vim-patch:8.2.3184: cannot add a digraph with a leading space

Problem:    Cannot add a digraph with a leading space.  It is not easy to list
            existing digraphs.
Solution:   Add setdigraph(), setdigraphlist(), getdigraph() and
            getdigraphlist(). (closes vim/vim#8580)
https://github.com/vim/vim/commit/6106504e9edc8500131f7a36e59bc146f90180fa

Use GA_APPEND_VIA_PTR in registerdigraph().
Use tv_list_append_*() in getdigraphlist_appendpair().
Put the error messages in digraph.c.
E196 is N/A.
Remove mentions about 'encoding' being non-Unicode.
Nvim doesn't support setting encoding=japan, so skip a test.

8 files changed, 407 insertions, 26 deletions

diff --git a/runtime/doc/builtin.txt b/runtime/doc/builtin.txt
index 61010be4ef..dd4c25af70 100644
--- a/runtime/doc/builtin.txt
+++ b/runtime/doc/builtin.txt
@@ -185,6 +185,8 @@ getcompletion({pat}, {type} [, {filtered}])
 getcurpos([{winnr}])		List	position of the cursor
 getcursorcharpos([{winnr}])	List	character position of the cursor
 getcwd([{winnr} [, {tabnr}]])	String	get the current working directory
+getdigraph({chars})		String	get the digraph of {chars}
+getdigraphlist([{listall}])	List	get all |digraph|s
 getenv({name})			String	return environment variable
 getfontname([{name}])		String	name of font being used
 getfperm({fname})		String	file permissions of file {fname}
@@ -402,6 +404,8 @@ setcharpos({expr}, {list})	Number	set the {expr} position to {list}
 setcharsearch({dict})		Dict	set character search from {dict}
 setcmdpos({pos})		Number	set cursor position in command-line
 setcursorcharpos({list})	Number	move cursor to position in {list}
+setdigraph({chars}, {digraph})	Boolean  register |digraph|
+setdigraphlist({digraphlist})	Boolean  register multiple |digraph|s
 setenv({name}, {val})		none	set environment variable
 setfperm({fname}, {mode}	Number	set {fname} file permissions to {mode}
 setline({lnum}, {line})		Number	set line {lnum} to {line}
@@ -2891,6 +2895,45 @@ getcwd([{winnr} [, {tabnr}]])				*getcwd()*
 
 		Can also be used as a |method|: >
 			GetWinnr()->getcwd()
+<
+
+getdigraph({chars})				*getdigraph()* *E1214*
+		Return the digraph of {chars}.  This should be a string with
+		exactly two characters.  If {chars} are not just two
+		characters, or the digraph of {chars} does not exist, an error
+		is given and an empty string is returned.
+
+		Also see |getdigraphlist()|.
+
+		Examples: >
+		" Get a built-in digraph
+		:echo getdigraph('00')		" Returns '∞'
+
+		" Get a user-defined digraph
+		:call setdigraph('aa', 'あ')
+		:echo getdigraph('aa')		" Returns 'あ'
+<
+		Can also be used as a |method|: >
+			GetChars()->getdigraph()
+<
+
+getdigraphlist([{listall}])			*getdigraphlist()*
+		Return a list of digraphs.  If the {listall} argument is given
+		and it is TRUE, return all digraphs, including the default
+		digraphs.  Otherwise, return only user-defined digraphs.
+
+		Also see |getdigraph()|.
+
+		Examples: >
+		" Get user-defined digraphs
+		:echo getdigraphlist()
+
+		" Get all the digraphs, including default digraphs
+		:echo digraphlist(1)
+<
+		Can also be used as a |method|: >
+			GetNumber()->getdigraphlist()
+<
 
 getenv({name})						*getenv()*
 		Return the value of environment variable {name}.  The {name}
@@ -6734,6 +6777,46 @@ setcursorcharpos({list})
 		Can also be used as a |method|: >
 			GetCursorPos()->setcursorcharpos()
 
