diff options
| -rw-r--r-- | runtime/doc/insert.txt | 9 | ||||
| -rw-r--r-- | runtime/doc/map.txt | 10 | ||||
| -rw-r--r-- | runtime/spell/en.ascii.spl | bin | 566603 -> 566616 bytes | |||
| -rw-r--r-- | runtime/spell/en.utf-8.spl | bin | 569129 -> 569151 bytes | |||
| -rw-r--r-- | runtime/spell/hu/hu_HU.diff | 78 | ||||
| -rw-r--r-- | runtime/spell/main.aap | 9 | ||||
| -rw-r--r-- | src/spell.c | 2117 |
7 files changed, 1487 insertions, 736 deletions
diff --git a/runtime/doc/insert.txt b/runtime/doc/insert.txt index 43b3d4f416..ae90eb954c 100644 --- a/runtime/doc/insert.txt +++ b/runtime/doc/insert.txt @@ -1,4 +1,4 @@ -*insert.txt* For Vim version 7.0aa. Last change: 2005 Aug 11 +*insert.txt* For Vim version 7.0aa. Last change: 2005 Aug 17 VIM REFERENCE MANUAL by Bram Moolenaar @@ -899,8 +899,11 @@ CTRL-X CTRL-O Guess what kind of item is in front of the cursor and Spelling suggestions *compl-spelling* -The word in front of the cursor is located and correctly spelled words are -suggested to replace it. The word doesn't actually have to be badly spelled. +A word before or at the cursor is located and correctly spelled words are +suggested to replace it. If there is a badly spelled word in the line, before +or under the cursor, the cursor is moved to after it. Otherwise the word just +before the cursor is used for suggestions, even though it isn't badly spelled. + NOTE: CTRL-S suspends display in many Unix terminals. Use 's' instead. Type CTRL-Q to resume displaying. diff --git a/runtime/doc/map.txt b/runtime/doc/map.txt index a135a481ab..5fb03e9165 100644 --- a/runtime/doc/map.txt +++ b/runtime/doc/map.txt @@ -281,10 +281,7 @@ last defined. Example: > n <C-W>* * <C-W><C-S>* Last set from /home/abcd/.vimrc -When the map was defined by hand there is no "Last set" message. When the map -was defined while executing a function, user command or autocommand, the -script in which it was defined is reported. -{not available when compiled without the +eval feature} +See |:verbose-cmd| for more information. *map_backslash* Note that only CTRL-V is mentioned here as a special character for mappings @@ -877,10 +874,7 @@ last defined. Example: > TOhtml 0 % :call Convert2HTML(<line1>, <line2>) Last set from /usr/share/vim/vim-7.0/plugin/tohtml.vim < -When the command was defined by hand there is no "Last set" message. When the -command was defined while executing a function, user command or autocommand, -the script in which it was defined is reported. -{not available when compiled without the +eval feature} +See |:verbose-cmd| for more information. *E174* *E182* :com[mand][!] [{attr}...] {cmd} {rep} diff --git a/runtime/spell/en.ascii.spl b/runtime/spell/en.ascii.spl Binary files differindex 9d2b068c44..360e1ae9ce 100644 --- a/runtime/spell/en.ascii.spl +++ b/runtime/spell/en.ascii.spl diff --git a/runtime/spell/en.utf-8.spl b/runtime/spell/en.utf-8.spl Binary files differindex 8b2edc6923..9ba853cebf 100644 --- a/runtime/spell/en.utf-8.spl +++ b/runtime/spell/en.utf-8.spl diff --git a/runtime/spell/hu/hu_HU.diff b/runtime/spell/hu/hu_HU.diff index de22c794fe..2b9a6ddf0b 100644 --- a/runtime/spell/hu/hu_HU.diff +++ b/runtime/spell/hu/hu_HU.diff @@ -1,5 +1,5 @@ *** hu_HU.orig.aff Tue Aug 16 18:21:10 2005 ---- hu_HU.aff Tue Aug 16 19:42:34 2005 +--- hu_HU.aff Fri Aug 19 21:28:45 2005 *************** *** 57,62 **** @@ -16,8 +16,9 @@ ! #VERSION Magyar 0.99.4.2 SET ISO8859-2 *************** -*** 65,77 **** - COMPOUNDFLAG Y +*** 64,77 **** + COMPOUNDMIN 2 +! COMPOUNDFLAG Y ! COMPOUNDWORD 2 y ! COMPOUNDSYLLABLE 6 aáeéiíoóöõuúüû ! SYLLABLENUM klmc @@ -30,24 +31,85 @@ ! WORDCHARS -.