ++k; | |
| 323 | 3250 } |
| 344 | 3251 s = smp[n].sm_rules; |
| 3252 pri = 5; /* default priority */ | |
| 323 | 3253 |
| 3254 p0 = *s; | |
| 3255 k0 = k; | |
| 3256 while (*s == '-' && k > 1) | |
| 3257 { | |
| 3258 k--; | |
| 3259 s++; | |
| 3260 } | |
| 3261 if (*s == '<') | |
| 3262 s++; | |
| 344 | 3263 if (VIM_ISDIGIT(*s)) |
| 323 | 3264 { |
| 3265 /* determine priority */ | |
| 3266 pri = *s - '0'; | |
| 3267 s++; | |
| 3268 } | |
| 3269 if (*s == '^' && *(s + 1) == '^') | |
| 3270 s++; | |
| 3271 | |
| 3272 if (*s == NUL | |
| 3273 || (*s == '^' | |
| 324 | 3274 && (i == 0 || !(word[i - 1] == ' ' |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3275 || spell_iswordp(word + i - 1, curwin))) |
| 323 | 3276 && (*(s + 1) != '$' |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3277 || (!spell_iswordp(word + i + k0, curwin)))) |
| 323 | 3278 || (*s == '$' && i > 0 |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3279 && spell_iswordp(word + i - 1, curwin) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3280 && (!spell_iswordp(word + i + k0, curwin)))) |
| 323 | 3281 { |
| 3282 /* search for followup rules, if: */ | |
| 3283 /* followup and k > 1 and NO '-' in searchstring */ | |
| 3284 c0 = word[i + k - 1]; | |
| 3285 n0 = slang->sl_sal_first[c0]; | |
| 3286 | |
| 3287 if (slang->sl_followup && k > 1 && n0 >= 0 | |
| 344 | 3288 && p0 != '-' && word[i + k] != NUL) |
| 323 | 3289 { |
| 3290 /* test follow-up rule for "word[i + k]" */ | |
| 344 | 3291 for ( ; (s = smp[n0].sm_lead)[0] == c0; ++n0) |
| 323 | 3292 { |
| 344 | 3293 /* Quickly skip entries that don't match the word. |
| 3294 * */ | |
| 3295 k0 = smp[n0].sm_leadlen; | |
| 3296 if (k0 > 1) | |
| 323 | 3297 { |
| 344 | 3298 if (word[i + k] != s[1]) |
| 3299 continue; | |
| 3300 if (k0 > 2) | |
| 3301 { | |
| 3302 pf = word + i + k + 1; | |
| 3303 for (j = 2; j < k0; ++j) | |
| 3304 if (*pf++ != s[j]) | |
| 3305 break; | |
| 3306 if (j < k0) | |
| 3307 continue; | |
| 3308 } | |
| 323 | 3309 } |
| 344 | 3310 k0 += k - 1; |
| 3311 | |
| 375 | 3312 if ((pf = smp[n0].sm_oneof) != NULL) |
| 323 | 3313 { |
| 344 | 3314 /* Check for match with one of the chars in |
| 375 | 3315 * "sm_oneof". */ |
| 344 | 3316 while (*pf != NUL && *pf != word[i + k0]) |
| 3317 ++pf; | |
| 3318 if (*pf == NUL) | |
| 3319 continue; | |
| 3320 ++k0; | |
| 323 | 3321 } |
| 344 | 3322 |
| 3323 p0 = 5; | |
| 3324 s = smp[n0].sm_rules; | |
| 323 | 3325 while (*s == '-') |
| 3326 { | |
| 344 | 3327 /* "k0" gets NOT reduced because |
| 3328 * "if (k0 == k)" */ | |
| 323 | 3329 s++; |
| 3330 } | |
| 3331 if (*s == '<') | |
| 3332 s++; | |
| 344 | 3333 if (VIM_ISDIGIT(*s)) |
| 323 | 3334 { |
| 3335 p0 = *s - '0'; | |
| 3336 s++; | |
| 3337 } | |
| 3338 | |
| 3339 if (*s == NUL | |
| 3340 /* *s == '^' cuts */ | |
| 3341 || (*s == '$' | |
| 376 | 3342 && !spell_iswordp(word + i + k0, |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3343 curwin))) |
| 323 | 3344 { |
| 3345 if (k0 == k) | |
| 3346 /* this is just a piece of the string */ | |
| 3347 continue; | |
| 3348 | |
| 3349 if (p0 < pri) | |
| 3350 /* priority too low */ | |
| 3351 continue; | |
| 3352 /* rule fits; stop search */ | |
| 3353 break; | |
| 3354 } | |
| 3355 } | |
| 3356 | |
| 344 | 3357 if (p0 >= pri && smp[n0].sm_lead[0] == c0) |
| 323 | 3358 continue; |
| 3359 } | |
| 3360 | |
| 3361 /* replace string */ | |
| 344 | 3362 s = smp[n].sm_to; |
| 389 | 3363 if (s == NULL) |
| 3364 s = (char_u *)""; | |
| 344 | 3365 pf = smp[n].sm_rules; |
| 3366 p0 = (vim_strchr(pf, '<') != NULL) ? 1 : 0; | |
| 323 | 3367 if (p0 == 1 && z == 0) |
| 3368 { | |
| 3369 /* rule with '<' is used */ | |
| 344 | 3370 if (reslen > 0 && *s != NUL && (res[reslen - 1] == c |
| 3371 || res[reslen - 1] == *s)) | |
| 3372 reslen--; | |
| 323 | 3373 z0 = 1; |
| 3374 z = 1; | |
| 3375 k0 = 0; | |
| 372 | 3376 while (*s != NUL && word[i + k0] != NUL) |
| 323 | 3377 { |
| 3378 word[i + k0] = *s; | |
| 3379 k0++; | |
| 3380 s++; | |
| 3381 } | |
| 3382 if (k > k0) | |
| 1619 | 3383 STRMOVE(word + i + k0, word + i + k); |
| 323 | 3384 |
| 3385 /* new "actual letter" */ | |
| 3386 c = word[i]; | |
| 3387 } | |
| 3388 else | |
| 3389 { | |
| 3390 /* no '<' rule used */ | |
| 3391 i += k - 1; | |
| 3392 z = 0; | |
