Mercurial > vim

changeset 5477:853a76c7fded v7.4.088

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
updated for version 7.4.088 Problem: When spell checking is enabled Asian characters are always marked as error. Solution: When 'spelllang' contains "cjk" do not mark Asian characters as error. (Ken Takata)
author Bram Moolenaar <bram@vim.org>
date Tue, 12 Nov 2013 04:44:01 +0100
parents 63314ae20838
children 8598fbd3bc5e
files runtime/doc/options.txt runtime/doc/spell.txt src/mbyte.c src/option.c src/spell.c src/structs.h src/version.c
diffstat 7 files changed, 76 insertions(+), 37 deletions(-) [+]
line wrap: on
line diff
--- a/runtime/doc/options.txt
+++ b/runtime/doc/options.txt
@@ -6555,6 +6555,9 @@ A jump table for the options with a shor
 	region by listing them: "en_us,en_ca" supports both US and Canadian
 	English, but not words specific for Australia, New Zealand or Great
 	Britain.
+	If the name "cjk" is included East Asian characters are excluded from
+	spell checking.  This is useful when editing text that also has Asian
+	words.
 							*E757*
 	As a special case the name of a .spl file can be given as-is.  The
 	first "_xx" in the name is removed and used as the region name
--- a/runtime/doc/spell.txt
+++ b/runtime/doc/spell.txt
@@ -269,6 +269,13 @@ In a table:
 	latin1		yi		transliterated Yiddish
 	utf-8		yi-tr		transliterated Yiddish
 
+							*spell-cjk*
+Chinese, Japanese and other East Asian characters are normally marked as
+errors, because spell checking of these characters is not supported. If
+'spelllang' includes "cjk", these characters are not marked as errors.  This
+is useful when editing text with spell checking while some Asian words are
+present.
+
 
 SPELL FILES						*spell-load*
 
--- a/src/mbyte.c
+++ b/src/mbyte.c
@@ -947,8 +947,8 @@ dbcs_class(lead, trail)
 		{
 		    case 0x2121: /* ZENKAKU space */
 			return 0;
-		    case 0x2122: /* KU-TEN (Japanese comma) */
-		    case 0x2123: /* TOU-TEN (Japanese period) */
+		    case 0x2122: /* TOU-TEN (Japanese comma) */
+		    case 0x2123: /* KU-TEN (Japanese period) */
 		    case 0x2124: /* ZENKAKU comma */
 		    case 0x2125: /* ZENKAKU period */
 			return 1;
@@ -2477,9 +2477,9 @@ utf_class(c)
     /* sorted list of non-overlapping intervals */
     static struct clinterval
     {
-	unsigned short first;
-	unsigned short last;
-	unsigned short class;
+	unsigned int first;
+	unsigned int last;
+	unsigned int class;
     } classes[] =
     {
 	{0x037e, 0x037e, 1},		/* Greek question mark */
@@ -2544,6 +2544,10 @@ utf_class(c)
 	{0xff1a, 0xff20, 1},		/* half/fullwidth ASCII */
 	{0xff3b, 0xff40, 1},		/* half/fullwidth ASCII */
 	{0xff5b, 0xff65, 1},		/* half/fullwidth ASCII */
+	{0x20000, 0x2a6df, 0x4e00},	/* CJK Ideographs */
+	{0x2a700, 0x2b73f, 0x4e00},	/* CJK Ideographs */
+	{0x2b740, 0x2b81f, 0x4e00},	/* CJK Ideographs */
+	{0x2f800, 0x2fa1f, 0x4e00},	/* CJK Ideographs */
     };
     int bot = 0;
     int top = sizeof(classes) / sizeof(struct clinterval) - 1;
@@ -2563,9 +2567,9 @@ utf_class(c)
     while (top >= bot)
     {
 	mid = (bot + top) / 2;
-	if (classes[mid].last < c)
+	if (classes[mid].last < (unsigned int)c)
 	    bot = mid + 1;
-	else if (classes[mid].first > c)
+	else if (classes[mid].first > (unsigned int)c)
 	    top = mid - 1;
 	else
 	    return (int)classes[mid].class;
--- a/src/option.c
+++ b/src/option.c
@@ -7122,6 +7122,11 @@ did_set_string_option(opt_idx, varp, new
 	if (varp == &(curwin->w_s->b_p_spl))
 	{
 	    char_u	fname[200];
+	    char_u	*q = curwin->w_s->b_p_spl;
+
+	    /* Skip the first name if it is "cjk". */
+	    if (STRNCMP(q, "cjk,", 4) == 0)
+		q += 4;
 