+
+setdigraph({chars}, {digraph})		*setdigraph()* *E1205*
+		Add digraph {chars} to the list.  {chars} must be a string
+		with two characters.  {digraph} is a string with one utf-8
+		encoded character. Be careful, composing characters are NOT
+		ignored.  This function is similar to |:digraphs| command, but
+		useful to add digraphs start with a white space.
+
+		The function result is v:true if |digraph| is registered.  If
+		this fails an error message is given and v:false is returned.
+
+		If you want to define multiple digraphs at once, you can use
+		|setdigraphlist()|.
+
+		Example: >
+			call setdigraph('  ', 'あ')
+<
+		Can be used as a |method|: >
+			GetString()->setdigraph('あ')
+<
+
+setdigraphlist({digraphlist})			*setdigraphlist()*
+		Similar to |setdigraph()| but this function can add multiple
+		digraphs at once.  {digraphlist} is a list composed of lists,
+		where each list contains two strings with {chars} and
+		{digraph} as in |setdigraph()|.
+		Example: >
+		    call setdigraphlist([['aa', 'あ'], ['ii', 'い']])
+<
+		It is similar to the following: >
+		    for [chars, digraph] in [['aa', 'あ'], ['ii', 'い']]
+			  call setdigraph(chars, digraph)
+		    endfor
+<		Except that the function returns after the first error,
+		following digraphs will not be added.
+
+		Can be used as a |method|: >
+		    GetList()->setdigraphlist()
+<
+
 setenv({name}, {val})						*setenv()*
 		Set environment variable {name} to {val}.  Example: >
 			call setenv('HOME', '/home/myhome')
diff --git a/runtime/doc/digraph.txt b/runtime/doc/digraph.txt
index dd7e9a331a..8b815ebee9 100644
--- a/runtime/doc/digraph.txt
+++ b/runtime/doc/digraph.txt
@@ -35,6 +35,9 @@ An alternative is using the 'keymap' option.
 <			Avoid defining a digraph with '_' (underscore) as the
 			first character, it has a special meaning in the
 			future.
+			NOTE: This command cannot add a digraph that starts
+			with a white space.  If you want to add such digraph,
+			you can use |setdigraph()| instead.
 
 Example of the output of ":digraphs": >
  TH Þ  222  ss ß  223  a! à  224  a' á  225  a> â  226  a? ã  227  a: ä  228
diff --git a/runtime/doc/usr_41.txt b/runtime/doc/usr_41.txt
index bf024315f6..42037a030b 100644
--- a/runtime/doc/usr_41.txt
+++ b/runtime/doc/usr_41.txt
@@ -964,6 +964,10 @@ Mappings:				    *mapping-functions*
 	mapcheck()		check if a matching mapping exists
 	maparg()		get rhs of a mapping
 	wildmenumode()		check if the wildmode is active
+	getdigraph()		get |digraph|
+	getdigraphlist()	get all |digraph|s
+	setdigraph()		register |digraph|
+	setdigraphlist()	register multiple |digraph|s
 
 Signs:						*sign-functions*
 	sign_define()		define or update a sign
diff --git a/src/nvim/digraph.c b/src/nvim/digraph.c
index 6e5389c979..4046869871 100644
--- a/src/nvim/digraph.c
+++ b/src/nvim/digraph.c
@@ -12,6 +12,7 @@
 #include "nvim/ascii.h"
 #include "nvim/charset.h"
 #include "nvim/digraph.h"
+#include "nvim/eval/typval.h"
 #include "nvim/ex_cmds2.h"
 #include "nvim/ex_docmd.h"
 #include "nvim/ex_getln.h"
@@ -34,6 +35,12 @@ typedef struct digraph {
   result_T result;
 } digr_T;
 
+static char e_digraph_must_be_just_two_characters_str[]
+    = N_("E1214: Digraph must be just two characters: %s");
+static char e_digraph_argument_must_be_one_character_str[]
+    = N_("E1215: Digraph must be one character: %s");
+static char e_setdigraphlist_argument_must_be_list_of_lists_with_two_items[]
+    = N_("E1216: setdigraphlist() argument must be a list of lists with two items");
 
