Merge pull request #3655 from bfredl/enc_default

Default to encoding=utf-8

runtime/doc/options.txt

@@ -2130,7 +2130,7 @@ A jump table for the options with a short description can be found at |Q_op|.
 'edcompatible' 'ed'	Removed. |vim-differences| {Nvim}
 
 					*'encoding'* *'enc'* *E543*
-'encoding' 'enc'	string (default: "utf-8" or value from $LANG)
+'encoding' 'enc'	string (default: "utf-8")
 			global
 			{only available when compiled with the |+multi_byte|
 			feature}
@@ -2152,10 +2152,6 @@ A jump table for the options with a short description can be found at |Q_op|.
 	can use: >
 		if has("multi_byte_encoding")
 <
-	Normally 'encoding' will be equal to your current locale.  This will
-	be the default if Vim recognizes your environment settings, otherwise
-	"utf-8" is used.
-
 	When you set this option, it fires the |EncodingChanged| autocommand
 	event so that you can set up fonts if necessary.
 
@@ -2172,9 +2168,6 @@ A jump table for the options with a short description can be found at |Q_op|.
 	setting 'encoding' to one of these values instead of utf-8 only has
 	effect for encoding used for files when 'fileencoding' is empty.
 
-	When 'encoding' is set to a Unicode encoding, and 'fileencodings' was
-	not set yet, the default for 'fileencodings' is changed.
-
 			*'endofline'* *'eol'* *'noendofline'* *'noeol'*
 'endofline' 'eol'	boolean	(default on)
 			local to buffer
@@ -2345,9 +2338,7 @@ A jump table for the options with a short description can be found at |Q_op|.
 	old short name was 'fe', which is no longer used.
 
 					*'fileencodings'* *'fencs'*
-'fileencodings' 'fencs'	string (default: "ucs-bom",
+'fileencodings' 'fencs'	string (default: "ucs-bom,utf-8,default,latin1")
-				    "ucs-bom,utf-8,default,latin1" when
-				    'encoding' is set to a Unicode value)
 			global
 			{only available when compiled with the |+multi_byte|
 			feature}
@@ -2387,9 +2378,8 @@ A jump table for the options with a short description can be found at |Q_op|.
 	because Vim cannot detect an error, thus the encoding is always
 	accepted.
 	The special value "default" can be used for the encoding from the
-	environment.  This is the default value for 'encoding'.  It is useful
+	environment.  It is useful when 'encoding' is set to "utf-8" and
-	when 'encoding' is set to "utf-8" and your environment uses a
+	your environment uses a non-latin1 encoding, such as Russian.
-	non-latin1 encoding, such as Russian.
 	When 'encoding' is "utf-8" and a file contains an illegal byte
 	sequence it won't be recognized as UTF-8.  You can use the |8g8|
 	command to find the illegal byte sequence.
@@ -3776,10 +3766,8 @@ A jump table for the options with a short description can be found at |Q_op|.
 	change 'iskeyword' instead.
 
 						*'iskeyword'* *'isk'*
-'iskeyword' 'isk'	string (Vim default for
+'iskeyword' 'isk'	string (default: @,48-57,_,192-255
-				  Win32:     @,48-57,_,128-167,224-235
+				Vi default: @,48-57,_)
-				  otherwise: @,48-57,_,192-255
-				Vi default:  @,48-57,_)
 			local to buffer
 	Keywords are used in searching and recognizing with many commands:
 	"w", "*", "[i", etc.  It is also used for "\k" in a |pattern|.  See
@@ -3791,8 +3779,7 @@ A jump table for the options with a short description can be found at |Q_op|.
 	When the 'lisp' option is on the '-' character is always included.
 
 						*'isprint'* *'isp'*
-'isprint' 'isp'	string	(default for MS-DOS, Win32, and Macintosh:
+'isprint' 'isp'	string	(default: "@,161-255")
-				"@,~-255"; otherwise: "@,161-255")
 			global
 	The characters given by this option are displayed directly on the
 	screen.  It is also used for "\p" in a |pattern|.  The characters from

src/nvim/globals.h

@@ -798,6 +798,8 @@ EXTERN bool enc_utf8 INIT(= false);             /* UTF-8 encoded Unicode */
 EXTERN int enc_latin1like INIT(= TRUE);         /* 'encoding' is latin1 comp. */
 EXTERN int has_mbyte INIT(= 0);                 /* any multi-byte encoding */
 