| 344 | 3393 while (*s != NUL && s[1] != NUL && reslen < MAXWLEN) |
| 323 | 3394 { |
| 344 | 3395 if (reslen == 0 || res[reslen - 1] != *s) |
| 372 | 3396 res[reslen++] = *s; |
| 323 | 3397 s++; |
| 3398 } | |
| 3399 /* new "actual letter" */ | |
| 3400 c = *s; | |
| 344 | 3401 if (strstr((char *)pf, "^^") != NULL) |
| 323 | 3402 { |
| 3403 if (c != NUL) | |
| 372 | 3404 res[reslen++] = c; |
| 1619 | 3405 STRMOVE(word, word + i + 1); |
| 323 | 3406 i = 0; |
| 3407 z0 = 1; | |
| 3408 } | |
| 3409 } | |
| 3410 break; | |
| 3411 } | |
| 3412 } | |
| 3413 } | |
|
11129
f4ea50924c6d
patch 8.0.0452: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11127
diff
changeset
|
3414 else if (VIM_ISWHITE(c)) |
| 324 | 3415 { |
| 3416 c = ' '; | |
| 3417 k = 1; | |
| 3418 } | |
| 323 | 3419 |
| 3420 if (z0 == 0) | |
| 3421 { | |
| 344 | 3422 if (k && !p0 && reslen < MAXWLEN && c != NUL |
| 3423 && (!slang->sl_collapse || reslen == 0 | |
| 3424 || res[reslen - 1] != c)) | |
| 323 | 3425 /* condense only double letters */ |
| 372 | 3426 res[reslen++] = c; |
| 323 | 3427 |
| 3428 i++; | |
| 3429 z = 0; | |
| 3430 k = 0; | |
| 3431 } | |
| 3432 } | |
| 3433 | |
| 344 | 3434 res[reslen] = NUL; |
| 323 | 3435 } |
| 3436 | |
| 372 | 3437 /* |
| 3438 * Turn "inword" into its sound-a-like equivalent in "res[MAXWLEN]". | |
| 3439 * Multi-byte version of spell_soundfold(). | |
| 3440 */ | |
| 3441 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3442 spell_soundfold_wsal(slang_T *slang, char_u *inword, char_u *res) |
| 372 | 3443 { |
| 375 | 3444 salitem_T *smp = (salitem_T *)slang->sl_sal.ga_data; |
| 372 | 3445 int word[MAXWLEN]; |
| 3446 int wres[MAXWLEN]; | |
| 3447 int l; | |
| 3448 char_u *s; | |
| 3449 int *ws; | |
| 3450 char_u *t; | |
| 3451 int *pf; | |
| 3452 int i, j, z; | |
| 3453 int reslen; | |
| 3454 int n, k = 0; | |
| 3455 int z0; | |
| 3456 int k0; | |
| 3457 int n0; | |
| 3458 int c; | |
| 3459 int pri; | |
| 3460 int p0 = -333; | |
| 3461 int c0; | |
| 3462 int did_white = FALSE; | |
| 3520 | 3463 int wordlen; |
| 3464 | |
| 372 | 3465 |
| 3466 /* | |
| 3467 * Convert the multi-byte string to a wide-character string. | |
| 3468 * Remove accents, if wanted. We actually remove all non-word characters. | |
| 3469 * But keep white space. | |
| 3470 */ | |
| 3520 | 3471 wordlen = 0; |
| 372 | 3472 for (s = inword; *s != NUL; ) |
| 3473 { | |
| 3474 t = s; | |
| 474 | 3475 c = mb_cptr2char_adv(&s); |
| 372 | 3476 if (slang->sl_rem_accents) |
| 3477 { | |
|
11129
f4ea50924c6d
patch 8.0.0452: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11127
diff
changeset
|
3478 if (enc_utf8 ? utf_class(c) == 0 : VIM_ISWHITE(c)) |
| 372 | 3479 { |
| 3480 if (did_white) | |
| 3481 continue; | |
| 3482 c = ' '; | |
| 3483 did_white = TRUE; | |
| 3484 } | |
| 3485 else | |
| 3486 { | |
| 3487 did_white = FALSE; | |
| 5477 | 3488 if (!spell_iswordp_nmw(t, curwin)) |
| 372 | 3489 continue; |
| 3490 } | |
| 3491 } | |
| 3520 | 3492 word[wordlen++] = c; |
| 3493 } | |
| 3494 word[wordlen] = NUL; | |
| 372 | 3495 |
| 3496 /* | |
| 3520 | 3497 * This algorithm comes from Aspell phonet.cpp. |
| 372 | 3498 * Converted from C++ to C. Added support for multi-byte chars. |
| 3499 * Changed to keep spaces. | |
| 3500 */ | |
| 3501 i = reslen = z = 0; | |
| 3502 while ((c = word[i]) != NUL) | |
| 3503 { | |
| 3504 /* Start with the first rule that has the character in the word. */ | |
| 3505 n = slang->sl_sal_first[c & 0xff]; | |
| 3506 z0 = 0; | |
| 3507 | |
| 3508 if (n >= 0) | |
| 3509 { | |
|
2455
9367de3e2e1b
Fix: crash in spell checking with a 0x300 character.
Bram Moolenaar <bram@vim.org>
parents:
2454
diff
changeset
|
3510 /* Check all rules for the same index byte. |
|
9367de3e2e1b
Fix: crash in spell checking with a 0x300 character.
Bram Moolenaar <bram@vim.org>
parents:
2454
diff
changeset
|
3511 * If c is 0x300 need extra check for the end of the array, as |
|
9367de3e2e1b
Fix: crash in spell checking with a 0x300 character.
Bram Moolenaar <bram@vim.org>
parents:
2454
diff
changeset
|
3512 * (c & 0xff) is NUL. */ |
|
9367de3e2e1b
Fix: crash in spell checking with a 0x300 character.
Bram Moolenaar <bram@vim.org>
parents:
2454
diff
changeset
|
3513 for (; ((ws = smp[n].sm_lead_w)[0] & 0xff) == (c & 0xff) |
|
9367de3e2e1b
Fix: crash in spell checking with a 0x300 character.
