xref: /vim-8.2.3635/src/spell.h (revision fb094e14)
1 /* vi:set ts=8 sts=4 sw=4 noet:
2  *
3  * VIM - Vi IMproved	by Bram Moolenaar
4  *
5  * Do ":help uganda"  in Vim to read copying and usage conditions.
6  * Do ":help credits" in Vim to see a list of people who contributed.
7  * See README.txt for an overview of the Vim source code.
8  */
9 
10 /*
11  * spell.h: common code for spell checking, used by spell.c and spellfile.c.
12  */
13 
14 /* Use SPELL_PRINTTREE for debugging: dump the word tree after adding a word.
15  * Only use it for small word lists! */
16 #if 0
17 # define SPELL_PRINTTREE
18 #endif
19 
20 /* Use SPELL_COMPRESS_ALLWAYS for debugging: compress the word tree after
21  * adding a word.  Only use it for small word lists! */
22 #if 0
23 # define SPELL_COMPRESS_ALLWAYS
24 #endif
25 
26 /* Use DEBUG_TRIEWALK to print the changes made in suggest_trie_walk() for a
27  * specific word. */
28 #if 0
29 # define DEBUG_TRIEWALK
30 #endif
31 
32 #define MAXWLEN 254		/* Assume max. word len is this many bytes.
33 				   Some places assume a word length fits in a
34 				   byte, thus it can't be above 255.
35 				   Must be >= PFD_NOTSPECIAL. */
36 
37 /* Type used for indexes in the word tree need to be at least 4 bytes.  If int
38  * is 8 bytes we could use something smaller, but what? */
39 #if VIM_SIZEOF_INT > 3
40 typedef int idx_T;
41 #else
42 typedef long idx_T;
43 #endif
44 
45 #ifdef FEAT_MBYTE
46 typedef int salfirst_T;
47 #else
48 typedef short salfirst_T;
49 #endif
50 
51 /*
52  * Structure used to store words and other info for one language, loaded from
53  * a .spl file.
54  * The main access is through the tree in "sl_fbyts/sl_fidxs", storing the
55  * case-folded words.  "sl_kbyts/sl_kidxs" is for keep-case words.
56  *
57  * The "byts" array stores the possible bytes in each tree node, preceded by
58  * the number of possible bytes, sorted on byte value:
59  *	<len> <byte1> <byte2> ...
60  * The "idxs" array stores the index of the child node corresponding to the
61  * byte in "byts".
62  * Exception: when the byte is zero, the word may end here and "idxs" holds
63  * the flags, region mask and affixID for the word.  There may be several
64  * zeros in sequence for alternative flag/region/affixID combinations.
65  */
66 typedef struct slang_S slang_T;
67 struct slang_S
68 {
69     slang_T	*sl_next;	/* next language */
70     char_u	*sl_name;	/* language name "en", "en.rare", "nl", etc. */
71     char_u	*sl_fname;	/* name of .spl file */
72     int		sl_add;		/* TRUE if it's a .add file. */
73 
74     char_u	*sl_fbyts;	/* case-folded word bytes */
75     idx_T	*sl_fidxs;	/* case-folded word indexes */
76     char_u	*sl_kbyts;	/* keep-case word bytes */
77     idx_T	*sl_kidxs;	/* keep-case word indexes */
78     char_u	*sl_pbyts;	/* prefix tree word bytes */
79     idx_T	*sl_pidxs;	/* prefix tree word indexes */
80 
81     char_u	*sl_info;	/* infotext string or NULL */
82 
83     char_u	sl_regions[17];	/* table with up to 8 region names plus NUL */
84 
85     char_u	*sl_midword;	/* MIDWORD string or NULL */
86 
