Mercurial > vim

changeset 2338:da6ec32d8d8f vim73

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Added strwidth() and strchars() functions.
author Bram Moolenaar <bram@vim.org>
date Sun, 18 Jul 2010 15:31:08 +0200
parents a0f87fc19d1d
children 01e4b4d37842
files runtime/doc/eval.txt runtime/doc/todo.txt src/eval.c src/gui.c src/gui_mac.c src/gui_w16.c src/gui_w32.c src/mbyte.c src/proto/mbyte.pro src/screen.c
diffstat 10 files changed, 99 insertions(+), 42 deletions(-) [+]
line wrap: on
line diff
--- a/runtime/doc/eval.txt
+++ b/runtime/doc/eval.txt
@@ -1909,6 +1909,7 @@ split( {expr} [, {pat} [, {keepempty}]])
 sqrt( {expr}			Float	squar root of {expr}
 str2float( {expr})		Float	convert String to Float
 str2nr( {expr} [, {base}])	Number	convert String to Number
+strchars( {expr})		Number	character length of the String {expr}
 strftime( {format}[, {time}])	String	time in specified format
 stridx( {haystack}, {needle}[, {start}])
 				Number	index of {needle} in {haystack}
@@ -1919,6 +1920,7 @@ strpart( {src}, {start}[, {len}])
 strridx( {haystack}, {needle} [, {start}])
 				Number	last index of {needle} in {haystack}
 strtrans( {expr})		String	translate string to make it printable
+strwidth( {expr})		Number	display cell length of the String {expr}
 submatch( {nr})			String	specific match in ":substitute"
 substitute( {expr}, {pat}, {sub}, {flags})
 				String	all {pat} in {expr} replaced with {sub}
@@ -5362,6 +5364,12 @@ str2nr( {expr} [, {base}])				*str2nr()*
 		Text after the number is silently ignored.
 
 
+strchars({expr})					*strchars()*
+		The result is a Number, which is the number of characters
+		String {expr} occupies.  Composing characters are counted
+		separately.
+		Also see |strlen()| and |strwidth()|.
+
 strftime({format} [, {time}])				*strftime()*
 		The result is a String, which is a formatted date and time, as
 		specified by the {format} string.  The given {time} is used,
@@ -5424,7 +5432,7 @@ strlen({expr})	The result is a Number, w
 <
 		If the argument is a Number it is first converted to a String.
 		For other types an error is given.
-		Also see |len()|.
+		Also see |len()|, |strchars()| and |strwidth()|.
 
 strpart({src}, {start}[, {len}])			*strpart()*
 		The result is a String, which is part of {src}, starting from
@@ -5467,6 +5475,14 @@ strtrans({expr})					*strtrans()*
 <		This displays a newline in register a as "^@" instead of
 		starting a new line.
 
+strwidth({expr})					*strwidth()*
+		The result is a Number, which is the number of display cells
+		String {expr} occupies.  A Tab character is counted as one
+		cell (on the screen it depends on the position).
+		When {expr} contains characters with East Asian Width Class
+		Ambiguous, this function's return value depends on 'ambiwidth'.
+		Also see |strlen()| and |strchars()|.
+
 submatch({nr})						*submatch()*
 		Only for an expression in a |:substitute| command.  Returns
 		the {nr}'th submatch of the matched text.  When {nr} is 0
--- a/runtime/doc/todo.txt
+++ b/runtime/doc/todo.txt
@@ -1193,8 +1193,8 @@ 8   Add functions:
 			Search in 'runtimepath'?
 			More docs needed about how to use this.
 			How to get the messages into the .po files?
-    charlen()		Like strlen() but counting characters instead of
-			bytes.
+    strchars()		Like strlen() and strwidth() but counting characters
+			instead of bytes.
     confirm()		add "flags" argument, with 'v' for vertical
 			    layout and 'c' for console dialog. (Haegg)
 			    Flemming Madsen has a patch for the 'c' flag
--- a/src/eval.c
+++ b/src/eval.c
@@ -700,6 +700,7 @@ static void f_sqrt __ARGS((typval_T *arg
 static void f_str2float __ARGS((typval_T *argvars, typval_T *rettv));
 #endif
 static void f_str2nr __ARGS((typval_T *argvars, typval_T *rettv));
+static void f_strchars __ARGS((typval_T *argvars, typval_T *rettv));
 #ifdef HAVE_STRFTIME
 static void f_strftime __ARGS((typval_T *argvars, typval_T *rettv));
 #endif
@@ -709,6 +710,7 @@ static void f_strlen __ARGS((typval_T *a
 static void f_strpart __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_strridx __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_strtrans __ARGS((typval_T *argvars, typval_T *rettv));
+static void f_strwidth __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_submatch __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_substitute __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_synID __ARGS((typval_T *argvars, typval_T *rettv));
@@ -7856,6 +7858,7 @@ static struct fst
     {"str2float",	1, 1, f_str2float},
 #endif
     {"str2nr",		1, 2, f_str2nr},
+    {"strchars",	1, 1, f_strchars},
 #ifdef HAVE_STRFTIME
     {"strftime",	1, 2, f_strftime},
 #endif
@@ -7865,6 +7868,7 @@ static struct fst
     {"strpart",		2, 3, f_strpart},
     {"strridx",		2, 3, f_strridx},
     {"strtrans",	1, 1, f_strtrans},
+    {"strwidth",	1, 1, f_strwidth},
     {"submatch",	1, 1, f_submatch},
     {"substitute",	4, 4, f_substitute},
     {"synID",		3, 3, f_synID},
@@ -16778,6 +16782,48 @@ f_strlen(argvars, rettv)
 }
 
