patch 7.4.1042v7.4.1042

Problem:    g-CTRL-G shows the word count, but there is no way to get the word
            count in a script.
Solution:   Add the wordcount() function. (Christian Brabandt)

11 files changed, 293 insertions, 59 deletions

diff --git a/runtime/doc/editing.txt b/runtime/doc/editing.txt
index 307e20d07..1d104bc72 100644
--- a/runtime/doc/editing.txt
+++ b/runtime/doc/editing.txt
@@ -78,7 +78,9 @@ g CTRL-G		Prints the current position of the cursor in five
 			than one position on the screen (<Tab> or special
 			character), both the "real" column and the screen
 			column are shown, separated with a dash.
-			See also 'ruler' option.  {not in Vi}
+			Also see the 'ruler' option and the |wordcount()|
+			function.
+			{not in Vi}
 
 							*v_g_CTRL-G*
 {Visual}g CTRL-G	Similar to "g CTRL-G", but Word, Character, Line, and
diff --git a/runtime/doc/eval.txt b/runtime/doc/eval.txt
index e40116b06..10780bf64 100644
--- a/runtime/doc/eval.txt
+++ b/runtime/doc/eval.txt
@@ -1,4 +1,4 @@
-*eval.txt*	For Vim version 7.4.  Last change: 2016 Jan 02
+*eval.txt*	For Vim version 7.4.  Last change: 2016 Jan 03
 
 
 		  VIM REFERENCE MANUAL	  by Bram Moolenaar
@@ -2075,6 +2075,7 @@ winrestcmd()			String	returns command to restore window sizes
 winrestview( {dict})		none	restore view of current window
 winsaveview()			Dict	save view of current window
 winwidth( {nr})			Number	width of window {nr}
+wordcount()			Dict	get byte/char/word statistics
 writefile( {list}, {fname} [, {flags}])
 				Number	write list of lines to file {fname}
 xor( {expr}, {expr})		Number  bitwise XOR
@@ -6744,6 +6745,28 @@ winwidth({nr})						*winwidth()*
   :  exe "normal 50\<C-W>|"
   :endif
 <
+wordcount()						*wordcount()*
+		The result is a dictionary of byte/chars/word statistics for
+		the current buffer.  This is the same info as provided by
+		|g_CTRL-G|
+		The return value includes:
+			bytes		Number of bytes in the buffer
+			chars		Number of chars in the buffer
+			words		Number of words in the buffer
+			cursor_bytes    Number of bytes before cursor position
+					(not in Visual mode)
+			cursor_chars    Number of chars before cursor position
+					(not in Visual mode)
+			cursor_words    Number of words before cursor position
+					(not in Visual mode)
+			visual_bytes    Number of bytes visually selected
+			                (only in Visual mode)
+			visual_chars    Number of chars visually selected
+			                (only in Visual mode)
+			visual_words    Number of chars visually selected
+			                (only in Visual mode)
+
+
 							*writefile()*
 writefile({list}, {fname} [, {flags}])
 		Write |List| {list} to file {fname}.  Each list item is
diff --git a/runtime/doc/usr_41.txt b/runtime/doc/usr_41.txt
index eb1d5f7fb..2194dc14b 100644
--- a/runtime/doc/usr_41.txt
+++ b/runtime/doc/usr_41.txt
@@ -923,6 +923,7 @@ Various:					*various-functions*
 	mzeval()		evaluate |MzScheme| expression
 	py3eval()		evaluate Python expression (|+python3|)
 	pyeval()		evaluate Python expression (|+python|)
+	wordcount()             get byte/word/char count of buffer
 
 ==============================================================================
 *41.7*	Defining a function
diff --git a/src/eval.c b/src/eval.c
index 743c7436c..feaa71c33 100644
--- a/src/eval.c
+++ b/src/eval.c
@@ -780,6 +780,7 @@ static void f_winrestview __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_winsaveview __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_winwidth __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_writefile __ARGS((typval_T *argvars, typval_T *rettv));
+static void f_wordcount __ARGS((typval_T *argvars, typval_T *rettv));
 static void f_xor __ARGS((typval_T *argvars, typval_T *rettv));
 
 static int list2fpos __ARGS((typval_T *arg, pos_T *posp, int *fnump, colnr_T *curswantp));
@@ -8387,6 +8388,7 @@ static struct fst
     {"winrestview",	1, 1, f_winrestview},
     {"winsaveview",	0, 0, f_winsaveview},
     {"winwidth",	1, 1, f_winwidth},
+    {"wordcount",	0, 0, f_wordcount},
     {"writefile",	2, 3, f_writefile},
     {"xor",		2, 2, f_xor},
 };
@@ -20220,6 +20222,19 @@ f_winwidth(argvars, rettv)
 }
 