§%°0123456789 ! HU_KOTOHANGZO Z ---- 65,80 ---- - COMPOUNDFLAG Y +--- 64,116 ---- + COMPOUNDMIN 2 ! #COMPOUNDWORD 2 y +! COMPOUNDMAX 2 +! # I don't understand what the "y" is for; if it's to disable compounding simply +! # remove the compound flag from the word. +! ! #COMPOUNDSYLLABLE 6 aáeéiíoóöõuúüû +! COMPOUNDSYLMAX 6 +! SYLLABLE a/á/e/é/i/í/o/ó/ö/õ/u/ú/ü/û +! # Strange that every vowel is counted as a syllable, that's how the hunspell +! # code works. +! ! #SYLLABLENUM klmc +! # Don't understand what this is for +! +! #COMPOUNDFLAG Y ! #COMPOUNDFIRST v ! #COMPOUNDLAST x +! COMPOUNDFLAGS Y+ +! COMPOUNDFLAGS vY*x +! COMPOUNDFLAGS Y+x +! COMPOUNDFLAGS vY+ +! ! #FORBIDDENWORD w -! BAD w +! # I don't understand what FORBIDDENWORD is needed for, using NEEDAFFIX +! # (ONLYROOT) should be sufficient. +! ! #ONLYROOT u +! NEEDAFFIX u +! ! #ACCENT áéíóöõúüû aeiooouuu +! MAP 5 +! MAP aáä +! MAP eé +! MAP ií +! MAP oóöõ +! MAP uúüû +! ! #CHECKNUM +! # Vim always handles numbers in the same way. +! ! #WORDCHARS -.§%°0123456789 +! FOL ±¢³µ¶¨¹º»¼¾¿±²³´µ¶·¸¹º»¼½¾¿àáâãäåæçèéêëìíîïðñòóôõöøùúûüýþßÿ-§%° +! LOW ±¢³µ¶¨¹º»¼¾¿±²³´µ¶·¸¹º»¼½¾¿àáâãäåæçèéêëìíîïðñòóôõöøùúûüýþßÿ-§%° +! UPP ¡¢£¥¦¨©ª«¬®¯±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞßÿ-§%° +! MIDWORD . +! ! #HU_KOTOHANGZO Z ! +! # There are soooo many affixes. Postpone the prefixes to keep the time needed +! # for generating the .spl within reasonable limits. ! PFXPOSTPONE *************** +*** 81,96 **** + +! REP 89 +! REP í i +! REP i í +! REP ó o +! REP o ó +! REP o õ +! REP ú u +! REP u ú +! REP u û +! REP û ü +! REP ü û + REP j ly + REP ly j +- REP a ä # Handel->Händel + REP S © # Skoda->©koda +--- 120,124 ---- + +! REP 78 + REP j ly + REP ly j + REP S © # Skoda->©koda +*************** *** 173,241 **** - # character conversion table @@ -118,11 +180,11 @@ - CHR 123 U3 Û - SFX z Y 6 ---- 176,177 ---- +--- 201,202 ---- *************** *** 17678,17681 **** PFX D 0 leg . - - 1 - ---- 17614 ---- +--- 17639 ---- diff --git a/runtime/spell/main.aap b/runtime/spell/main.aap index 1e37c1cc3d..5aaaaa1022 100644 --- a/runtime/spell/main.aap +++ b/runtime/spell/main.aap @@ -4,19 +4,22 @@ # aap generate all the .spl files # aap diff create all the diff files -LANG = af bg ca cs da de el en eo fr fo gl he hr it nl ny pl sk yi +LANG = af bg ca cs da de el en eo fr fo gl he hr it nl ny pl sk yi hu +# "hu" is at the end, because it takes so much time. +# # TODO: # Finnish doesn't work, the dictionary fi_FI.zip file contains hyphenation... diff: $*LANG/diff - :print done + :print Done. @for l in string.split(_no.LANG): :child $l/main.aap # The existing .spl files need to be generated when the spell file format # changes. Depending on the Vim executable does that, but results in doing it -# much too often. Generate a dummy .spl file and check if it changed. +# much too often. Generate a dummy .spl file and let the .spl depend on it, so +# that they are outdated when it changes. :child check/main.aap *.spl: check/check.latin1.spl diff --git a/src/spell.c b/src/spell.c index a3fc24f720..30b08ae982 100644 --- a/src/spell.c +++ b/src/spell.c @@ -56,16 +56,6 @@ # define SPELL_PRINTTREE #endif -/* SPELL_COMPRESS_CNT is after how many allocated blocks we compress the tree - * to limit the amount of memory used (esp. for Italian and Hungarian). The - * amount of memory used for nodes then is SPELL_COMPRESS_CNT times - * SBLOCKSIZE. - * Then compress again after allocating SPELL_COMPRESS_INC more blocks or - * adding SPELL_COMPRESS_ADDED words and running out of memory again. */ -#define SPELL_COMPRESS_CNT 30000 -#define SPELL_COMPRESS_INC 100 -#define SPELL_COMPRESS_ADDED 500000 - /* * Use this to adjust the score after finding suggestions, based on the * suggested word sounding like the bad word. This is much faster than doing @@ -78,94 +68,102 @@ /* * Vim spell file format: <HEADER> - * <SUGGEST> + * <SECTIONS> * <LWORDTREE> * <KWORDTREE> * <PREFIXTREE> * - * <HEADER>: <fileID> - * <regioncnt> <regionname> ... - * <charflagslen> <charflags> - * <fcharslen> <fchars> - * <midwordlen> <midword> - * <compoundlen> <compoundtype> <compoundinfo> - * <prefcondcnt> <prefcond> ... + * <HEADER>: <fileID> <versionnr> * - * <fileID> 10 bytes "VIMspell10" - * <regioncnt> 1 byte