+/// Encoding used when 'fencs' is set to "default"
+EXTERN char_u *fenc_default INIT(= NULL);
 
 /*
  * To speed up BYTELEN() we fill a table with the byte lengths whenever

src/nvim/mbyte.c

@@ -568,11 +568,6 @@ char_u * mb_init(void)
   /* When enc_utf8 is set or reset, (de)allocate ScreenLinesUC[] */
   screenalloc(false);
 
-  /* When using Unicode, set default for 'fileencodings'. */
-  if (enc_utf8 && !option_was_set((char_u *)"fencs"))
-    set_string_option_direct((char_u *)"fencs", -1,
-        (char_u *)"ucs-bom,utf-8,default,latin1", OPT_FREE, 0);
-
 #ifdef HAVE_WORKING_LIBINTL
   /* GNU gettext 0.10.37 supports this feature: set the codeset used for
    * translated messages independently from the current locale. */
@@ -2417,11 +2412,8 @@ char_u *enc_canonize(char_u *enc) FUNC_ATTR_NONNULL_RET
   int i;
 
   if (STRCMP(enc, "default") == 0) {
-    /* Use the default encoding as it's found by set_init_1(). */
+    // Use the default encoding as found by set_init_1().
-    char_u *r = get_encoding_default();
+    return vim_strsave(fenc_default);
-    if (r == NULL)
-      r = (char_u *)"latin1";
-    return vim_strsave(r);
   }
 
   /* copy "enc" to allocated memory, with room for two '-' */

src/nvim/option.c

@@ -233,12 +233,6 @@ typedef struct vimoption {
 #define P_CURSWANT    0x2000000U /* update curswant required; not needed when
                                   * there is a redraw flag */
 
-#define ISK_LATIN1  (char_u *)"@,48-57,_,192-255"
-
-/* 'isprint' for latin1 is also used for MS-Windows cp1252, where 0x80 is used
- * for the currency sign. */
-# define ISP_LATIN1 (char_u *)"@,161-255"
-
 #define HIGHLIGHT_INIT \
   "8:SpecialKey,~:EndOfBuffer,z:TermCursor,Z:TermCursorNC,@:NonText," \
   "d:Directory,e:ErrorMsg,i:IncSearch,l:Search,m:MoreMsg,M:ModeMsg,n:LineNr," \
@@ -776,59 +770,18 @@ void set_init_1(void)
   /* Parse default for 'listchars'. */
   (void)set_chars_option(&p_lcs);
 
-  /* enc_locale() will try to find the encoding of the current locale. */
+  // enc_locale() will try to find the encoding of the current locale.
+  // This will be used when 'default' is used as encoding specifier
+  // in 'fileencodings'
   char_u *p = enc_locale();
-  if (p != NULL) {
+  if (p == NULL) {
-    char_u *save_enc;
+      // use utf-8 as 'default' if locale encoding can't be detected.
-
+      p = vim_strsave((char_u *)"utf-8");
-    /* Try setting 'encoding' and check if the value is valid.
-     * If not, go back to the default "utf-8". */
-    save_enc = p_enc;
-    p_enc = (char_u *) p;
-    if (STRCMP(p_enc, "gb18030") == 0) {
-      /* We don't support "gb18030", but "cp936" is a good substitute
-       * for practical purposes, thus use that.  It's not an alias to
-       * still support conversion between gb18030 and utf-8. */
-      p_enc = vim_strsave((char_u *)"cp936");
-      xfree(p);
-    }
-    if (mb_init() == NULL) {
-      opt_idx = findoption((char_u *)"encoding");
-      if (opt_idx >= 0) {
-        options[opt_idx].def_val[VI_DEFAULT] = p_enc;
-        options[opt_idx].flags |= P_DEF_ALLOCED;
-      }
-
-#if defined(MSWIN) || defined(MACOS)
-      if (STRCMP(p_enc, "latin1") == 0
-          || enc_utf8
-          ) {
-        /* Adjust the default for 'isprint' and 'iskeyword' to match
-         * latin1. */
-        set_string_option_direct((char_u *)"isp", -1,
-            ISP_LATIN1, OPT_FREE, SID_NONE);
-        set_string_option_direct((char_u *)"isk", -1,
-            ISK_LATIN1, OPT_FREE, SID_NONE);
-        opt_idx = findoption((char_u *)"isp");
-        if (opt_idx >= 0)
-          options[opt_idx].def_val[VIM_DEFAULT] = ISP_LATIN1;
-        opt_idx = findoption((char_u *)"isk");
-        if (opt_idx >= 0)
-          options[opt_idx].def_val[VIM_DEFAULT] = ISK_LATIN1;
-        (void)init_chartab();
-      }
-#endif
-
-    } else {
-      xfree(p_enc);
-      // mb_init() failed; fallback to utf8 and try again.
-      p_enc = save_enc;
-      mb_init();
-    }
-  } else {
-    // enc_locale() failed; initialize the default (utf8).
-    mb_init();
   }
+  fenc_default = p;
+
+  // Initialize multibyte (utf-8) handling
+  mb_init();
 