 /*
+ * "wordcount()" function
+ */
+    static void
+f_wordcount(argvars, rettv)
+    typval_T	*argvars UNUSED;
+    typval_T	*rettv;
+{
+    if (rettv_dict_alloc(rettv) == FAIL)
+	return;
+    cursor_pos_info(rettv->vval.v_dict);
+}
+
+/*
  * Write list of strings to file
  */
     static int
diff --git a/src/normal.c b/src/normal.c
index 89c67d9c8..640c0603e 100644
--- a/src/normal.c
+++ b/src/normal.c
@@ -8270,7 +8270,7 @@ nv_g_cmd(cap)
      * "g CTRL-G": display info about cursor position
      */
     case Ctrl_G:
-	cursor_pos_info();
+	cursor_pos_info(NULL);
 	break;
 
     /*
diff --git a/src/ops.c b/src/ops.c
index 31d7eb110..e710f66ef 100644
--- a/src/ops.c
+++ b/src/ops.c
@@ -6963,15 +6963,18 @@ line_count_info(line, wc, cc, limit, eol_size)
  * Give some info about the position of the cursor (for "g CTRL-G").
  * In Visual mode, give some info about the selected region.  (In this case,
  * the *_count_cursor variables store running totals for the selection.)
+ * When "dict" is not NULL store the info there instead of showing it.
  */
     void
-cursor_pos_info()
+cursor_pos_info(dict)
+    dict_T	*dict;
 {
     char_u	*p;
     char_u	buf1[50];
     char_u	buf2[40];
     linenr_T	lnum;
     long	byte_count = 0;
+    long	bom_count  = 0;
     long	byte_count_cursor = 0;
     long	char_count = 0;
     long	char_count_cursor = 0;
@@ -6989,7 +6992,11 @@ cursor_pos_info()
      */
     if (curbuf->b_ml.ml_flags & ML_EMPTY)
     {
-	MSG(_(no_lines_msg));
+	if (dict == NULL)
+	{
+	    MSG(_(no_lines_msg));
+	    return;
+	}
     }
     else
     {
@@ -7122,74 +7129,98 @@ cursor_pos_info()
 	if (!curbuf->b_p_eol && (curbuf->b_p_bin || !curbuf->b_p_fixeol))
 	    byte_count -= eol_size;
 