 	    /*
 	     * Source the spell/LANG.vim in 'runtimepath'.
@@ -7129,11 +7134,10 @@ did_set_string_option(opt_idx, varp, new
 	     * Use the first name in 'spelllang' up to '_region' or
 	     * '.encoding'.
 	     */
-	    for (p = curwin->w_s->b_p_spl; *p != NUL; ++p)
+	    for (p = q; *p != NUL; ++p)
 		if (vim_strchr((char_u *)"_.,", *p) != NULL)
 		    break;
-	    vim_snprintf((char *)fname, 200, "spell/%.*s.vim",
-				 (int)(p - curwin->w_s->b_p_spl), curwin->w_s->b_p_spl);
+	    vim_snprintf((char *)fname, 200, "spell/%.*s.vim", (int)(p - q), q);
 	    source_runtime(fname, TRUE);
 	}
 #endif
--- a/src/spell.c
+++ b/src/spell.c
@@ -754,9 +754,9 @@ static int	    did_set_spelltab;
 static void clear_spell_chartab __ARGS((spelltab_T *sp));
 static int set_spell_finish __ARGS((spelltab_T	*new_st));
 static int spell_iswordp __ARGS((char_u *p, win_T *wp));
-static int spell_iswordp_nmw __ARGS((char_u *p));
-#ifdef FEAT_MBYTE
-static int spell_mb_isword_class __ARGS((int cl));
+static int spell_iswordp_nmw __ARGS((char_u *p, win_T *wp));
+#ifdef FEAT_MBYTE
+static int spell_mb_isword_class __ARGS((int cl, win_T *wp));
 static int spell_iswordp_w __ARGS((int *p, win_T *wp));
 #endif
 static int write_spell_prefcond __ARGS((FILE *fd, garray_T *gap));
@@ -1149,7 +1149,7 @@ spell_check(wp, ptr, attrp, capcol, doco
 
 	/* When we are at a non-word character there is no error, just
 	 * skip over the character (try looking for a word after it). */
-	else if (!spell_iswordp_nmw(ptr))
+	else if (!spell_iswordp_nmw(ptr, wp))
 	{
 	    if (capcol != NULL && wp->w_s->b_cap_prog != NULL)
 	    {
@@ -1561,7 +1561,7 @@ find_word(mip, mode)
 			 * accept a no-caps word, even when the dictionary
 			 * word specifies ONECAP. */
 			mb_ptr_back(mip->mi_word, p);
-			if (spell_iswordp_nmw(p)
+			if (spell_iswordp_nmw(p, mip->mi_win)
 				? capflags == WF_ONECAP
 				: (flags & WF_ONECAP) != 0
 						     && capflags != WF_ONECAP)
@@ -4234,7 +4234,9 @@ did_set_spelllang(wp)
     if (spl_copy == NULL)
 	goto theend;
 
-    /* loop over comma separated language names. */
+    wp->w_s->b_cjk = 0;
+
+    /* Loop over comma separated language names. */
     for (splp = spl_copy; *splp != NUL; )
     {
 	/* Get one language name. */
@@ -4242,6 +4244,12 @@ did_set_spelllang(wp)
 	region = NULL;
 	len = (int)STRLEN(lang);
 
+	if (STRCMP(lang, "cjk") == 0)
+	{
+	    wp->w_s->b_cjk = 1;
+	    continue;
+	}
+
 	/* If the name ends in ".spl" use it as the name of the spell file.
 	 * If there is a region name let "region" point to it and remove it
 	 * from the name. */
@@ -4601,7 +4609,7 @@ captype(word, end)
     int		past_second = FALSE;	/* past second word char */
 
     /* find first letter */
-    for (p = word; !spell_iswordp_nmw(p); mb_ptr_adv(p))
+    for (p = word; !spell_iswordp_nmw(p, curwin); mb_ptr_adv(p))
 	if (end == NULL ? *p == NUL : p >= end)
 	    return 0;	    /* only non-word characters, illegal word */
 #ifdef FEAT_MBYTE
@@ -4617,7 +4625,7 @@ captype(word, end)
      * But a word with an upper char only at start is a ONECAP.
      */
     for ( ; end == NULL ? *p != NUL : p < end; mb_ptr_adv(p))
-	if (spell_iswordp_nmw(p))
+	if (spell_iswordp_nmw(p, curwin))
 	{
 	    c = PTR2CHAR(p);
 	    if (!SPELL_ISUPPER(c))
@@ -9907,7 +9915,7 @@ spell_iswordp(p, wp)
 
 	c = mb_ptr2char(s);
 	if (c > 255)
-	    return spell_mb_isword_class(mb_get_class(s));
+	    return spell_mb_isword_class(mb_get_class(s), wp);
 	return spelltab.st_isw[c];
     }
 #endif
@@ -9920,8 +9928,9 @@ spell_iswordp(p, wp)
  * Unlike spell_iswordp() this doesn't check for "midword" characters.
  */
     static int
-spell_iswordp_nmw(p)
+spell_iswordp_nmw(p, wp)
     char_u	*p;
+    win_T	*wp;
 {
 #ifdef FEAT_MBYTE
     int		c;
@@ -9930,7 +9939,7 @@ spell_iswordp_nmw(p)
     {
 	c = mb_ptr2char(p);
 	if (c > 255)
-	    return spell_mb_isword_class(mb_get_class(p));
+	    return spell_mb_isword_class(mb_get_class(p), wp);
 	return spelltab.st_isw[c];
     }
 #endif
@@ -9942,11 +9951,16 @@ spell_iswordp_nmw(p)
  * Return TRUE if word class indicates a word character.
  * Only for characters above 255.
  * Unicode subscript and superscript are not considered word characters.
- */
-    static int
-spell_mb_isword_class(cl)
-    int cl;
-{
+ * See also dbcs_class() and utf_class() in mbyte.c.
+ */
+    static int
+spell_mb_isword_class(cl, wp)
+    int		cl;
+    win_T	*wp;
+{
+    if (wp->w_s->b_cjk)
+	/* East Asian characters are not considered word characters. */
+	return cl == 2 || cl == 0x2800;
     return cl >= 2 && cl != 0x2070 && cl != 0x2080;
 }
 