   // Don't change &encoding when resetting to defaults with ":set all&".
   opt_idx = findoption((char_u *)"encoding");
@@ -4669,16 +4622,6 @@ char_u *get_highlight_default(void)
   return (char_u *)NULL;
 }
 
-char_u *get_encoding_default(void)
-{
-  int i;
-
-  i = findoption((char_u *)"enc");
-  if (i >= 0)
-    return options[i].def_val[VI_DEFAULT];
-  return (char_u *)NULL;
-}
-
 /*
  * Translate a string like "t_xx", "<t_xx>" or "<S-Tab>" to a key number.
  */

src/nvim/options.lua

@@ -748,7 +748,7 @@ return {
       type='string', list='comma', scope={'global'},
       vi_def=true,
       varname='p_fencs',
-      defaults={if_true={vi="ucs-bom"}}
+      defaults={if_true={vi="ucs-bom,utf-8,default,latin1"}}
     },
     {
       full_name='fileformat', abbreviation='ff',
@@ -1285,7 +1285,7 @@ return {
       vim=true,
       alloced=true,
       varname='p_isk',
-      defaults={if_true={vi="@,48-57,_", vim=macros('ISK_LATIN1')}}
+      defaults={if_true={vi="@,48-57,_", vim="@,48-57,_,192-255"}}
     },
     {
       full_name='isprint', abbreviation='isp',
@@ -1294,10 +1294,7 @@ return {
       vi_def=true,
       redraw={'all_windows'},
       varname='p_isp',
-      defaults={
+      defaults={if_true={vi="@,161-255"}
-        condition='MSWIN',
-        if_true={vi="@,~-255"},
-        if_false={vi=macros("ISP_LATIN1")}
       }
     },
     {

test/functional/ex_cmds/encoding_spec.lua

@@ -21,20 +21,21 @@ describe('&encoding', function()
     eq(3, eval('strwidth("Bär")'))
   end)
 
-  it('is not changed by `set all&`', function()
+  it('can be changed before startup', function()
-    -- we need to set &encoding to something non-default
+    clear('set enc=latin1')
-    -- use 'latin1' when enc&vi is 'utf-8', 'utf-8' otherwise
+    execute('set encoding=utf-8')
-    execute('set fenc=default')
+    -- error message expected
-    local enc_default, enc_other, width = eval('&fenc'), 'utf-8', 3
+    feed('<cr>')
-    if enc_default == 'utf-8' then
+    eq('latin1', eval('&encoding'))
-      enc_other = 'latin1'
+    eq(4, eval('strwidth("Bär")'))
-      width = 4 -- utf-8 string 'Bär' will count as 4 latin1 chars
+  end)
-    end
 
-    clear('set enc=' .. enc_other)
+  it('is not changed by `set all&`', function()
+    -- we need to set &encoding to something non-default. Use 'latin1'
+    clear('set enc=latin1')
     execute('set all&')
-    eq(enc_other, eval('&encoding'))
+    eq('latin1', eval('&encoding'))
-    eq(width, eval('strwidth("Bär")'))
+    eq(4, eval('strwidth("Bär")'))
   end)
 
 end)

test/functional/helpers.lua

@@ -8,7 +8,7 @@ local Session = require('nvim.session')
 
 local nvim_prog = os.getenv('NVIM_PROG') or 'build/bin/nvim'
 local nvim_argv = {nvim_prog, '-u', 'NONE', '-i', 'NONE', '-N',
-                   '--cmd', 'set shortmess+=I background=light noswapfile noautoindent laststatus=1 encoding=utf-8 undodir=. directory=. viewdir=. backupdir=.',
+                   '--cmd', 'set shortmess+=I background=light noswapfile noautoindent laststatus=1 undodir=. directory=. viewdir=. backupdir=.',
                    '--embed'}
 
 -- Formulate a path to the directory containing nvim.  We use this to