Bram Moolenaar <bram@vim.org>
parents:
2454
diff
changeset
|
3514 && ws[0] != NUL; ++n) |
| 372 | 3515 { |
| 3516 /* Quickly skip entries that don't match the word. Most | |
| 3517 * entries are less then three chars, optimize for that. */ | |
| 375 | 3518 if (c != ws[0]) |
| 3519 continue; | |
| 372 | 3520 k = smp[n].sm_leadlen; |
| 3521 if (k > 1) | |
| 3522 { | |
| 3523 if (word[i + 1] != ws[1]) | |
| 3524 continue; | |
| 3525 if (k > 2) | |
| 3526 { | |
| 3527 for (j = 2; j < k; ++j) | |
| 3528 if (word[i + j] != ws[j]) | |
| 3529 break; | |
| 3530 if (j < k) | |
| 3531 continue; | |
| 3532 } | |
| 3533 } | |
| 3534 | |
| 375 | 3535 if ((pf = smp[n].sm_oneof_w) != NULL) |
| 372 | 3536 { |
| 375 | 3537 /* Check for match with one of the chars in "sm_oneof". */ |
| 372 | 3538 while (*pf != NUL && *pf != word[i + k]) |
| 3539 ++pf; | |
| 3540 if (*pf == NUL) | |
| 3541 continue; | |
| 3542 ++k; | |
| 3543 } | |
| 3544 s = smp[n].sm_rules; | |
| 3545 pri = 5; /* default priority */ | |
| 3546 | |
| 3547 p0 = *s; | |
| 3548 k0 = k; | |
| 3549 while (*s == '-' && k > 1) | |
| 3550 { | |
| 3551 k--; | |
| 3552 s++; | |
| 3553 } | |
| 3554 if (*s == '<') | |
| 3555 s++; | |
| 3556 if (VIM_ISDIGIT(*s)) | |
| 3557 { | |
| 3558 /* determine priority */ | |
| 3559 pri = *s - '0'; | |
| 3560 s++; | |
| 3561 } | |
| 3562 if (*s == '^' && *(s + 1) == '^') | |
| 3563 s++; | |
| 3564 | |
| 3565 if (*s == NUL | |
| 3566 || (*s == '^' | |
| 3567 && (i == 0 || !(word[i - 1] == ' ' | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3568 || spell_iswordp_w(word + i - 1, curwin))) |
| 372 | 3569 && (*(s + 1) != '$' |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3570 || (!spell_iswordp_w(word + i + k0, curwin)))) |
| 372 | 3571 || (*s == '$' && i > 0 |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3572 && spell_iswordp_w(word + i - 1, curwin) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3573 && (!spell_iswordp_w(word + i + k0, curwin)))) |
| 372 | 3574 { |
| 3575 /* search for followup rules, if: */ | |
| 3576 /* followup and k > 1 and NO '-' in searchstring */ | |
| 3577 c0 = word[i + k - 1]; | |
| 3578 n0 = slang->sl_sal_first[c0 & 0xff]; | |
| 3579 | |
| 3580 if (slang->sl_followup && k > 1 && n0 >= 0 | |
| 3581 && p0 != '-' && word[i + k] != NUL) | |
| 3582 { | |
| 375 | 3583 /* Test follow-up rule for "word[i + k]"; loop over |
| 3584 * all entries with the same index byte. */ | |
| 372 | 3585 for ( ; ((ws = smp[n0].sm_lead_w)[0] & 0xff) |
| 3586 == (c0 & 0xff); ++n0) | |
| 3587 { | |
| 3588 /* Quickly skip entries that don't match the word. | |
| 375 | 3589 */ |
| 3590 if (c0 != ws[0]) | |
| 3591 continue; | |
| 372 | 3592 k0 = smp[n0].sm_leadlen; |
| 3593 if (k0 > 1) | |
| 3594 { | |
| 3595 if (word[i + k] != ws[1]) | |
| 3596 continue; | |
| 3597 if (k0 > 2) | |
| 3598 { | |
| 3599 pf = word + i + k + 1; | |
| 3600 for (j = 2; j < k0; ++j) | |
| 3601 if (*pf++ != ws[j]) | |
| 3602 break; | |
| 3603 if (j < k0) | |
| 3604 continue; | |
| 3605 } | |
| 3606 } | |
| 3607 k0 += k - 1; | |
| 3608 | |
| 375 | 3609 if ((pf = smp[n0].sm_oneof_w) != NULL) |
| 372 | 3610 { |
| 3611 /* Check for match with one of the chars in | |
| 375 | 3612 * "sm_oneof". */ |
| 372 | 3613 while (*pf != NUL && *pf != word[i + k0]) |
| 3614 ++pf; | |
| 3615 if (*pf == NUL) | |
| 3616 continue; | |
| 3617 ++k0; | |
| 3618 } | |
| 3619 | |
| 3620 p0 = 5; | |
| 3621 s = smp[n0].sm_rules; | |
| 3622 while (*s == '-') | |
| 3623 { | |
| 3624 /* "k0" gets NOT reduced because | |
| 3625 * "if (k0 == k)" */ | |
| 3626 s++; | |
| 3627 } | |
| 3628 if (*s == '<') | |
| 3629 s++; | |
| 3630 if (VIM_ISDIGIT(*s)) | |
| 3631 { | |
| 3632 p0 = *s - '0'; | |
| 3633 s++; | |
| 3634 } | |
| 3635 | |
| 3636 if (*s == NUL | |
| 3637 /* *s == '^' cuts */ | |
| 3638 || (*s == '$' | |
| 376 | 3639 && !spell_iswordp_w(word + i + k0, |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3640 curwin))) |
| 372 | 3641 { |
| 3642 if (k0 == k) | |
| 3643 /* this is just a piece of the string */ | |
| 3644 continue; | |
| 3645 | |
| 3646 if (p0 < pri) | |
| 3647 /* priority too low */ | |
| 3648 continue; | |
| 3649 /* rule fits; stop search */ | |
| 3650 break; | |
| 3651 } | |
| 3652 } | |
| 3653 | |
| 3654 if (p0 >= pri && (smp[n0].sm_lead_w[0] & 0xff) | |
| 3655 == (c0 & 0xff)) | |
| 3656 continue; | |
| 3657 } | |
| 3658 | |
| 3659 /* replace string */ | |
| 3660 ws = smp[n].sm_to_w; | |
| 3661 s = smp[n].sm_rules; | |
| 3662 p0 = (vim_strchr(s, '<') != NULL) ? 1 : 0; | |
| 3663 if (p0 == 1 && z == 0) | |
| 3664 { | |
| 3665 /* rule with '<' is used */ | |
| 389 | 3666 if (reslen > 0 && ws != NULL && *ws != NUL |
| 3667 && (wres[reslen - 1] == c | |
| 372 | 3668 || wres[reslen - 1] == *ws)) |
| 3669 reslen--; | |
| 3670 z0 = 1; | |
| 3671 z = 1; | |
| 3672 k0 = 0; | |
| 389 | 3673 if (ws != NULL) |
| 3674 while (*ws != NUL && word[i + k0] != NUL) | |
| 3675 { | |
| 3676 word[i + k0] = *ws; | |
| 3677 k0++; | |
| 3678 ws++; | |
| 3679 } | |
| 372 | 3680 if (k > k0) |
| 3681 mch_memmove(word + i + k0, word + i + k, | |
| 3520 | 3682 sizeof(int) * (wordlen - (i + k) + 1)); |
| 372 | 3683 |
| 3684 /* new "actual letter" */ | |
| 3685 c = word[i]; | |
| 3686 } | |
| 3687 else | |
| 3688 { | |
| 3689 /* no '<' rule used */ | |
| 3690 i += k - 1; | |
| 3691 z = 0; | |
| 389 | 3692 if (ws != NULL) |
| 3693 while (*ws != NUL && ws[1] != NUL | |