@@ -9971,9 +9985,10 @@ spell_iswordp_w(p, wp)
     if (*s > 255)
     {
 	if (enc_utf8)
-	    return spell_mb_isword_class(utf_class(*s));
+	    return spell_mb_isword_class(utf_class(*s), wp);
 	if (enc_dbcs)
-	    return dbcs_class((unsigned)*s >> 8, *s & 0xff) >= 2;
+	    return spell_mb_isword_class(
+				dbcs_class((unsigned)*s >> 8, *s & 0xff), wp);
 	return 0;
     }
     return spelltab.st_isw[*s];
@@ -10193,13 +10208,13 @@ spell_suggest(count)
 	line = ml_get_curline();
 	p = line + curwin->w_cursor.col;
 	/* Backup to before start of word. */
-	while (p > line && spell_iswordp_nmw(p))
+	while (p > line && spell_iswordp_nmw(p, curwin))
 	    mb_ptr_back(line, p);
 	/* Forward to start of word. */
-	while (*p != NUL && !spell_iswordp_nmw(p))
+	while (*p != NUL && !spell_iswordp_nmw(p, curwin))
 	    mb_ptr_adv(p);
 
-	if (!spell_iswordp_nmw(p))		/* No word found. */
+	if (!spell_iswordp_nmw(p, curwin))		/* No word found. */
 	{
 	    beep_flush();
 	    return;
@@ -10436,7 +10451,7 @@ check_need_cap(lnum, col)
 	for (;;)
 	{
 	    mb_ptr_back(line, p);
-	    if (p == line || spell_iswordp_nmw(p))
+	    if (p == line || spell_iswordp_nmw(p, curwin))
 		break;
 	    if (vim_regexec(&regmatch, p, 0)
 					 && regmatch.endp[0] == line + endcol)
@@ -11645,7 +11660,7 @@ suggest_trie_walk(su, lp, fword, soundfo
 
 		/* When appending a compound word after a word character don't
 		 * use Onecap. */
-		if (p != NULL && spell_iswordp_nmw(p))
+		if (p != NULL && spell_iswordp_nmw(p, curwin))
 		    c &= ~WF_ONECAP;
 		make_case_word(tword + sp->ts_splitoff,
 					      preword + sp->ts_prewordlen, c);
@@ -11895,7 +11910,8 @@ suggest_trie_walk(su, lp, fword, soundfo
 			 * character when the word ends.  But only when the
 			 * good word can end. */
 			if (((!try_compound && !spell_iswordp_nmw(fword
-							       + sp->ts_fidx))
+							       + sp->ts_fidx,
+							       curwin))
 				    || fword_ends)
 				&& fword[sp->ts_fidx] != NUL
 				&& goodword_ends)
@@ -14226,7 +14242,7 @@ spell_soundfold_sal(slang, inword, res)
 	    }
 	    else
 	    {
-		if (spell_iswordp_nmw(s))
+		if (spell_iswordp_nmw(s, curwin))
 		    *t++ = *s;
 		++s;
 	    }
@@ -14521,7 +14537,7 @@ spell_soundfold_wsal(slang, inword, res)
 	    else
 	    {
 		did_white = FALSE;
-		if (!spell_iswordp_nmw(t))
+		if (!spell_iswordp_nmw(t, curwin))
 		    continue;
 	    }
 	}
@@ -16045,7 +16061,7 @@ spell_word_start(startcol)
     for (p = line + startcol; p > line; )
     {
 	mb_ptr_back(line, p);
-	if (spell_iswordp_nmw(p))
+	if (spell_iswordp_nmw(p, curwin))
 	    break;
     }
 
--- a/src/structs.h
+++ b/src/structs.h
@@ -1310,6 +1310,9 @@ typedef struct {
     regprog_T	*b_cap_prog;	/* program for 'spellcapcheck' */
     char_u	*b_p_spf;	/* 'spellfile' */
     char_u	*b_p_spl;	/* 'spelllang' */
+# ifdef FEAT_MBYTE
+    int		b_cjk;		/* all CJK letters as OK */
+# endif
 #endif
 #if !defined(FEAT_SYN_HL) && !defined(FEAT_SPELL)
     int		dummy;
--- a/src/version.c
+++ b/src/version.c
@@ -739,6 +739,8 @@ static char *(features[]) =
 static int included_patches[] =
 {   /* Add new patch number below this line */
 /**/
+    88,
+/**/
     87,
 /**/
     86,