number of regions following (8 supported) - * <regionname> 2 bytes Region name: ca, au, etc. Lower case. - * First <regionname> is region 1. + * <fileID> 8 bytes "VIMspell" + * <versionnr> 1 byte VIMSPELLVERSION * - * <charflagslen> 1 byte Number of bytes in <charflags> (should be 128). - * <charflags> N bytes List of flags (first one is for character 128): - * 0x01 word character CF_WORD - * 0x02 upper-case character CF_UPPER - * <fcharslen> 2 bytes Number of bytes in <fchars>. - * <fchars> N bytes Folded characters, first one is for character 128. * - * <midwordlen> 2 bytes Number of bytes in <midword>. - * <midword> N bytes Characters that are word characters only when used - * in the middle of a word. + * Sections make it possible to add information to the .spl file without + * making it incompatible with previous versions. There are two kinds of + * sections: + * 1. Not essential for correct spell checking. E.g. for making suggestions. + * These are skipped when not supported. + * 2. Optional information, but essential for spell checking when present. + * E.g. conditions for affixes. When this section is present but not + * supported an error message is given. * - * <compoundlen> 2 bytes Number of bytes following for compound info (can - * be used to skip it when it's not understood). + * <SECTIONS>: <section> ... <sectionend> * - * <compoundtype 1 byte 1: compound words using <comp1minlen> and - * <comp1flags> + * <section>: <sectionID> <sectionflags> <sectionlen> (section contents) * - * <comp1minlen> 1 byte minimal word length for compounding + * <sectionID> 1 byte number from 0 to 254 identifying the section * - * <comp1flags> N bytes flags used for compounding words + * <sectionflags> 1 byte SNF_REQUIRED: this section is required for correct + * spell checking * + * <sectionlen> 4 bytes length of section contents, MSB first * - * <prefcondcnt> 2 bytes Number of <prefcond> items following. - * - * <prefcond> : <condlen> <condstr> + * <sectionend> 1 byte SN_END * - * <condlen> 1 byte Length of <condstr>. * - * <condstr> N bytes Condition for the prefix. + * sectionID == SN_REGION: <regionname> ... + * <regionname> 2 bytes Up to 8 region names: ca, au, etc. Lower case. + * First <regionname> is region 1. * + * sectionID == SN_CHARFLAGS: <charflagslen> <charflags> + * <folcharslen> <folchars> + * <charflagslen> 1 byte Number of bytes in <charflags> (should be 128). + * <charflags> N bytes List of flags (first one is for character 128): + * 0x01 word character CF_WORD + * 0x02 upper-case character CF_UPPER + * <folcharslen> 2 bytes Number of bytes in <folchars>. + * <folchars> N bytes Folded characters, first one is for character 128. * - * <SUGGEST> : <repcount> <rep> ... - * <salflags> <salcount> <sal> ... - * <maplen> <mapstr> + * sectionID == SN_MIDWORD: <midword> + * <midword> N bytes Characters that are word characters only when used + * in the middle of a word. * - * <repcount> 2 bytes number of <rep> items, MSB first. + * sectionID == SN_PREFCOND: <prefcondcnt> <prefcond> ... + * <prefcondcnt> 2 bytes Number of <prefcond> items following. + * <prefcond> : <condlen> <condstr> + * <condlen> 1 byte Length of <condstr>. + * <condstr> N bytes Condition for the prefix. * + * sectionID == SN_REP: <repcount> <rep> ... + * <repcount> 2 bytes number of <rep> items, MSB first. * <rep> : <repfromlen> <repfrom> <reptolen> <repto> + * <repfromlen> 1 byte length of <repfrom> + * <repfrom> N bytes "from" part of replacement + * <reptolen> 1 byte length of <repto> + * <repto> N bytes "to" part of replacement * - * <repfromlen> 1 byte length of <repfrom> - * - * <repfrom> N bytes "from" part of replacement - * - * <reptolen> 1 byte length of <repto> - * - * <repto> N bytes "to" part of replacement - * - * <salflags> 1 byte flags for soundsalike conversion: + * sectionID == SN_SAL: <salflags> <salcount> <sal> ... + * <salflags> 1 byte flags for soundsalike conversion: * SAL_F0LLOWUP * SAL_COLLAPSE * SAL_REM_ACCENTS - * SAL_SOFO: SOFOFROM and SOFOTO used instead of SAL - * - * <salcount> 2 bytes number of <sal> items following - * + * <salcount> 2 bytes number of <sal> items following * <sal> : <salfromlen> <salfrom> <saltolen> <salto> + * <salfromlen> 1 byte length of <salfrom> + * <salfrom> N bytes "from" part of soundsalike + * <saltolen> 1 byte length of <salto> + * <salto> N bytes "to" part of soundsalike * - * <salfromlen> 1-2 bytes length of <salfrom> (2 bytes for SAL_SOFO) - * - * <salfrom> N bytes "from" part of soundsalike - * - * <saltolen> 1-2 bytes length of <salto> (2 bytes for SAL_SOFO) - * - * <salto> N bytes "to" part of soundsalike + * sectionID == SN_SOFO: <sofofromlen> <sofofrom> <sofotolen> <sofoto> + * <sofofromlen> 2 bytes length of <sofofrom> + * <sofofrom> N bytes "from" part of soundfold + * <sofotolen> 2 bytes length of <sofoto> + * <sofoto> N bytes "to" part of soundfold * - * <maplen> 2 bytes length of <mapstr>, MSB first - * - * <mapstr> N bytes String with sequences of similar characters, + * sectionID == SN_MAP: <mapstr> + * <mapstr> N bytes String with sequences of similar characters, * separated by slashes. * + * sectionID == SN_COMPOUND: <compmax> <compminlen> <compsylmax> <compflags> + * <compmax> 1 byte Maximum nr of words in compound word. + * <compminlen> 1 byte Minimal word length for compounding. + * <compsylmax> 1 byte Maximum nr of syllables in compound word. + * <compflags> N bytes Flags from COMPOUNDFLAGS items, separated by + * slashes. + * + * sectionID == SN_SYLLABLE: <syllable> + * <syllable> N bytes String from SYLLABLE item. * * <LWORDTREE>: <wordtree> * @@ -333,6 +331,12 @@ typedef int salfirst_T; typedef short salfirst_T; #endif +/* Values for SP_*ERROR are negative, positive values are used by + * read_cnt_string(). */ +#define SP_TRUNCERROR -1 /* spell file truncated error */ +#define SP_FORMERROR -2 /* format error in spell file */ +#define SP_ERROR -3 /* other error while reading spell file */ + /* * Structure used to store words and other info for one language, loaded from * a .spl file. @@ -367,8 +371,14 @@ struct slang_S char_u *sl_midword; /* MIDWORD string or NULL */ - int sl_compminlen; /* COMPOUNDMIN */ - char_u *sl_compflags; /* COMPOUNDFLAGS (NULL when no compounding) */ + int sl_compmax; /* COMPOUNDMAX (default: MAXWLEN) */ + int sl_compminlen; /* COMPOUNDMIN (default: MAXWLEN) */ + int sl_compsylmax; /* COMPOUNDSYLMAX (default: MAXWLEN) */ + regprog_T *sl_compprog; /* COMPOUNDFLAGS turned into a regexp progrm + * (NULL when no compounding) */ + char_u *sl_compstartflags; /* flags for first compound word */ + char_u *sl_syllable; /* SYLLABLE repeatable chars or NULL */ + garray_T sl_syl_items; /* syllable items */ int sl_prefixcnt; /* number of items in "sl_prefprog" */ regprog_T **sl_prefprog; /* table with regprogs for prefixes */ @@ -402,7 +412,6 @@ static slang_T *first_lang = NULL; #define SAL_F0LLOWUP 1 #define SAL_COLLAPSE 2 #define SAL_REM_ACCENTS 4 -#define SAL_SOFO 8 /* SOFOFROM and SOFOTO instead of SAL */ /* * Structure used in "b_langp", filled from 'spelllang'. @@ -417,6 +426,25 @@ typedef struct langp_S #define REGION_ALL 0xff /* word valid in all regions */ +#define VIMSPELLMAGIC "VIMspell" /* string at start of Vim spell file */ +#define VIMSPELLMAGICL 8 +#define VIMSPELLVERSION 50 + +/* Section IDs. Only renumber them when VIMSPELLVERSION changes! */ +#define SN_REGION 0 /* <regionname> section */ +#define SN_CHARFLAGS 1 /* charflags section */ +#define SN_MIDWORD 2 /* <midword> section */ +#define SN_PREFCOND 3 /* <prefcond> section */ +#define SN_REP 4 /* REP items section */ +#define SN_SAL 5 /* SAL items section */ +#define SN_SOFO 6 /* soundfolding section */ +#define SN_MAP 7 /* MAP items section */ +#define SN_COMPOUND 8 /* compound words section */ +#define SN_SYLLABLE 9 /* syllable section */ +#define SN_END 255 /* end of sections */ + +#define SNF_REQUIRED 