-	if (VIsual_active)
+	if (dict == NULL)
 	{
-	    if (VIsual_mode == Ctrl_V && curwin->w_curswant < MAXCOL)
+	    if (VIsual_active)
 	    {
-		getvcols(curwin, &min_pos, &max_pos, &min_pos.col,
-								&max_pos.col);
-		vim_snprintf((char *)buf1, sizeof(buf1), _("%ld Cols; "),
-			(long)(oparg.end_vcol - oparg.start_vcol + 1));
+		if (VIsual_mode == Ctrl_V && curwin->w_curswant < MAXCOL)
+		{
+		    getvcols(curwin, &min_pos, &max_pos, &min_pos.col,
+								    &max_pos.col);
+		    vim_snprintf((char *)buf1, sizeof(buf1), _("%ld Cols; "),
+			    (long)(oparg.end_vcol - oparg.start_vcol + 1));
+		}
+		else
+		    buf1[0] = NUL;
+
+		if (char_count_cursor == byte_count_cursor
+						    && char_count == byte_count)
+		    vim_snprintf((char *)IObuff, IOSIZE,
+			    _("Selected %s%ld of %ld Lines; %ld of %ld Words; %ld of %ld Bytes"),
+			    buf1, line_count_selected,
+			    (long)curbuf->b_ml.ml_line_count,
+			    word_count_cursor, word_count,
+			    byte_count_cursor, byte_count);
+		else
+		    vim_snprintf((char *)IObuff, IOSIZE,
+			    _("Selected %s%ld of %ld Lines; %ld of %ld Words; %ld of %ld Chars; %ld of %ld Bytes"),
+			    buf1, line_count_selected,
+			    (long)curbuf->b_ml.ml_line_count,
+			    word_count_cursor, word_count,
+			    char_count_cursor, char_count,
+			    byte_count_cursor, byte_count);
 	    }
 	    else
-		buf1[0] = NUL;
-
-	    if (char_count_cursor == byte_count_cursor
-						  && char_count == byte_count)
-		vim_snprintf((char *)IObuff, IOSIZE,
-			_("Selected %s%ld of %ld Lines; %ld of %ld Words; %ld of %ld Bytes"),
-			buf1, line_count_selected,
+	    {
+		p = ml_get_curline();
+		validate_virtcol();
+		col_print(buf1, sizeof(buf1), (int)curwin->w_cursor.col + 1,
+			(int)curwin->w_virtcol + 1);
+		col_print(buf2, sizeof(buf2), (int)STRLEN(p),
+				    linetabsize(p));
+
+		if (char_count_cursor == byte_count_cursor
+			&& char_count == byte_count)
+		    vim_snprintf((char *)IObuff, IOSIZE,
+			_("Col %s of %s; Line %ld of %ld; Word %ld of %ld; Byte %ld of %ld"),
+			(char *)buf1, (char *)buf2,
+			(long)curwin->w_cursor.lnum,
 			(long)curbuf->b_ml.ml_line_count,
 			word_count_cursor, word_count,
 			byte_count_cursor, byte_count);
-	    else
-		vim_snprintf((char *)IObuff, IOSIZE,
-			_("Selected %s%ld of %ld Lines; %ld of %ld Words; %ld of %ld Chars; %ld of %ld Bytes"),
-			buf1, line_count_selected,
+		else
+		    vim_snprintf((char *)IObuff, IOSIZE,
+			_("Col %s of %s; Line %ld of %ld; Word %ld of %ld; Char %ld of %ld; Byte %ld of %ld"),
+			(char *)buf1, (char *)buf2,
+			(long)curwin->w_cursor.lnum,
 			(long)curbuf->b_ml.ml_line_count,
 			word_count_cursor, word_count,
 			char_count_cursor, char_count,
 			byte_count_cursor, byte_count);
-	}
-	else
-	{
-	    p = ml_get_curline();
-	    validate_virtcol();
-	    col_print(buf1, sizeof(buf1), (int)curwin->w_cursor.col + 1,
-		    (int)curwin->w_virtcol + 1);
-	    col_print(buf2, sizeof(buf2), (int)STRLEN(p),
-				linetabsize(p));
-
-	    if (char_count_cursor == byte_count_cursor
-		    && char_count == byte_count)
-		vim_snprintf((char *)IObuff, IOSIZE,
-		    _("Col %s of %s; Line %ld of %ld; Word %ld of %ld; Byte %ld of %ld"),
-		    (char *)buf1, (char *)buf2,
-		    (long)curwin->w_cursor.lnum,
-		    (long)curbuf->b_ml.ml_line_count,
-		    word_count_cursor, word_count,
-		    byte_count_cursor, byte_count);
-	    else
-		vim_snprintf((char *)IObuff, IOSIZE,
-		    _("Col %s of %s; Line %ld of %ld; Word %ld of %ld; Char %ld of %ld; Byte %ld of %ld"),
-		    (char *)buf1, (char *)buf2,
-		    (long)curwin->w_cursor.lnum,
-		    (long)curbuf->b_ml.ml_line_count,
-		    word_count_cursor, word_count,
-		    char_count_cursor, char_count,
-		    byte_count_cursor, byte_count);
+	    }
 	}
 