| 3694 && reslen < MAXWLEN) | |
| 3695 { | |
| 3696 if (reslen == 0 || wres[reslen - 1] != *ws) | |
| 3697 wres[reslen++] = *ws; | |
| 3698 ws++; | |
| 3699 } | |
| 372 | 3700 /* new "actual letter" */ |
| 389 | 3701 if (ws == NULL) |
| 3702 c = NUL; | |
| 3703 else | |
| 3704 c = *ws; | |
| 372 | 3705 if (strstr((char *)s, "^^") != NULL) |
| 3706 { | |
| 3707 if (c != NUL) | |
| 3708 wres[reslen++] = c; | |
| 3709 mch_memmove(word, word + i + 1, | |
| 3520 | 3710 sizeof(int) * (wordlen - (i + 1) + 1)); |
| 372 | 3711 i = 0; |
| 3712 z0 = 1; | |
| 3713 } | |
| 3714 } | |
| 3715 break; | |
| 3716 } | |
| 3717 } | |
| 3718 } | |
|
11129
f4ea50924c6d
patch 8.0.0452: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11127
diff
changeset
|
3719 else if (VIM_ISWHITE(c)) |
| 372 | 3720 { |
| 3721 c = ' '; | |
| 3722 k = 1; | |
| 3723 } | |
| 3724 | |
| 3725 if (z0 == 0) | |
| 3726 { | |
| 3727 if (k && !p0 && reslen < MAXWLEN && c != NUL | |
| 3728 && (!slang->sl_collapse || reslen == 0 | |
| 3729 || wres[reslen - 1] != c)) | |
| 3730 /* condense only double letters */ | |
| 3731 wres[reslen++] = c; | |
| 3732 | |
| 3733 i++; | |
| 3734 z = 0; | |
| 3735 k = 0; | |
| 3736 } | |
| 3737 } | |
| 3738 | |
| 3739 /* Convert wide characters in "wres" to a multi-byte string in "res". */ | |
| 3740 l = 0; | |
| 3741 for (n = 0; n < reslen; ++n) | |
| 3742 { | |
| 3743 l += mb_char2bytes(wres[n], res + l); | |
| 3744 if (l + MB_MAXBYTES > MAXWLEN) | |
| 3745 break; | |
| 3746 } | |
| 3747 res[l] = NUL; | |
| 3748 } | |
| 3749 | |
| 324 | 3750 /* |
| 714 | 3751 * ":spellinfo" |
| 3752 */ | |
| 3753 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3754 ex_spellinfo(exarg_T *eap UNUSED) |
| 714 | 3755 { |
| 3756 int lpi; | |
| 3757 langp_T *lp; | |
| 3758 char_u *p; | |
| 3759 | |
| 3760 if (no_spell_checking(curwin)) | |
| 3761 return; | |
| 3762 | |
| 3763 msg_start(); | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3764 for (lpi = 0; lpi < curwin->w_s->b_langp.ga_len && !got_int; ++lpi) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3765 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3766 lp = LANGP_ENTRY(curwin->w_s->b_langp, lpi); |
|
15543
dd725a8ab112
patch 8.1.0779: argument for message functions is inconsistent
Bram Moolenaar <Bram@vim.org>
parents:
15470
diff
changeset
|
3767 msg_puts("file: "); |
|
dd725a8ab112
patch 8.1.0779: argument for message functions is inconsistent
Bram Moolenaar <Bram@vim.org>
parents:
15470
diff
changeset
|
3768 msg_puts((char *)lp->lp_slang->sl_fname); |
| 714 | 3769 msg_putchar('\n'); |
| 3770 p = lp->lp_slang->sl_info; | |
| 3771 if (p != NULL) | |
| 3772 { | |
|
15543
dd725a8ab112
patch 8.1.0779: argument for message functions is inconsistent
Bram Moolenaar <Bram@vim.org>
parents:
15470
diff
changeset
|
3773 msg_puts((char *)p); |
| 714 | 3774 msg_putchar('\n'); |
| 3775 } | |
| 3776 } | |
| 3777 msg_end(); | |
| 3778 } | |
| 3779 | |
| 625 | 3780 #define DUMPFLAG_KEEPCASE 1 /* round 2: keep-case tree */ |
| 3781 #define DUMPFLAG_COUNT 2 /* include word count */ | |
| 701 | 3782 #define DUMPFLAG_ICASE 4 /* ignore case when finding matches */ |
| 710 | 3783 #define DUMPFLAG_ONECAP 8 /* pattern starts with capital */ |
| 3784 #define DUMPFLAG_ALLCAP 16 /* pattern is all capitals */ | |
| 625 | 3785 |
| 351 | 3786 /* |
| 3787 * ":spelldump" | |
| 3788 */ | |
| 3789 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3790 ex_spelldump(exarg_T *eap) |
| 351 | 3791 { |
| 5382 | 3792 char_u *spl; |
| 3793 long dummy; | |
| 3794 | |
| 701 | 3795 if (no_spell_checking(curwin)) |
| 3796 return; | |
| 5382 | 3797 get_option_value((char_u*)"spl", &dummy, &spl, OPT_LOCAL); |
| 3798 | |
| 3799 /* Create a new empty buffer in a new window. */ | |
| 701 | 3800 do_cmdline_cmd((char_u *)"new"); |
| 5382 | 3801 |
| 3802 /* enable spelling locally in the new window */ | |
| 3803 set_option_value((char_u*)"spell", TRUE, (char_u*)"", OPT_LOCAL); | |
|
7447
ad432f8f68fb
commit https://github.com/vim/vim/commit/887c1fea4a114e7170091942d0446c8882701b5b
Christian Brabandt <cb@256bit.org>
parents:
6949
diff
changeset
|
3804 set_option_value((char_u*)"spl", dummy, spl, OPT_LOCAL); |
| 5382 | 3805 vim_free(spl); |
| 3806 | |
|
11121
778c10516955
patch 8.0.0448: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
10950
diff
changeset
|
3807 if (!BUFEMPTY()) |
| 701 | 3808 return; |
| 3809 | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3810 spell_dump_compl(NULL, 0, NULL, eap->forceit ? DUMPFLAG_COUNT : 0); |
| 701 | 3811 |
| 3812 /* Delete the empty line that we started with. */ | |
| 3813 if (curbuf->b_ml.ml_line_count > 1) | |
| 3814 ml_delete(curbuf->b_ml.ml_line_count, FALSE); | |
| 3815 | |
| 3816 redraw_later(NOT_VALID); | |
| 3817 } | |
| 3818 | |
| 3819 /* | |
| 3820 * Go through all possible words and: | |
| 3821 * 1. When "pat" is NULL: dump a list of all words in the current buffer. | |
| 3822 * "ic" and "dir" are not used. | |
| 3823 * 2. When "pat" is not NULL: add matching words to insert mode completion. | |
| 3824 */ | |
| 3825 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3826 spell_dump_compl( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3827 char_u *pat, /* leading part of the word */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3828 int ic, /* ignore case */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3829 int *dir, /* direction for adding matches */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