1 /* <sectionflags>: required section */ + /* Result values. Lower number is accepted over higher one. */ #define SP_BANNED -1 #define SP_OK 0 @@ -424,9 +452,6 @@ typedef struct langp_S #define SP_LOCAL 2 #define SP_BAD 3 -#define VIMSPELLMAGIC "VIMspell10" /* string at start of Vim spell file */ -#define VIMSPELLMAGICL 10 - /* file used for "zG" and "zW" */ static char_u *int_wordlist = NULL; @@ -472,7 +497,7 @@ typedef struct suggest_S /* score for various changes */ #define SCORE_SPLIT 149 /* split bad word */ #define SCORE_ICASE 52 /* slightly different case */ -#define SCORE_REGION 70 /* word is for different region */ +#define SCORE_REGION 200 /* word is for different region */ #define SCORE_RARE 180 /* rare word */ #define SCORE_SWAP 90 /* swap two characters */ #define SCORE_SWAP3 110 /* swap two characters in three */ @@ -527,6 +552,8 @@ typedef struct matchinf_S /* for when checking a compound word */ int mi_compoff; /* start of following word offset */ + char_u mi_compflags[MAXWLEN]; /* flags for compound words used */ + int mi_complen; /* nr of compound words used */ /* others */ int mi_result; /* result so far: SP_BAD, SP_OK, etc. */ @@ -559,7 +586,7 @@ static int spell_iswordp_nmw __ARGS((char_u *p)); #ifdef FEAT_MBYTE static int spell_iswordp_w __ARGS((int *p, buf_T *buf)); #endif -static void write_spell_prefcond __ARGS((FILE *fd, garray_T *gap)); +static int write_spell_prefcond __ARGS((FILE *fd, garray_T *gap)); /* * For finding suggestions: At each node in the tree these states are tried: @@ -608,8 +635,9 @@ typedef struct trystate_S char_u ts_isdiff; /* DIFF_ values */ char_u ts_fcharstart; /* index in fword where badword char started */ #endif - char_u ts_save_prewordlen; /* saved "prewordlen" */ - char_u ts_save_splitoff; /* su_splitoff saved here */ + char_u ts_prewordlen; /* length of word in "preword[]" */ + char_u ts_splitoff; /* index in "tword" after last split */ + char_u ts_complen; /* nr of compound words used */ char_u ts_save_badflags; /* su_badflags saved here */ } trystate_T; @@ -634,7 +662,7 @@ static slang_T *slang_alloc __ARGS((char_u *lang)); static void slang_free __ARGS((slang_T *lp)); static void slang_clear __ARGS((slang_T *lp)); static void find_word __ARGS((matchinf_T *mip, int mode)); -static int can_compound __ARGS((slang_T *slang, int flags)); +static int can_compound __ARGS((slang_T *slang, char_u *word, char_u *flags)); static int valid_word_prefix __ARGS((int totprefcnt, int arridx, int flags, char_u *word, slang_T *slang, int cond_req)); static void find_prefix __ARGS((matchinf_T *mip)); static int fold_more __ARGS((matchinf_T *mip)); @@ -646,6 +674,16 @@ static void int_wordlist_spl __ARGS((char_u *fname)); static void spell_load_cb __ARGS((char_u *fname, void *cookie)); static slang_T *spell_load_file __ARGS((char_u *fname, char_u *lang, slang_T *old_lp, int silent)); static char_u *read_cnt_string __ARGS((FILE *fd, int cnt_bytes, int *lenp)); +static char_u *read_string __ARGS((FILE *fd, int cnt)); +static int read_region_section __ARGS((FILE *fd, slang_T *slang, int len)); +static int read_charflags_section __ARGS((FILE *fd)); +static int read_prefcond_section __ARGS((FILE *fd, slang_T *lp)); +static int read_rep_section __ARGS((FILE *fd, slang_T *slang)); +static int read_sal_section __ARGS((FILE *fd, slang_T *slang)); +static int read_sofo_section __ARGS((FILE *fd, slang_T *slang)); +static int read_compound __ARGS((FILE *fd, slang_T *slang, int len)); +static int init_syl_tab __ARGS((slang_T *slang)); +static int count_syllables __ARGS((slang_T *slang, char_u *word)); static int set_sofo __ARGS((slang_T *lp, char_u *from, char_u *to)); static void set_sal_first __ARGS((slang_T *lp)); #ifdef FEAT_MBYTE @@ -658,9 +696,8 @@ static int find_region __ARGS((char_u *rp, char_u *region)); static int captype __ARGS((char_u *word, char_u *end)); static int badword_captype __ARGS((char_u *word, char_u *end)); static void