87     hashtab_T	sl_wordcount;	/* hashtable with word count, wordcount_T */
88 
89     int		sl_compmax;	/* COMPOUNDWORDMAX (default: MAXWLEN) */
90     int		sl_compminlen;	/* COMPOUNDMIN (default: 0) */
91     int		sl_compsylmax;	/* COMPOUNDSYLMAX (default: MAXWLEN) */
92     int		sl_compoptions;	/* COMP_* flags */
93     garray_T	sl_comppat;	/* CHECKCOMPOUNDPATTERN items */
94     regprog_T	*sl_compprog;	/* COMPOUNDRULE turned into a regexp progrm
95 				 * (NULL when no compounding) */
96     char_u	*sl_comprules;	/* all COMPOUNDRULE concatenated (or NULL) */
97     char_u	*sl_compstartflags; /* flags for first compound word */
98     char_u	*sl_compallflags; /* all flags for compound words */
99     char_u	sl_nobreak;	/* When TRUE: no spaces between words */
100     char_u	*sl_syllable;	/* SYLLABLE repeatable chars or NULL */
101     garray_T	sl_syl_items;	/* syllable items */
102 
103     int		sl_prefixcnt;	/* number of items in "sl_prefprog" */
104     regprog_T	**sl_prefprog;	/* table with regprogs for prefixes */
105 
106     garray_T	sl_rep;		/* list of fromto_T entries from REP lines */
107     short	sl_rep_first[256];  /* indexes where byte first appears, -1 if
108 				       there is none */
109     garray_T	sl_sal;		/* list of salitem_T entries from SAL lines */
110     salfirst_T	sl_sal_first[256];  /* indexes where byte first appears, -1 if
111 				       there is none */
112     int		sl_followup;	/* SAL followup */
113     int		sl_collapse;	/* SAL collapse_result */
114     int		sl_rem_accents;	/* SAL remove_accents */
115     int		sl_sofo;	/* SOFOFROM and SOFOTO instead of SAL items:
116 				 * "sl_sal_first" maps chars, when has_mbyte
117 				 * "sl_sal" is a list of wide char lists. */
118     garray_T	sl_repsal;	/* list of fromto_T entries from REPSAL lines */
119     short	sl_repsal_first[256];  /* sl_rep_first for REPSAL lines */
120     int		sl_nosplitsugs;	/* don't suggest splitting a word */
121     int		sl_nocompoundsugs; /* don't suggest compounding */
122 
123     /* Info from the .sug file.  Loaded on demand. */
124     time_t	sl_sugtime;	/* timestamp for .sug file */
125     char_u	*sl_sbyts;	/* soundfolded word bytes */
126     idx_T	*sl_sidxs;	/* soundfolded word indexes */
127     buf_T	*sl_sugbuf;	/* buffer with word number table */
128     int		sl_sugloaded;	/* TRUE when .sug file was loaded or failed to
129 				   load */
130 
131     int		sl_has_map;	/* TRUE if there is a MAP line */
132 #ifdef FEAT_MBYTE
133     hashtab_T	sl_map_hash;	/* MAP for multi-byte chars */
134     int		sl_map_array[256]; /* MAP for first 256 chars */
135 #else
136     char_u	sl_map_array[256]; /* MAP for first 256 chars */
137 #endif
138     hashtab_T	sl_sounddone;	/* table with soundfolded words that have
139 				   handled, see add_sound_suggest() */
140 };
141 
142 #ifdef VMS
143 # define SPL_FNAME_TMPL  "%s_%s.spl"
144 # define SPL_FNAME_ADD   "_add."
145 # define SPL_FNAME_ASCII "_ascii."
146 #else
147 # define SPL_FNAME_TMPL  "%s.%s.spl"
148 # define SPL_FNAME_ADD   ".add."
149 # define SPL_FNAME_ASCII ".ascii."