+	/* Don't shorten this message, the user asked for it. */
 #ifdef FEAT_MBYTE
-	byte_count = bomb_size();
-	if (byte_count > 0)
+	bom_count = bomb_size();
+	if (bom_count > 0)
 	    sprintf((char *)IObuff + STRLEN(IObuff), _("(+%ld for BOM)"),
-								  byte_count);
+								bom_count);
 #endif
-	/* Don't shorten this message, the user asked for it. */
-	p = p_shm;
-	p_shm = (char_u *)"";
-	msg(IObuff);
-	p_shm = p;
+	if (dict == NULL)
+	{
+	    p = p_shm;
+	    p_shm = (char_u *)"";
+	    msg(IObuff);
+	    p_shm = p;
+	}
+    }
+    if (dict != NULL)
+    {
+	dict_add_nr_str(dict, "words", (long)word_count, NULL);
+	dict_add_nr_str(dict, "chars", (long)char_count, NULL);
+	dict_add_nr_str(dict, "bytes", (long)byte_count + bom_count, NULL);
+	if (VIsual_active)
+	{
+	    dict_add_nr_str(dict, "visual_bytes", (long)byte_count_cursor, NULL);
+	    dict_add_nr_str(dict, "visual_chars", (long)char_count_cursor, NULL);
+	    dict_add_nr_str(dict, "visual_words", (long)word_count_cursor, NULL);
+	}
+	else
+	{
+	    dict_add_nr_str(dict, "cursor_bytes", (long)byte_count_cursor, NULL);
+	    dict_add_nr_str(dict, "cursor_chars", (long)char_count_cursor, NULL);
+	    dict_add_nr_str(dict, "cursor_words", (long)word_count_cursor, NULL);
+	}
     }
 }
diff --git a/src/proto/ops.pro b/src/proto/ops.pro
index f9b53c08f..aaffa0202 100644
--- a/src/proto/ops.pro
+++ b/src/proto/ops.pro
@@ -58,5 +58,5 @@ void write_reg_contents __ARGS((int name, char_u *str, int maxlen, int must_appe
 void write_reg_contents_lst __ARGS((int name, char_u **strings, int maxlen, int must_append, int yank_type, long block_len));
 void write_reg_contents_ex __ARGS((int name, char_u *str, int maxlen, int must_append, int yank_type, long block_len));
 void clear_oparg __ARGS((oparg_T *oap));
-void cursor_pos_info __ARGS((void));
+void cursor_pos_info __ARGS((dict_T *eval));
 /* vim: set ft=c : */
diff --git a/src/testdir/Make_all.mak b/src/testdir/Make_all.mak
index c5796bd51..a67619f46 100644
--- a/src/testdir/Make_all.mak
+++ b/src/testdir/Make_all.mak
@@ -120,6 +120,7 @@ SCRIPTS_ALL = \
 	test_tagcase.out \
 	test_textobjects.out \
 	test_utf8.out \
+	test_wordcount.out \
 	test_writefile.out
 