 /*
+ * "strchars()" function
+ */
+    static void
+f_strchars(argvars, rettv)
+    typval_T	*argvars;
+    typval_T	*rettv;
+{
+    char_u		*s = get_tv_string(&argvars[0]);
+#ifdef FEAT_MBYTE
+    varnumber_T		len = 0;
+
+    while (*s != NUL)
+    {
+	mb_cptr2char_adv(&s);
+	++len;
+    }
+    rettv->vval.v_number = len;
+#else
+    rettv->vval.v_number = (varnumber_T)(STRLEN(s));
+#endif
+}
+
+/*
+ * "strwidth()" function
+ */
+    static void
+f_strwidth(argvars, rettv)
+    typval_T	*argvars;
+    typval_T	*rettv;
+{
+    char_u	*s = get_tv_string(&argvars[0]);
+
+    rettv->vval.v_number = (varnumber_T)(
+#ifdef FEAT_MBYTE
+	    mb_string2cells(s, -1)
+#else
+	    STRLEN(s)
+#endif
+	    );
+}
+
+/*
  * "strpart()" function
  */
     static void
--- a/src/gui.c
+++ b/src/gui.c
@@ -2329,14 +2329,9 @@ gui_outstr_nowrap(s, len, flags, fg, bg,
 # ifdef FEAT_MBYTE
 	if (enc_dbcs == DBCS_JPNU)
 	{
-	    int		clen = 0;
-	    int		i;
-
 	    /* Get the length in display cells, this can be different from the
 	     * number of bytes for "euc-jp". */
-	    for (i = 0; i < len; i += (*mb_ptr2len)(s + i))
-		clen += (*mb_ptr2cells)(s + i);
-	    len = clen;
+	    len = mb_string2cells(s, len);
 	}
 # endif
     }
--- a/src/gui_mac.c
+++ b/src/gui_mac.c
@@ -3983,13 +3983,8 @@ draw_string_QD(int row, int col, char_u 
 	/* Multibyte computation taken from gui_w32.c */
 	if (has_mbyte)
 	{
-	    int cell_len = 0;
-	    int n;
-
 	    /* Compute the length in display cells. */
-	    for (n = 0; n < len; n += MB_BYTE2LEN(s[n]))
-		cell_len += (*mb_ptr2cells)(s + n);
-	    rc.right = FILL_X(col + cell_len);
+	    rc.right = FILL_X(col + mb_string2cells(s, len));
 	}
 	else
 #endif
@@ -4087,13 +4082,8 @@ draw_string_ATSUI(int row, int col, char
 	/* Multibyte computation taken from gui_w32.c */
 	if (has_mbyte)
 	{
-	    int cell_len = 0;
-	    int n;
-
 	    /* Compute the length in display cells. */
-	    for (n = 0; n < len; n += MB_BYTE2LEN(s[n]))
-		cell_len += (*mb_ptr2cells)(s + n);
-	    rc.right = FILL_X(col + cell_len);
+	    rc.right = FILL_X(col + mb_string2cells(s, len));
 	}
 	else
 	    rc.right = FILL_X(col + len) + (col + len == Columns);
--- a/src/gui_w16.c
+++ b/src/gui_w16.c
@@ -664,12 +664,8 @@ gui_mch_draw_string(
 #ifdef FEAT_MBYTE
 	if (has_mbyte)
 	{
-	    int cell_len = 0;
-
 	    /* Compute the length in display cells. */
-	    for (n = 0; n < len; n += MB_BYTE2LEN(text[n]))
-		cell_len += (*mb_ptr2cells)(text + n);
-	    rc.right = FILL_X(col + cell_len);
+	    rc.right = FILL_X(col + mb_string2cells(text, len));
 	}
 	else
 #endif
--- a/src/gui_w32.c
+++ b/src/gui_w32.c
@@ -2261,12 +2261,8 @@ gui_mch_draw_string(
 #ifdef FEAT_MBYTE
 	if (has_mbyte)
 	{
-	    int cell_len = 0;
-
 	    /* Compute the length in display cells. */
-	    for (n = 0; n < len; n += MB_BYTE2LEN(text[n]))
-		cell_len += (*mb_ptr2cells)(text + n);
-	    rc.right = FILL_X(col + cell_len);
+	    rc.right = FILL_X(col + mb_string2cells(text, len));
 	}
 	else
 #endif
--- a/src/mbyte.c
+++ b/src/mbyte.c
@@ -1579,6 +1579,23 @@ dbcs_char2cells(c)
 }
 