 #ifdef INCLUDE_GENERATED_DECLARATIONS
 # include "digraph.c.generated.h"
@@ -1607,6 +1614,44 @@ int getdigraph(int char1, int char2, bool meta_char)
   return retval;
 }
 
+/// Add a digraph to the digraph table.
+static void registerdigraph(int char1, int char2, int n)
+{
+  // If the digraph already exists, replace "result".
+  digr_T *dp = (digr_T *)user_digraphs.ga_data;
+  for (int i = 0; i < user_digraphs.ga_len; i++) {
+    if ((int)dp->char1 == char1 && (int)dp->char2 == char2) {
+      dp->result = n;
+      return;
+    }
+    dp++;
+  }
+
+  // Add a new digraph to the table.
+  dp = GA_APPEND_VIA_PTR(digr_T, &user_digraphs);
+  dp->char1 = (char_u)char1;
+  dp->char2 = (char_u)char2;
+  dp->result = n;
+}
+
+/// Check the characters are valid for a digraph.
+/// If they are valid, returns true; otherwise, give an error message and
+/// returns false.
+bool check_digraph_chars_valid(int char1, int char2)
+{
+  if (char2 == 0) {
+    char_u msg[MB_MAXBYTES + 1];
+    msg[utf_char2bytes(char1, msg)] = NUL;
+    semsg(_(e_digraph_must_be_just_two_characters_str), msg);
+    return false;
+  }
+  if (char1 == ESC || char2 == ESC) {
+    emsg(_("E104: Escape not allowed in digraph"));
+    return false;
+  }
+  return true;
+}
+
 /// Add the digraphs in the argument to the digraph table.
 /// format: {c1}{c2} char {c1}{c2} char ...
 ///
@@ -1622,15 +1667,10 @@ void putdigraph(char_u *str)
     char_u char1 = *str++;
     char_u char2 = *str++;
 
-    if (char2 == 0) {
-      emsg(_(e_invarg));
+    if (!check_digraph_chars_valid(char1, char2)) {
       return;
     }
 
-    if ((char1 == ESC) || (char2 == ESC)) {
-      emsg(_("E104: Escape not allowed in digraph"));
-      return;
-    }
     str = skipwhite(str);
 
     if (!ascii_isdigit(*str)) {
@@ -1639,25 +1679,7 @@ void putdigraph(char_u *str)
     }
     int n = getdigits_int(&str, true, 0);
 
-    // If the digraph already exists, replace the result.
-    digr_T *dp = (digr_T *)user_digraphs.ga_data;
-
-    int i;
-    for (i = 0; i < user_digraphs.ga_len; i++) {
-      if (((int)dp->char1 == char1) && ((int)dp->char2 == char2)) {
-        dp->result = n;
-        break;
-      }
-      dp++;
-    }
-
-    // Add a new digraph to the table.
-    if (i == user_digraphs.ga_len) {
-      dp = GA_APPEND_VIA_PTR(digr_T, &user_digraphs);
-      dp->char1 = char1;
-      dp->char2 = char2;
-      dp->result = n;
-    }
+    registerdigraph(char1, char2, n);
   }
 }
 