150 #endif
151 
152 /* Flags used for a word.  Only the lowest byte can be used, the region byte
153  * comes above it. */
154 #define WF_REGION   0x01	/* region byte follows */
155 #define WF_ONECAP   0x02	/* word with one capital (or all capitals) */
156 #define WF_ALLCAP   0x04	/* word must be all capitals */
157 #define WF_RARE	    0x08	/* rare word */
158 #define WF_BANNED   0x10	/* bad word */
159 #define WF_AFX	    0x20	/* affix ID follows */
160 #define WF_FIXCAP   0x40	/* keep-case word, allcap not allowed */
161 #define WF_KEEPCAP  0x80	/* keep-case word */
162 
163 /* for <flags2>, shifted up one byte to be used in wn_flags */
164 #define WF_HAS_AFF  0x0100	/* word includes affix */
165 #define WF_NEEDCOMP 0x0200	/* word only valid in compound */
166 #define WF_NOSUGGEST 0x0400	/* word not to be suggested */
167 #define WF_COMPROOT 0x0800	/* already compounded word, COMPOUNDROOT */
168 #define WF_NOCOMPBEF 0x1000	/* no compounding before this word */
169 #define WF_NOCOMPAFT 0x2000	/* no compounding after this word */
170 
171 /* flags for <pflags> */
172 #define WFP_RARE	    0x01	/* rare prefix */
173 #define WFP_NC		    0x02	/* prefix is not combining */
174 #define WFP_UP		    0x04	/* to-upper prefix */
175 #define WFP_COMPPERMIT	    0x08	/* prefix with COMPOUNDPERMITFLAG */
176 #define WFP_COMPFORBID	    0x10	/* prefix with COMPOUNDFORBIDFLAG */
177 
178 /* Flags for postponed prefixes in "sl_pidxs".  Must be above affixID (one
179  * byte) and prefcondnr (two bytes). */
180 #define WF_RAREPFX  (WFP_RARE << 24)	/* rare postponed prefix */
181 #define WF_PFX_NC   (WFP_NC << 24)	/* non-combining postponed prefix */
182 #define WF_PFX_UP   (WFP_UP << 24)	/* to-upper postponed prefix */
183 #define WF_PFX_COMPPERMIT (WFP_COMPPERMIT << 24) /* postponed prefix with
184 						  * COMPOUNDPERMITFLAG */
185 #define WF_PFX_COMPFORBID (WFP_COMPFORBID << 24) /* postponed prefix with
186 						  * COMPOUNDFORBIDFLAG */
187 
188 /* flags for <compoptions> */
189 #define COMP_CHECKDUP		1	/* CHECKCOMPOUNDDUP */
190 #define COMP_CHECKREP		2	/* CHECKCOMPOUNDREP */
191 #define COMP_CHECKCASE		4	/* CHECKCOMPOUNDCASE */
192 #define COMP_CHECKTRIPLE	8	/* CHECKCOMPOUNDTRIPLE */
193 
194 /* Info from "REP", "REPSAL" and "SAL" entries in ".aff" file used in si_rep,
195  * si_repsal, sl_rep, and si_sal.  Not for sl_sal!
196  * One replacement: from "ft_from" to "ft_to". */
197 typedef struct fromto_S
198 {
199     char_u	*ft_from;
200     char_u	*ft_to;
201 } fromto_T;
202 
203 /* Info from "SAL" entries in ".aff" file used in sl_sal.
204  * The info is split for quick processing by spell_soundfold().
205  * Note that "sm_oneof" and "sm_rules" point into sm_lead. */
206 typedef struct salitem_S
207 {
208     char_u	*sm_lead;	/* leading letters */
209     int		sm_leadlen;	/* length of "sm_lead" */
210     char_u	*sm_oneof;	/* letters from () or NULL */
211     char_u	*sm_rules;	/* rules like ^, $, priority */
212     char_u	*sm_to;		/* replacement. */
213 #ifdef FEAT_MBYTE
214     int		*sm_lead_w;	/* wide character copy of "sm_lead" */
215     int		*sm_oneof_w;	/* wide character copy of "sm_oneof" */
216     int		*sm_to_w;	/* wide character copy of "sm_to" */
217 #endif
218 } salitem_T;
219 
220 /* Values for SP_*ERROR are negative, positive values are used by
221  * read_cnt_string(). */
222 #define	SP_TRUNCERROR	-1	/* spell file truncated error */
223 #define	SP_FORMERROR	-2	/* format error in spell file */
224 #define SP_OTHERERROR	-3	/* other error while reading spell file */
225 
226 /*
227  * Structure used in "b_langp", filled from 'spelllang'.
228  */
229 typedef struct langp_S
230 {
231     slang_T	*lp_slang;	/* info for this language */
232     slang_T	*lp_sallang;	/* language used for sound folding or NULL */
233     slang_T	*lp_replang;	/* language used for REP items or NULL */
234     int		lp_region;	/* bitmask for region or REGION_ALL */
235 } langp_T;
236 
237 #define LANGP_ENTRY(ga, i)	(((langp_T *)(ga).ga_data) + (i))
238 
239 #define VIMSUGMAGIC "VIMsug"	/* string at start of Vim .sug file */
240 #define VIMSUGMAGICL 6
241 #define VIMSUGVERSION 1
242 
243 /*
244  * The tables used for recognizing word characters according to spelling.