spell_reload_one __ARGS((char_u *fname, int added_word)); -static int set_spell_charflags __ARGS((char_u *flags, int cnt, char_u *upp)); +static void set_spell_charflags __ARGS((char_u *flags, int cnt, char_u *upp)); static int set_spell_chartab __ARGS((char_u *fol, char_u *low, char_u *upp)); -static void write_spell_chartab __ARGS((FILE *fd)); static int spell_casefold __ARGS((char_u *p, int len, char_u *buf, int buflen)); static int check_need_cap __ARGS((linenr_T lnum, colnr_T col)); static void spell_find_suggest __ARGS((char_u *badptr, suginfo_T *su, int maxcount, int banbadword, int need_cap)); @@ -777,11 +814,13 @@ spell_check(wp, ptr, attrp, capcol) be passed to functions quickly. */ int nrlen = 0; /* found a number first */ int c; + int wrongcaplen = 0; /* A word never starts at a space or a control character. Return quickly * then, skipping over the character. */ if (*ptr <= ' ') return 1; + vim_memset(&mi, 0, sizeof(matchinf_T)); /* A number is always OK. Also skip hexadecimal numbers 0xFF99 and * 0X99FF. But when a word character follows do check spelling to find @@ -818,10 +857,7 @@ spell_check(wp, ptr, attrp, capcol) /* Check word starting with capital letter. */ c = PTR2CHAR(ptr); if (!SPELL_ISUPPER(c)) - { - *attrp = highlight_attr[HLF_SPC]; - return (int)(mi.mi_fend - ptr); - } + wrongcaplen = (int)(mi.mi_fend - ptr); } } if (capcol != NULL) @@ -832,12 +868,10 @@ spell_check(wp, ptr, attrp, capcol) mi.mi_end = mi.mi_fend; /* Check caps type later. */ - mi.mi_capflags = 0; - mi.mi_cend = NULL; mi.mi_buf = wp->w_buffer; - /* Include one non-word character so that we can check for the - * word end. */ + /* case-fold the word with one non-word character, so that we can check + * for the word end. */ if (*mi.mi_fend != NUL) mb_ptr_adv(mi.mi_fend); @@ -897,6 +931,10 @@ spell_check(wp, ptr, attrp, capcol) #endif return 1; } + else if (mi.mi_end == ptr) + /* Always include at least one character. Required for when there + * is a mixup in "midword". */ + mb_ptr_adv(mi.mi_end); if (mi.mi_result == SP_BAD || mi.mi_result == SP_BANNED) *attrp = highlight_attr[HLF_SPB]; @@ -906,6 +944,13 @@ spell_check(wp, ptr, attrp, capcol) *attrp = highlight_attr[HLF_SPL]; } + if (wrongcaplen > 0 && (mi.mi_result == SP_OK || mi.mi_result == SP_RARE)) + { + /* Report SpellCap only when the word isn't badly spelled. */ + *attrp = highlight_attr[HLF_SPC]; + return wrongcaplen; + } + return (int)(mi.mi_end - ptr); } @@ -1085,7 +1130,7 @@ find_word(mip, mode) #endif if (spell_iswordp(ptr + wlen, mip->mi_buf)) { - if (slang->sl_compflags == NULL) + if (slang->sl_compprog == NULL) continue; /* next char is a word character */ word_ends = FALSE; } @@ -1157,16 +1202,45 @@ find_word(mip, mode) if (mode == FIND_COMPOUND || mode == FIND_KEEPCOMPOUND || !word_ends) { - /* Makes you wonder why someone puts a compound flag on a word + /* If there is no flag or the word is shorter than + * COMPOUNDMIN reject it quickly. + * Makes you wonder why someone puts a compound flag on a word * that's too short... Myspell compatibility requires this * anyway. */ - if (wlen < slang->sl_compminlen) + if (((unsigned)flags >> 24) == 0 || wlen < slang->sl_compminlen) + continue; + + /* Limit the number of compound words to COMPOUNDMAX. */ + if (!word_ends && mip->mi_complen + 2 > slang->sl_compmax) continue; - /* The word doesn't end or it comes after another: it must - * have a compound flag. */ - if (!can_compound(slang, flags)) + /* At start of word quickly check if compounding is possible + * with this flag. */ + if (mip->mi_complen == 0 + && vim_strchr(slang->sl_compstartflags, + ((unsigned)flags >> 24)) == NULL) continue; + + /* If the word ends the sequence of compound flags of the + * words must match with one of the COMPOUNDFLAGS items and + * the number of syllables must not be too large. */ + mip->mi_compflags[mip->mi_complen] = ((unsigned)flags >> 24); + mip->mi_compflags[mip->mi_complen + 1] = NUL; + if (word_ends) + { + char_u fword[MAXWLEN]; + + if (slang->sl_compsylmax < MAXWLEN) + { + /* "fword" is only needed for checking syllables. */ + if (ptr == mip->mi_word) + (void)spell_casefold(ptr, wlen, fword, MAXWLEN); + else + vim_strncpy(fword, ptr, endlen[endidxcnt]); + } + if (!can_compound(slang, fword, mip->mi_compflags)) + continue; + } } if (!word_ends) @@ -1193,13 +1267,17 @@ find_word(mip, mode) } } #endif + ++mip->mi_complen; find_word(mip, FIND_COMPOUND); + --mip->mi_complen; if (mip->mi_result == SP_OK) break; /* Find following word in keep-case tree. */ mip->mi_compoff = wlen; + ++mip->mi_complen; find_word(mip, FIND_KEEPCOMPOUND); + --mip->mi_complen; if (mip->mi_result == SP_OK) break; continue; @@ -1239,16 +1317,29 @@ find_word(mip, mode) } /* - * Return TRUE if "flags" has a valid compound flag. - * TODO: check flags in a more advanced way. + * Return TRUE if "flags" is a valid sequence of compound flags and + * "word[len]" does not have too many syllables. */ static int -can_compound(slang, flags) +can_compound(slang, word, flags) slang_T *slang; - int flags; + char_u *word; + char_u *flags; { - return slang->sl_compflags != NULL - && *slang->sl_compflags == ((unsigned)flags >> 24); + regmatch_T regmatch; + + if (slang->sl_compprog == NULL) + return FALSE; + regmatch.regprog = slang->sl_compprog; + regmatch.rm_ic = FALSE; + if (!vim_regexec(®match, flags, 0)) + return FALSE; + + /* Count the number of syllables. This may be slow, do it last. */ + if (slang->sl_compsylmax < MAXWLEN + && count_syllables(slang, word) > slang->sl_compsylmax) + return FALSE; + return TRUE; } /* @@ -1480,6 +1571,8 @@ no_spell_checking() /* * Move to next spell error. * "curline" is TRUE for "z?": find word under/after cursor in the same line. + * For Insert mode completion "dir" is BACKWARD and "curline" is TRUE: move + * to after badly spelled word before the cursor. * Return OK if found, FAIL otherwise. */ int @@ -1599,6 +1692,10 @@ spell_move_to(dir, allwords, curline) vim_free(buf); return OK; } + else if (curline) + /* Insert mode completion: put cursor after + * the bad word. */ + found_pos.col += len; } } } @@ -1609,19 +1706,20 @@ spell_move_to(dir, allwords, curline) capcol -= len; } + if (dir == BACKWARD && found_pos.lnum != 0) + { + /* Use the last match in the line. */ + curwin->w_cursor = found_pos; + vim_free(buf); + return OK; + } + if (curline) break; /* only check cursor line */ /* Advance to next line. */ if (dir == BACKWARD) { - if (found_pos.lnum != 0) - { - /* Use the last match in the line. */ - curwin->w_cursor = found_pos; - vim_free(buf); - return OK; - } if (lnum == 1) break; --lnum; @@ -1715,8 +1813,8 @@ spell_load_lang(lang) } if (r == FAIL) - smsg((char_u *)_("Warning: Cannot find word list \"%s\""), - fname_enc + 6); + smsg((char_u *)_("Warning: Cannot find word list \"%s.%s.spl\" or \"%s.ascii.spl\""), + lang, spell_enc(), lang); else if (*langcp != NUL) { /* Load all the additions. */ @@ -1767,6 +1865,9 @@ slang_alloc(lang) { lp->sl_name = vim_strsave(lang); ga_init2(&lp->sl_rep, sizeof(fromto_T), 10); + lp->sl_compmax = MAXWLEN; + lp->sl_compminlen = MAXWLEN; + lp->sl_compsylmax = MAXWLEN; } return lp; } @@ -1853,8 +1954,14 @@ slang_clear(lp) vim_free(lp->sl_midword); lp->sl_midword = NULL; - vim_free(lp->sl_compflags); - lp->sl_compflags = NULL; + vim_free(lp->sl_compprog); + vim_free(lp->sl_compstartflags); + lp->sl_compprog = NULL; + lp->sl_compstartflags = NULL; + + vim_free(lp->sl_syllable); + lp->sl_syllable = NULL; + ga_clear(&lp->sl_syl_items); #ifdef FEAT_MBYTE { @@ -1870,6 +1977,11 @@ slang_clear(lp) } hash_clear(&lp->sl_map_hash); #endif + + lp->sl_compmax = MAXWLEN; + lp->sl_compminlen = MAXWLEN; + lp->sl_compsylmax = MAXWLEN; + lp->sl_regions[0] = NUL; } /* @@ -1902,7 +2014,7 @@ spell_load_file(fname, lang, old_lp, silent) int silent; /* no error if file doesn't exist */ { FILE *fd; - char_u buf[MAXWLEN + 1]; + char_u buf[VIMSPELLMAGICL]; char_u *p; char_u *bp; idx_T *ip; @@ -1912,15 +2024,10 @@ spell_load_file(fname, lang, old_lp, silent) int round; char_u *save_sourcing_name = sourcing_name; linenr_T save_sourcing_lnum = sourcing_lnum; - int cnt, ccnt; - char_u *fol; slang_T *lp = NULL; - garray_T *gap; - fromto_T *ftp; - salitem_T *smp; - short *first; idx_T idx; int c = 0; + int res; fd = mch_fopen((char *)fname, "r"); if (fd == NULL) @@ -1964,332 +2071,122 @@ spell_load_file(fname, lang, old_lp, silent) sourcing_lnum = 0; /* <HEADER>: <fileID> - * <regioncnt> <regionname> ... - * <charflagslen> <charflags> - * <fcharslen> <fchars> - * <midwordlen> <midword> - * <compoundlen> <compoundtype> <compoundinfo> - * <prefcondcnt> <prefcond> ... */ for (i = 0; i < VIMSPELLMAGICL; ++i) buf[i] = getc(fd); /* <fileID> */ if (STRNCMP(buf, VIMSPELLMAGIC, VIMSPELLMAGICL) != 0) { - EMSG(_("E757: Wrong file ID in spell file")); + EMSG(_("E757: This does not look like a spell file")); goto endFAIL; } - - cnt = getc(fd); /* <regioncnt> */ - if (cnt < 0) + c = getc(fd); /* <versionnr> */ + if (c < VIMSPELLVERSION) { -truncerr: - EMSG(_(e_spell_trunc)); + EMSG(_("E771: Old spell file, needs to be updated")); goto endFAIL; } - if (cnt > 8) + else if (c > VIMSPELLVERSION) { -formerr: - EMSG(_(e_format)); + EMSG(_("E772: Spell file is for newer version of Vim")); goto endFAIL; } - for (i = 0; i < cnt; ++i) - { - lp->sl_regions[i * 2] = getc(fd); /* <regionname> */ - lp->sl_regions[i * 2 + 1] = getc(fd); - } - lp->sl_regions[cnt * 2] = NUL; - /* <charflagslen> <charflags> */ - p = read_cnt_string(fd, 1, &cnt); - if (cnt < 0) - goto endFAIL; - /* <fcharslen> <fchars> */ - fol = read_cnt_string(fd, 2, &ccnt); - if (ccnt < 0) + /* + * <SECTIONS>: <section> ... <sectionend> + * <section>: <sectionID> <sectionflags> <sectionlen> (section contents) + */ + for (;;) { - vim_free(p); - goto endFAIL; - } - - /* Set the word-char flags and fill SPELL_ISUPPER() table. */ - if (p != NULL && fol != NULL) - i = set_spell_charflags(p, cnt, fol); - - vim_free(p); - vim_free(fol); - - /* When <charflagslen> is zero then <fcharlen> must also be zero. */ - if ((p == NULL) != (fol == NULL)) - goto formerr; - - /* <midwordlen> <midword> */ - lp->sl_midword = read_cnt_string(fd, 2, &cnt); - if (cnt < 0) - goto endFAIL; + n = getc(fd); /* <sectionID> or <sectionend> */ + if (n == SN_END) + break; + c = getc(fd); /* <sectionflags> */ + len = (getc(fd) << 24) + (getc(fd) << 16) + (getc(fd) << 8) + getc(fd); + /* <sectionlen> */ + if (len < 0) + goto truncerr; - /* <compoundlen> <compoundtype> <compoundinfo> */ - cnt = (getc(fd) << 8) + getc(fd); /* <compoundlen> */ - if (cnt < 0) - goto endFAIL; - if (cnt > 0) - { - --cnt; - c = getc(fd); /* <compoundtype> */ - if (c != 1) + res = 0; + switch (n) { - /* Unknown kind of compound words, skip the info. */ - while (cnt-- > 0) - getc(fd); - } - else if (cnt < 2) - goto formerr; - else - { - --cnt; - c = getc(fd); /* <comp1minlen> */ - if (c < 1 || c > 50) - c = 3; - lp->sl_compminlen = c; + case SN_REGION: + res = read_region_section(fd, lp, len); + break; - p = alloc(cnt + 1); - if (p == NULL) - goto endFAIL; - lp->sl_compflags = p; - while (cnt-- > 0) - *p++ = getc(fd); /* <comp1flags> */ - *p = NUL; - } - } + case SN_CHARFLAGS: + res = read_charflags_section(fd); + break; + case SN_MIDWORD: + lp->sl_midword = read_string(fd, len); /* <midword> */ + if (lp->sl_midword == NULL) + goto endFAIL; + break; - /* <prefcondcnt> <prefcond> ... */ - cnt = (getc(fd) << 8) + getc(fd); /* <prefcondcnt> */ - if (cnt > 0) - { - lp->sl_prefprog = (regprog_T **)alloc_clear( - (unsigned)sizeof(regprog_T *) * cnt); - if (lp->sl_prefprog == NULL) - goto endFAIL; - lp->sl_prefixcnt = cnt; + case SN_PREFCOND: + res = read_prefcond_section(fd, lp); + break; - for (i = 0; i < cnt; ++i) - { - /* <prefcond> : <condlen> <condstr> */ - n = getc(fd); /* <condlen> */ - if (n < 0 || n >= MAXWLEN) - goto formerr; - /* When <condlen |