3830 int dumpflags_arg) /* DUMPFLAG_* */ |
| 701 | 3831 { |
| 351 | 3832 langp_T *lp; |
| 3833 slang_T *slang; | |
| 3834 idx_T arridx[MAXWLEN]; | |
| 3835 int curi[MAXWLEN]; | |
| 3836 char_u word[MAXWLEN]; | |
| 3837 int c; | |
| 3838 char_u *byts; | |
| 3839 idx_T *idxs; | |
| 3840 linenr_T lnum = 0; | |
| 3841 int round; | |
| 3842 int depth; | |
| 3843 int n; | |
| 3844 int flags; | |
| 381 | 3845 char_u *region_names = NULL; /* region names being used */ |
| 3846 int do_region = TRUE; /* dump region names and numbers */ | |
| 3847 char_u *p; | |
| 500 | 3848 int lpi; |
| 701 | 3849 int dumpflags = dumpflags_arg; |
| 3850 int patlen; | |
| 3851 | |
| 710 | 3852 /* When ignoring case or when the pattern starts with capital pass this on |
| 3853 * to dump_word(). */ | |
| 3854 if (pat != NULL) | |
| 3855 { | |
| 3856 if (ic) | |
| 3857 dumpflags |= DUMPFLAG_ICASE; | |
| 3858 else | |
| 3859 { | |
| 3860 n = captype(pat, NULL); | |
| 3861 if (n == WF_ONECAP) | |
| 3862 dumpflags |= DUMPFLAG_ONECAP; | |
|
15605
62b3805506b3
patch 8.1.0810: too many #ifdefs
Bram Moolenaar <Bram@vim.org>
parents:
15543
diff
changeset
|
3863 else if (n == WF_ALLCAP && (int)STRLEN(pat) > mb_ptr2len(pat)) |
| 710 | 3864 dumpflags |= DUMPFLAG_ALLCAP; |
| 3865 } | |
| 3866 } | |
| 351 | 3867 |
| 381 | 3868 /* Find out if we can support regions: All languages must support the same |
| 3869 * regions or none at all. */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3870 for (lpi = 0; lpi < curwin->w_s->b_langp.ga_len; ++lpi) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3871 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3872 lp = LANGP_ENTRY(curwin->w_s->b_langp, lpi); |
| 381 | 3873 p = lp->lp_slang->sl_regions; |
| 3874 if (p[0] != 0) | |
| 3875 { | |
| 3876 if (region_names == NULL) /* first language with regions */ | |
| 3877 region_names = p; | |
| 3878 else if (STRCMP(region_names, p) != 0) | |
| 3879 { | |
| 3880 do_region = FALSE; /* region names are different */ | |
| 3881 break; | |
| 3882 } | |
| 3883 } | |
| 3884 } | |
| 3885 | |
| 3886 if (do_region && region_names != NULL) | |
| 3887 { | |
| 701 | 3888 if (pat == NULL) |
| 3889 { | |
| 3890 vim_snprintf((char *)IObuff, IOSIZE, "/regions=%s", region_names); | |
| 3891 ml_append(lnum++, IObuff, (colnr_T)0, FALSE); | |
| 3892 } | |
| 381 | 3893 } |
| 3894 else | |
| 3895 do_region = FALSE; | |
| 3896 | |
| 3897 /* | |
| 3898 * Loop over all files loaded for the entries in 'spelllang'. | |
| 3899 */ | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3900 for (lpi = 0; lpi < curwin->w_s->b_langp.ga_len; ++lpi) |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3901 { |
|
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
3902 lp = LANGP_ENTRY(curwin->w_s->b_langp, lpi); |
| 351 | 3903 slang = lp->lp_slang; |
| 500 | 3904 if (slang->sl_fbyts == NULL) /* reloading failed */ |
| 3905 continue; | |
| 351 | 3906 |
| 701 | 3907 if (pat == NULL) |
| 3908 { | |
| 3909 vim_snprintf((char *)IObuff, IOSIZE, "# file: %s", slang->sl_fname); | |
| 3910 ml_append(lnum++, IObuff, (colnr_T)0, FALSE); | |
| 3911 } | |
| 3912 | |
| 3913 /* When matching with a pattern and there are no prefixes only use | |
| 3914 * parts of the tree that match "pat". */ | |
| 3915 if (pat != NULL && slang->sl_pbyts == NULL) | |
| 835 | 3916 patlen = (int)STRLEN(pat); |
| 701 | 3917 else |
| 840 | 3918 patlen = -1; |
| 351 | 3919 |
| 3920 /* round 1: case-folded tree | |
| 3921 * round 2: keep-case tree */ | |
| 3922 for (round = 1; round <= 2; ++round) | |
| 3923 { | |
| 3924 if (round == 1) | |
| 3925 { | |
| 701 | 3926 dumpflags &= ~DUMPFLAG_KEEPCASE; |
| 351 | 3927 byts = slang->sl_fbyts; |
| 3928 idxs = slang->sl_fidxs; | |
| 3929 } | |
| 3930 else | |
| 3931 { | |
| 701 | 3932 dumpflags |= DUMPFLAG_KEEPCASE; |
| 351 | 3933 byts = slang->sl_kbyts; |
| 3934 idxs = slang->sl_kidxs; | |
| 3935 } | |
| 3936 if (byts == NULL) | |
| 3937 continue; /* array is empty */ | |
| 3938 | |
| 3939 depth = 0; | |
| 3940 arridx[0] = 0; | |
| 3941 curi[0] = 1; | |
| 701 | 3942 while (depth >= 0 && !got_int |
|
16142
570a296aa0b4
patch 8.1.1076: file for Insert mode is much too big
Bram Moolenaar <Bram@vim.org>
parents:
15636
diff
changeset
|
3943 && (pat == NULL || !ins_compl_interrupted())) |
| 351 | 3944 { |
| 3945 if (curi[depth] > byts[arridx[depth]]) | |
| 3946 { | |
| 3947 /* Done all bytes at this node, go up one level. */ | |
| 3948 --depth; | |
| 3949 line_breakcheck(); | |
|
10277
154d5a2e7395
commit https://github.com/vim/vim/commit/472e85970ee3a80abd824bef510df12e9cfe9e96
Christian Brabandt <cb@256bit.org>
parents:
10042
diff
changeset
|
3950 ins_compl_check_keys(50, FALSE); |
| 351 | 3951 } |
| 3952 else | |
| 3953 { | |
| 3954 /* Do one more byte at this node. */ | |
| 3955 n = arridx[depth] + curi[depth]; | |
| 3956 ++curi[depth]; | |
| 3957 c = byts[n]; | |
| 3958 if (c == 0) | |
| 3959 { | |
| 3960 /* End of word, deal with the word. | |
| 3961 * Don't use keep-case words in the fold-case tree, | |
| 3962 * they will appear in the keep-case tree. | |
| 3963 * Only use the word when the region matches. */ | |
| 3964 flags = (int)idxs[n]; | |
| 3965 if ((round == 2 || (flags & WF_KEEPCAP) == 0) | |
| 500 | 3966 && (flags & WF_NEEDCOMP) == 0 |
| 381 | 3967 && (do_region |