245  * These are only used for the first 256 characters of 'encoding'.
246  */
247 typedef struct spelltab_S
248 {
249     char_u  st_isw[256];	/* flags: is word char */
250     char_u  st_isu[256];	/* flags: is uppercase char */
251     char_u  st_fold[256];	/* chars: folded case */
252     char_u  st_upper[256];	/* chars: upper case */
253 } spelltab_T;
254 
255 /*
256  * Use our own character-case definitions, because the current locale may
257  * differ from what the .spl file uses.
258  * These must not be called with negative number!
259  */
260 #ifndef FEAT_MBYTE
261 /* Non-multi-byte implementation. */
262 # define SPELL_TOFOLD(c) ((c) < 256 ? (int)spelltab.st_fold[c] : (c))
263 # define SPELL_TOUPPER(c) ((c) < 256 ? (int)spelltab.st_upper[c] : (c))
264 # define SPELL_ISUPPER(c) ((c) < 256 ? spelltab.st_isu[c] : FALSE)
265 #else
266 # if defined(HAVE_WCHAR_H)
267 #  include <wchar.h>	    /* for towupper() and towlower() */
268 # endif
269 /* Multi-byte implementation.  For Unicode we can call utf_*(), but don't do
270  * that for ASCII, because we don't want to use 'casemap' here.  Otherwise use
271  * the "w" library function for characters above 255 if available. */
272 # ifdef HAVE_TOWLOWER
273 #  define SPELL_TOFOLD(c) (enc_utf8 && (c) >= 128 ? utf_fold(c) \
274 	    : (c) < 256 ? (int)spelltab.st_fold[c] : (int)towlower(c))
275 # else
276 #  define SPELL_TOFOLD(c) (enc_utf8 && (c) >= 128 ? utf_fold(c) \
277 	    : (c) < 256 ? (int)spelltab.st_fold[c] : (c))
278 # endif
279 
280 # ifdef HAVE_TOWUPPER
281 #  define SPELL_TOUPPER(c) (enc_utf8 && (c) >= 128 ? utf_toupper(c) \
282 	    : (c) < 256 ? (int)spelltab.st_upper[c] : (int)towupper(c))
283 # else
284 #  define SPELL_TOUPPER(c) (enc_utf8 && (c) >= 128 ? utf_toupper(c) \
285 	    : (c) < 256 ? (int)spelltab.st_upper[c] : (c))
286 # endif
287 
288 # ifdef HAVE_ISWUPPER
289 #  define SPELL_ISUPPER(c) (enc_utf8 && (c) >= 128 ? utf_isupper(c) \
290 	    : (c) < 256 ? spelltab.st_isu[c] : iswupper(c))
291 # else
292 #  define SPELL_ISUPPER(c) (enc_utf8 && (c) >= 128 ? utf_isupper(c) \
293 	    : (c) < 256 ? spelltab.st_isu[c] : (FALSE))
294 # endif
295 #endif
296 
297 #ifdef FEAT_SPELL
298 /* First language that is loaded, start of the linked list of loaded
299  * languages. */
300 # ifdef IN_SPELL_C
301 #  define SPELL_EXTERN
302 #  define SPELL_INIT(x) x
303 # else
304 #  define SPELL_EXTERN extern
305 #  define SPELL_INIT(x)
306 # endif
307 
308 SPELL_EXTERN slang_T	*first_lang SPELL_INIT(= NULL);
309 
310 /* file used for "zG" and "zW" */
311 SPELL_EXTERN char_u	*int_wordlist SPELL_INIT(= NULL);
312 
313 
314 SPELL_EXTERN char e_format[] SPELL_INIT(= N_("E759: Format error in spell file"));
315 
316 SPELL_EXTERN spelltab_T   spelltab;
317 SPELL_EXTERN int	  did_set_spelltab;
318 
319 #endif
320