 /*
+ * Return the number of cells occupied by string "p".
+ * Stop at a NUL character.  When "len" >= 0 stop at character "p[len]".
+ */
+    int
+mb_string2cells(p, len)
+    char_u  *p;
+    int	    len;
+{
+    int i;
+    int clen = 0;
+
+    for (i = 0; (len < 0 || i < len) && p[i] != NUL; i += (*mb_ptr2len)(p + i))
+	clen += (*mb_ptr2cells)(p + i);
+    return clen;
+}
+
+/*
  * mb_off2cells() function pointer.
  * Return number of display cells for char at ScreenLines[off].
  * We make sure that the offset used is less than "max_off".
@@ -4364,12 +4381,12 @@ im_commit_cb(GtkIMContext *context UNUSE
 	     const gchar *str,
 	     gpointer data UNUSED)
 {
-    int	slen = (int)STRLEN(str);
-    int	add_to_input = TRUE;
-    int	clen;
-    int	len = slen;
-    int	commit_with_preedit = TRUE;
-    char_u	*im_str, *p;
+    int		slen = (int)STRLEN(str);
+    int		add_to_input = TRUE;
+    int		clen;
+    int		len = slen;
+    int		commit_with_preedit = TRUE;
+    char_u	*im_str;
 
 #ifdef XIM_DEBUG
     xim_log("im_commit_cb(): %s\n", str);
@@ -4402,9 +4419,9 @@ im_commit_cb(GtkIMContext *context UNUSE
     }
     else
 	im_str = (char_u *)str;
-    clen = 0;
-    for (p = im_str; p < im_str + len; p += (*mb_ptr2len)(p))
-	clen += (*mb_ptr2cells)(p);
+
+    clen = mb_string2cells(im_str, len);
+
     if (input_conv.vc_type != CONV_NONE)
 	vim_free(im_str);
     preedit_start_col += clen;
--- a/src/proto/mbyte.pro
+++ b/src/proto/mbyte.pro
@@ -14,6 +14,7 @@ int utf_ptr2cells __ARGS((char_u *p));
 int dbcs_ptr2cells __ARGS((char_u *p));
 int latin_ptr2cells_len __ARGS((char_u *p, int size));
 int latin_char2cells __ARGS((int c));
+int mb_string2cells __ARGS((char_u *p, int len));
 int latin_off2cells __ARGS((unsigned off, unsigned max_off));
 int dbcs_off2cells __ARGS((unsigned off, unsigned max_off));
 int utf_off2cells __ARGS((unsigned off, unsigned max_off));
--- a/src/screen.c
+++ b/src/screen.c
@@ -6168,8 +6168,8 @@ win_redr_status(wp)
 		int	clen = 0, i;
 
 		/* Count total number of display cells. */
-		for (i = 0; p[i] != NUL; i += (*mb_ptr2len)(p + i))
-		    clen += (*mb_ptr2cells)(p + i);
+		clen = mb_string2cells(p, -1);
+
 		/* Find first character that will fit.
 		 * Going from start to end is much faster for DBCS. */
 		for (i = 0; p[i] != NUL && clen >= this_ru_col - 1;