| 3968 || (flags & WF_REGION) == 0 | |
| 390 | 3969 || (((unsigned)flags >> 16) |
| 351 | 3970 & lp->lp_region) != 0)) |
| 3971 { | |
| 3972 word[depth] = NUL; | |
| 381 | 3973 if (!do_region) |
| 3974 flags &= ~WF_REGION; | |
| 355 | 3975 |
| 3976 /* Dump the basic word if there is no prefix or | |
| 3977 * when it's the first one. */ | |
| 390 | 3978 c = (unsigned)flags >> 24; |
| 355 | 3979 if (c == 0 || curi[depth] == 2) |
| 701 | 3980 { |
| 3981 dump_word(slang, word, pat, dir, | |
| 3982 dumpflags, flags, lnum); | |
| 3983 if (pat == NULL) | |
| 3984 ++lnum; | |
| 3985 } | |
| 351 | 3986 |
| 3987 /* Apply the prefix, if there is one. */ | |
| 355 | 3988 if (c != 0) |
| 701 | 3989 lnum = dump_prefixes(slang, word, pat, dir, |
| 3990 dumpflags, flags, lnum); | |
| 351 | 3991 } |
| 3992 } | |
| 3993 else | |
| 3994 { | |
| 3995 /* Normal char, go one level deeper. */ | |
| 3996 word[depth++] = c; | |
| 3997 arridx[depth] = idxs[n]; | |
| 3998 curi[depth] = 1; | |
| 701 | 3999 |
| 4000 /* Check if this characters matches with the pattern. | |
| 4001 * If not skip the whole tree below it. | |
| 710 | 4002 * Always ignore case here, dump_word() will check |
| 4003 * proper case later. This isn't exactly right when | |
| 4004 * length changes for multi-byte characters with | |
| 4005 * ignore case... */ | |
| 4006 if (depth <= patlen | |
| 4007 && MB_STRNICMP(word, pat, depth) != 0) | |
| 701 | 4008 --depth; |
| 351 | 4009 } |
| 4010 } | |
| 4011 } | |
| 4012 } | |
| 4013 } | |
| 4014 } | |
| 4015 | |
| 4016 /* | |
| 4017 * Dump one word: apply case modifications and append a line to the buffer. | |
| 701 | 4018 * When "lnum" is zero add insert mode completion. |
| 4019 */ | |
| 4020 static void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4021 dump_word( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4022 slang_T *slang, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4023 char_u *word, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4024 char_u *pat, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4025 int *dir, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4026 int dumpflags, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4027 int wordflags, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4028 linenr_T lnum) |
| 351 | 4029 { |
| 4030 int keepcap = FALSE; | |
| 4031 char_u *p; | |
| 625 | 4032 char_u *tw; |
| 351 | 4033 char_u cword[MAXWLEN]; |
| 381 | 4034 char_u badword[MAXWLEN + 10]; |
| 4035 int i; | |
| 710 | 4036 int flags = wordflags; |
| 4037 | |
| 4038 if (dumpflags & DUMPFLAG_ONECAP) | |
| 4039 flags |= WF_ONECAP; | |
| 4040 if (dumpflags & DUMPFLAG_ALLCAP) | |
| 4041 flags |= WF_ALLCAP; | |
| 351 | 4042 |
| 625 | 4043 if ((dumpflags & DUMPFLAG_KEEPCASE) == 0 && (flags & WF_CAPMASK) != 0) |
| 351 | 4044 { |
| 4045 /* Need to fix case according to "flags". */ | |
| 4046 make_case_word(word, cword, flags); | |
| 4047 p = cword; | |
| 4048 } | |
| 4049 else | |
| 4050 { | |
| 4051 p = word; | |
| 625 | 4052 if ((dumpflags & DUMPFLAG_KEEPCASE) |
| 4053 && ((captype(word, NULL) & WF_KEEPCAP) == 0 | |
| 389 | 4054 || (flags & WF_FIXCAP) != 0)) |
| 351 | 4055 keepcap = TRUE; |
| 4056 } | |
| 625 | 4057 tw = p; |
| 351 | 4058 |
| 701 | 4059 if (pat == NULL) |
| 4060 { | |
| 4061 /* Add flags and regions after a slash. */ | |
| 4062 if ((flags & (WF_BANNED | WF_RARE | WF_REGION)) || keepcap) | |
| 4063 { | |
| 4064 STRCPY(badword, p); | |
| 4065 STRCAT(badword, "/"); | |
| 4066 if (keepcap) | |
| 4067 STRCAT(badword, "="); | |
| 4068 if (flags & WF_BANNED) | |
| 4069 STRCAT(badword, "!"); | |
| 4070 else if (flags & WF_RARE) | |
| 4071 STRCAT(badword, "?"); | |
| 4072 if (flags & WF_REGION) | |
| 4073 for (i = 0; i < 7; ++i) | |
| 4074 if (flags & (0x10000 << i)) | |
| 4075 sprintf((char *)badword + STRLEN(badword), "%d", i + 1); | |
| 4076 p = badword; | |
| 4077 } | |
| 4078 | |
| 4079 if (dumpflags & DUMPFLAG_COUNT) | |
| 4080 { | |
| 4081 hashitem_T *hi; | |
| 4082 | |
| 4083 /* Include the word count for ":spelldump!". */ | |
| 4084 hi = hash_find(&slang->sl_wordcount, tw); | |
| 4085 if (!HASHITEM_EMPTY(hi)) | |
| 4086 { | |
| 4087 vim_snprintf((char *)IObuff, IOSIZE, "%s\t%d", | |
| 625 | 4088 tw, HI2WC(hi)->wc_count); |
| 701 | 4089 p = IObuff; |
| 4090 } | |
| 4091 } | |
| 4092 | |
| 4093 ml_append(lnum, p, (colnr_T)0, FALSE); | |
| 4094 } | |
| 710 | 4095 else if (((dumpflags & DUMPFLAG_ICASE) |
| 4096 ? MB_STRNICMP(p, pat, STRLEN(pat)) == 0 | |
| 4097 : STRNCMP(p, pat, STRLEN(pat)) == 0) | |
| 701 | 4098 && ins_compl_add_infercase(p, (int)STRLEN(p), |
|
16239
5df26b29e809
patch 8.1.1124: insert completion flags are mixed up
Bram Moolenaar <Bram@vim.org>
parents:
16162
diff
changeset
|
4099 p_ic, NULL, *dir, FALSE) == OK) |
| 710 | 4100 /* if dir was BACKWARD then honor it just once */ |
| 4101 *dir = FORWARD; | |
| 351 | 4102 } |
| 4103 | |
| 4104 /* | |
| 372 | 4105 * For ":spelldump": Find matching prefixes for "word". Prepend each to |
| 4106 * "word" and append a line to the buffer. | |
| 701 | 4107 * When "lnum" is zero add insert mode completion. |
| 351 | 4108 * Return the updated line number. |
| 4109 */ | |