 
diff --git a/src/testdir/test_wordcount.in b/src/testdir/test_wordcount.in
new file mode 100644
index 000000000..b2db50530
--- /dev/null
+++ b/src/testdir/test_wordcount.in
@@ -0,0 +1,125 @@
+Test for wordcount() function
+
+STARTTEST
+:so small.vim
+:so mbyte.vim
+:set enc=utf8
+:new
+:fu DoRecordWin(...)
+:	wincmd k
+:       if exists("a:1")
+:           call cursor(a:1)
+:       endif
+:       let result=[]
+:       call add(result, g:test)
+:       call add(result, getline(1, '$'))
+:       call add(result, wordcount())
+:	wincmd j
+:       return result
+:endfu
+:fu PutInWindow(args)
+:       wincmd k
+:       %d _
+:       call append(1, a:args)
+:	wincmd j
+:endfu
+:fu Log()
+:   $put ='----'
+:   $put =remove(g:log,0)
+:   $put =string(g:log)
+:endfu
+:fu! STL()
+:    if mode() =~? 'V'
+:       let g:visual_stat=wordcount()
+:    endif
+:    return string(wordcount())
+:endfu
+:let g:test="Test 1: empty window"
+:let log=DoRecordWin()
+:call Log()
+:"
+:let g:test="Test 2: some words, cursor at start"
+:call PutInWindow('one two three')
+:let log=DoRecordWin([1,1,0])
+:call Log()
+:"
+:let g:test="Test 3: some words, cursor at end"
+:call PutInWindow('one two three')
+:let log=DoRecordWin([2,99,0])
+:call Log()
+:"
+:let g:test="Test 4: some words, cursor at end, ve=all"
+:set ve=all
+:call PutInWindow('one two three')
+:let log=DoRecordWin([2,99,0])
+:call Log()
+:set ve=
+:"
+:let g:test="Test 5: several lines with words"
+:call PutInWindow(['one two three', 'one two three', 'one two three'])
+:let log=DoRecordWin([4,99,0])
+:call Log()
+:"
+:let g:test="Test 6: one line with BOM set"
+:call PutInWindow('one two three')
+:wincmd k
+:set bomb
+:w! Xtest
+:wincmd j
+:let log=DoRecordWin([2,99,0])
+:call Log()
+:wincmd k
+:set nobomb
+:w!
+:wincmd j
+:"
+:let g:test="Test 7: one line with multibyte words"
+:call PutInWindow(['Äne M¤ne Müh'])
+:let log=DoRecordWin([2,99,0])
+:call Log()
+:"
+:let g:test="Test 8: several lines with multibyte words"
+:call PutInWindow(['Äne M¤ne Müh', 'und raus bist dü!'])
+:let log=DoRecordWin([3,99,0])
+:call Log()
+:"
+:let g:test="Test 9: visual mode, complete buffer"
+:call PutInWindow(['Äne M¤ne Müh', 'und raus bist dü!'])
+:wincmd k
+:set ls=2 stl=%{STL()}
+:" start visual mode quickly and select complete buffer
+:0
+V2jy
+:set stl= ls=1
+:let log=DoRecordWin([3,99,0])
+:let log[2]=g:visual_stat
+:call Log()
+:"
+:let g:test="Test 10: visual mode (empty)"
+:call PutInWindow(['Äne M¤ne Müh', 'und raus bist dü!'])
+:wincmd k
+:set ls=2 stl=%{STL()}
+:" start visual mode quickly and select complete buffer
+:0
+v$y
+:set stl= ls=1
+:let log=DoRecordWin([3,99,0])
+:let log[2]=g:visual_stat
+:call Log()
+:"
+:let g:test="Test 11: visual mode, single line"
+:call PutInWindow(['Äne M¤ne Müh', 'und raus bist dü!'])
+:wincmd k
+:set ls=2 stl=%{STL()}
+:" start visual mode quickly and select complete buffer
+:2
+0v$y
+:set stl= ls=1
+:let log=DoRecordWin([3,99,0])
+:let log[2]=g:visual_stat
+:call Log()
+:"
+:/^RESULT test/,$w! test.out
+:qa!
+ENDTEST
+RESULT test:
diff --git a/src/testdir/test_wordcount.ok b/src/testdir/test_wordcount.ok
new file mode 100644
index 000000000..4bc8a9390
--- /dev/null
+++ b/src/testdir/test_wordcount.ok
@@ -0,0 +1,34 @@
+RESULT test:
+----
+Test 1: empty window
+[[''], {'chars': 0, 'cursor_chars': 0, 'words': 0, 'cursor_words': 0, 'bytes': 0, 'cursor_bytes': 0}]
+----
+Test 2: some words, cursor at start
+[['', 'one two three'], {'chars': 15, 'cursor_chars': 1, 'words': 3, 'cursor_words': 0, 'bytes': 15, 'cursor_bytes': 1}]
+----
+Test 3: some words, cursor at end
+[['', 'one two three'], {'chars': 15, 'cursor_chars': 14, 'words': 3, 'cursor_words': 3, 'bytes': 15, 'cursor_bytes': 14}]
+----
+Test 4: some words, cursor at end, ve=all
+[['', 'one two three'], {'chars': 15, 'cursor_chars': 15, 'words': 3, 'cursor_words': 3, 'bytes': 15, 'cursor_bytes': 15}]
+----
+Test 5: several lines with words
+[['', 'one two three', 'one two three', 'one two three'], {'chars': 43, 'cursor_chars': 42, 'words': 9, 'cursor_words': 9, 'bytes': 43, 'cursor_bytes': 42}]
+----
+Test 6: one line with BOM set
+[['', 'one two three'], {'chars': 15, 'cursor_chars': 14, 'words': 3, 'cursor_words': 3, 'bytes': 18, 'cursor_bytes': 14}]
+----
+Test 7: one line with multibyte words
+[['', 'Äne M¤ne Müh'], {'chars': 14, 'cursor_chars': 13, 'words': 3, 'cursor_words': 3, 'bytes': 17, 'cursor_bytes': 16}]
+----
+Test 8: several lines with multibyte words
+[['', 'Äne M¤ne Müh', 'und raus bist dü!'], {'chars': 32, 'cursor_chars': 31, 'words': 7, 'cursor_words': 7, 'bytes': 36, 'cursor_bytes': 35}]
+----
+Test 9: visual mode, complete buffer
+[['', 'Äne M¤ne Müh', 'und raus bist dü!'], {'chars': 32, 'words': 7, 'bytes': 36, 'visual_chars': 32, 'visual_words': 7, 'visual_bytes': 36}]
+----
+Test 10: visual mode (empty)
+[['', 'Äne M¤ne Müh', 'und raus bist dü!'], {'chars': 32, 'words': 7, 'bytes': 36, 'visual_chars': 1, 'visual_words': 0, 'visual_bytes': 1}]
+----
+Test 11: visual mode, single line
+[['', 'Äne M¤ne Müh', 'und raus bist dü!'], {'chars': 32, 'words': 7, 'bytes': 36, 'visual_chars': 13, 'visual_words': 3, 'visual_bytes': 16}]
diff --git a/src/version.c b/src/version.c
index ab777e394..821b1bb0a 100644
--- a/src/version.c
+++ b/src/version.c
@@ -742,6 +742,8 @@ static char *(features[]) =
 static int included_patches[] =
 {   /* Add new patch number below this line */
 /**/
+    1042,
+/**/
     1041,
 /**/
     1040,