| 4110 static linenr_T | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4111 dump_prefixes( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4112 slang_T *slang, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4113 char_u *word, /* case-folded word */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4114 char_u *pat, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4115 int *dir, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4116 int dumpflags, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4117 int flags, /* flags with prefix ID */ |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4118 linenr_T startlnum) |
| 351 | 4119 { |
| 4120 idx_T arridx[MAXWLEN]; | |
| 4121 int curi[MAXWLEN]; | |
| 4122 char_u prefix[MAXWLEN]; | |
| 455 | 4123 char_u word_up[MAXWLEN]; |
| 4124 int has_word_up = FALSE; | |
| 351 | 4125 int c; |
| 4126 char_u *byts; | |
| 4127 idx_T *idxs; | |
| 4128 linenr_T lnum = startlnum; | |
| 4129 int depth; | |
| 4130 int n; | |
| 4131 int len; | |
| 4132 int i; | |
| 4133 | |
| 626 | 4134 /* If the word starts with a lower-case letter make the word with an |
| 455 | 4135 * upper-case letter in word_up[]. */ |
| 4136 c = PTR2CHAR(word); | |
| 4137 if (SPELL_TOUPPER(c) != c) | |
| 4138 { | |
| 4139 onecap_copy(word, word_up, TRUE); | |
| 4140 has_word_up = TRUE; | |
| 4141 } | |
| 4142 | |
| 351 | 4143 byts = slang->sl_pbyts; |
| 4144 idxs = slang->sl_pidxs; | |
| 4145 if (byts != NULL) /* array not is empty */ | |
| 4146 { | |
| 4147 /* | |
| 4148 * Loop over all prefixes, building them byte-by-byte in prefix[]. | |
| 390 | 4149 * When at the end of a prefix check that it supports "flags". |
| 351 | 4150 */ |
| 4151 depth = 0; | |
| 4152 arridx[0] = 0; | |
| 4153 curi[0] = 1; | |
| 4154 while (depth >= 0 && !got_int) | |
| 4155 { | |
| 390 | 4156 n = arridx[depth]; |
| 4157 len = byts[n]; | |
| 4158 if (curi[depth] > len) | |
| 351 | 4159 { |
| 4160 /* Done all bytes at this node, go up one level. */ | |
| 4161 --depth; | |
| 4162 line_breakcheck(); | |
| 4163 } | |
| 4164 else | |
| 4165 { | |
| 4166 /* Do one more byte at this node. */ | |
| 390 | 4167 n += curi[depth]; |
| 351 | 4168 ++curi[depth]; |
| 4169 c = byts[n]; | |
| 4170 if (c == 0) | |
| 4171 { | |
| 4172 /* End of prefix, find out how many IDs there are. */ | |
| 4173 for (i = 1; i < len; ++i) | |
| 4174 if (byts[n + i] != 0) | |
| 4175 break; | |
| 4176 curi[depth] += i - 1; | |
| 4177 | |
| 455 | 4178 c = valid_word_prefix(i, n, flags, word, slang, FALSE); |
| 4179 if (c != 0) | |
| 351 | 4180 { |
| 376 | 4181 vim_strncpy(prefix + depth, word, MAXWLEN - depth - 1); |
| 701 | 4182 dump_word(slang, prefix, pat, dir, dumpflags, |
| 455 | 4183 (c & WF_RAREPFX) ? (flags | WF_RARE) |
| 701 | 4184 : flags, lnum); |
| 4185 if (lnum != 0) | |
| 4186 ++lnum; | |
| 351 | 4187 } |
| 455 | 4188 |
| 4189 /* Check for prefix that matches the word when the | |
| 4190 * first letter is upper-case, but only if the prefix has | |
| 4191 * a condition. */ | |
| 4192 if (has_word_up) | |
| 4193 { | |
| 4194 c = valid_word_prefix(i, n, flags, word_up, slang, | |
| 4195 TRUE); | |
| 4196 if (c != 0) | |
| 4197 { | |
| 4198 vim_strncpy(prefix + depth, word_up, | |
| 4199 MAXWLEN - depth - 1); | |
| 701 | 4200 dump_word(slang, prefix, pat, dir, dumpflags, |
| 455 | 4201 (c & WF_RAREPFX) ? (flags | WF_RARE) |
| 701 | 4202 : flags, lnum); |
| 4203 if (lnum != 0) | |
| 4204 ++lnum; | |
| 455 | 4205 } |
| 4206 } | |
| 351 | 4207 } |
| 4208 else | |
| 4209 { | |
| 4210 /* Normal char, go one level deeper. */ | |
| 4211 prefix[depth++] = c; | |
| 4212 arridx[depth] = idxs[n]; | |
| 4213 curi[depth] = 1; | |
| 4214 } | |
| 4215 } | |
| 4216 } | |
| 4217 } | |
| 4218 | |
| 4219 return lnum; | |
| 4220 } | |
| 4221 | |
| 498 | 4222 /* |
| 626 | 4223 * Move "p" to the end of word "start". |
| 4224 * Uses the spell-checking word characters. | |
| 498 | 4225 */ |
| 4226 char_u * | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4227 spell_to_word_end(char_u *start, win_T *win) |
| 498 | 4228 { |
| 4229 char_u *p = start; | |
| 4230 | |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
4231 while (*p != NUL && spell_iswordp(p, win)) |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
4232 MB_PTR_ADV(p); |
| 498 | 4233 return p; |
| 4234 } | |
| 4235 | |
| 475 | 4236 /* |
| 626 | 4237 * For Insert mode completion CTRL-X s: |
| 4238 * Find start of the word in front of column "startcol". | |
| 4239 * We don't check if it is badly spelled, with completion we can only change | |
| 4240 * the word in front of the cursor. | |
| 475 | 4241 * Returns the column number of the word. |
| 4242 */ | |
| 4243 int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4244 spell_word_start(int startcol) |
| 475 | 4245 { |
| 4246 char_u *line; | |
| 4247 char_u *p; | |
| 4248 int col = 0; | |
| 4249 | |
| 498 | 4250 if (no_spell_checking(curwin)) |
| 475 | 4251 return startcol; |
| 4252 | |
| 4253 /* Find a word character before "startcol". */ | |
| 4254 line = ml_get_curline(); | |
| 4255 for (p = line + startcol; p > line; ) | |
| 4256 { | |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
4257 MB_PTR_BACK(line, p); |
| 5477 | 4258 if (spell_iswordp_nmw(p, curwin)) |
| 475 | 4259 break; |
| 4260 } | |
| 4261 | |
| 4262 /* Go back to start of the word. */ | |
| 4263 while (p > line) | |
| 4264 { | |
| 835 | 4265 col = (int)(p - line); |
|
11127
506f5d8b7d8b
patch 8.0.0451: some macros are in lower case
Christian Brabandt <cb@256bit.org>
parents:
11121
diff
changeset
|
4266 MB_PTR_BACK(line, p); |
|
2250
1bac28a53fae
Add the conceal patch from Vince Negri.
Bram Moolenaar <bram@vim.org>
parents:
2229
diff
changeset
|
4267 if (!spell_iswordp(p, curwin)) |
| 475 | 4268 break; |
| 4269 col = 0; | |
| 4270 } | |
| 4271 | |
| 535 | 4272 return col; |
| 4273 } | |
| 4274 | |
| 4275 /* | |
| 4276 * Need to check for 'spellcapcheck' now, the word is removed before | |
| 4277 * expand_spelling() is called. Therefore the ugly global variable. | |
| 4278 */ | |
| 4279 static int spell_expand_need_cap; | |
| 4280 | |
| 4281 void | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4282 spell_expand_check_cap(colnr_T col) |
| 535 | 4283 { |
| 475 | 4284 spell_expand_need_cap = check_need_cap(curwin->w_cursor.lnum, col); |
| 4285 } | |
| 4286 | |
| 4287 /* | |
| 4288 * Get list of spelling suggestions. | |
| 4289 * Used for Insert mode completion CTRL-X ?. | |
| 4290 * Returns the number of matches. The matches are in "matchp[]", array of | |
| 4291 * allocated strings. | |
| 4292 */ | |
| 4293 int | |
|
7835
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4294 expand_spelling( |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4295 linenr_T lnum UNUSED, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4296 char_u *pat, |
|
4d7ce6c03fda
commit https://github.com/vim/vim/commit/764b23c8fd3369cb05ae9122abf3ca16fec539d7
Christian Brabandt <cb@256bit.org>
parents:
7805
diff
changeset
|
4297 char_u ***matchp) |
| 475 | 4298 { |
| 4299 garray_T ga; | |
| 4300 | |
| 625 | 4301 spell_suggest_list(&ga, pat, 100, spell_expand_need_cap, TRUE); |
| 475 | 4302 *matchp = ga.ga_data; |
| 4303 return ga.ga_len; | |
| 4304 } | |
| 4305 | |
|
17940
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4306 /* |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4307 * Return TRUE if "val" is a valid 'spellang' value. |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4308 */ |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4309 int |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4310 valid_spellang(char_u *val) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4311 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4312 return valid_name(val, ".-_,@"); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4313 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4314 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4315 /* |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4316 * Return TRUE if "val" is a valid 'spellfile' value. |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4317 */ |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4318 int |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4319 valid_spellfile(char_u *val) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4320 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4321 char_u *s; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4322 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4323 for (s = val; *s != NUL; ++s) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4324 if (!vim_isfilec(*s) && *s != ',') |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4325 return FALSE; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4326 return TRUE; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4327 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4328 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4329 /* |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4330 * Handle side effects of setting 'spell'. |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4331 * Return an error message or NULL for success. |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4332 */ |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4333 char * |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4334 did_set_spell_option(int is_spellfile) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4335 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4336 char *errmsg = NULL; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4337 win_T *wp; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4338 int l; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4339 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4340 if (is_spellfile) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4341 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4342 l = (int)STRLEN(curwin->w_s->b_p_spf); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4343 if (l > 0 && (l < 4 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4344 || STRCMP(curwin->w_s->b_p_spf + l - 4, ".add") != 0)) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4345 errmsg = e_invarg; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4346 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4347 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4348 if (errmsg == NULL) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4349 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4350 FOR_ALL_WINDOWS(wp) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4351 if (wp->w_buffer == curbuf && wp->w_p_spell) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4352 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4353 errmsg = did_set_spelllang(wp); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4354 break; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4355 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4356 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4357 return errmsg; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4358 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4359 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4360 /* |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4361 * Set curbuf->b_cap_prog to the regexp program for 'spellcapcheck'. |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4362 * Return error message when failed, NULL when OK. |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4363 */ |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4364 char * |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4365 compile_cap_prog(synblock_T *synblock) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4366 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4367 regprog_T *rp = synblock->b_cap_prog; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4368 char_u *re; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4369 |
|
18305
f4db8631d9c5
patch 8.1.2147: crash when allocating memory fails
Bram Moolenaar <Bram@vim.org>
parents:
18251
diff
changeset
|
4370 if (synblock->b_p_spc == NULL || *synblock->b_p_spc == NUL) |
|
17940
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4371 synblock->b_cap_prog = NULL; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4372 else |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4373 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4374 // Prepend a ^ so that we only match at one column |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4375 re = concat_str((char_u *)"^", synblock->b_p_spc); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4376 if (re != NULL) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4377 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4378 synblock->b_cap_prog = vim_regcomp(re, RE_MAGIC); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4379 vim_free(re); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4380 if (synblock->b_cap_prog == NULL) |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4381 { |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4382 synblock->b_cap_prog = rp; // restore the previous program |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4383 return e_invarg; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4384 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4385 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4386 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4387 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4388 vim_regfree(rp); |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4389 return NULL; |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4390 } |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4391 |
|
079e10a49ea1
patch 8.1.1966: some code in options.c fits better elsewhere
Bram Moolenaar <Bram@vim.org>
parents:
17809
diff
changeset
|
4392 #endif // FEAT_SPELL |