@@ -1707,6 +1729,50 @@ void listdigraphs(bool use_headers)
   }
 }
 
+static void getdigraphlist_appendpair(digr_T *dp, list_T *l)
+{
+  list_T *l2 = tv_list_alloc(2);
+  tv_list_append_list(l, l2);
+
+  char_u buf[30];
+  buf[0] = dp->char1;
+  buf[1] = dp->char2;
+  buf[2] = NUL;
+  tv_list_append_string(l2, (char *)buf, -1);
+
+  char_u *p = buf;
+  p += utf_char2bytes(dp->result, p);
+  *p = NUL;
+  tv_list_append_string(l2, (char *)buf, -1);
+}
+
+void getdigraphlist_common(bool list_all, typval_T *rettv)
+{
+  tv_list_alloc_ret(rettv, (int)sizeof(digraphdefault) + user_digraphs.ga_len);
+
+  digr_T *dp;
+
+  if (list_all) {
+    dp = digraphdefault;
+    for (int i = 0; dp->char1 != NUL && !got_int; i++) {
+      digr_T tmp;
+      tmp.char1 = dp->char1;
+      tmp.char2 = dp->char2;
+      tmp.result = getexactdigraph(tmp.char1, tmp.char2, false);
+      if (tmp.result != 0 && tmp.result != tmp.char2) {
+        getdigraphlist_appendpair(&tmp, rettv->vval.v_list);
+      }
+      dp++;
+    }
+  }
+
+  dp = (digr_T *)user_digraphs.ga_data;
+  for (int i = 0; i < user_digraphs.ga_len && !got_int; i++) {
+    getdigraphlist_appendpair(dp, rettv->vval.v_list);
+    dp++;
+  }
+}
+
 struct dg_header_entry {
   int dg_start;
   const char *dg_header;
@@ -1797,6 +1863,147 @@ static void printdigraph(const digr_T *dp, result_T *previous)
   }
 }
 
+/// Get the two digraph characters from a typval.
+/// @return OK or FAIL.
+static int get_digraph_chars(typval_T *arg, int *char1, int *char2)
+{
+  char buf_chars[NUMBUFLEN];
+  const char *chars = tv_get_string_buf_chk(arg, buf_chars);
+  const char_u *p = (const char_u *)chars;
+
+  if (p != NULL) {
+    if (*p != NUL) {
+      *char1 = mb_cptr2char_adv(&p);
+      if (*p != NUL) {
+        *char2 = mb_cptr2char_adv(&p);
+        if (*p == NUL) {
+          if (check_digraph_chars_valid(*char1, *char2)) {
+            return OK;
+          }
+          return FAIL;
+        }
+      }
+    }
+  }
+  semsg(_(e_digraph_must_be_just_two_characters_str), chars);
+  return FAIL;
+}
+
+static bool setdigraph_common(typval_T *argchars, typval_T *argdigraph)
+{
+  int char1, char2;
+  if (get_digraph_chars(argchars, &char1, &char2) == FAIL) {
+    return false;
+  }
+
+  char buf_digraph[NUMBUFLEN];
+  const char *digraph = tv_get_string_buf_chk(argdigraph, buf_digraph);
+  if (digraph == NULL) {
+    return false;
+  }
+  const char_u *p = (const char_u *)digraph;
+  int n = mb_cptr2char_adv(&p);
+  if (*p != NUL) {
+    semsg(_(e_digraph_argument_must_be_one_character_str), digraph);
+    return false;
+  }
+
+  registerdigraph(char1, char2, n);
+  return true;
+}
+
+/// "getdigraph()" function
+void f_getdigraph(typval_T *argvars, typval_T *rettv, FunPtr fptr)
+{
+  rettv->v_type = VAR_STRING;
+  rettv->vval.v_string = NULL;  // Return empty string for failure
+  const char *digraphs = tv_get_string_chk(&argvars[0]);
+
+  if (digraphs == NULL) {
+    return;
+  }
+  if (STRLEN(digraphs) != 2) {
+    semsg(_(e_digraph_must_be_just_two_characters_str), digraphs);
+    return;
+  }
+  int code = getdigraph(digraphs[0], digraphs[1], false);
+
+  char_u buf[NUMBUFLEN];
+  buf[utf_char2bytes(code, buf)] = NUL;
+  rettv->vval.v_string = vim_strsave(buf);
+}
+
+/// "getdigraphlist()" function
+void f_getdigraphlist(typval_T *argvars, typval_T *rettv, FunPtr fptr)
+{
+  bool flag_list_all;
+
+  if (argvars[0].v_type == VAR_UNKNOWN) {
+    flag_list_all = false;
+  } else {
+    bool error = false;
+    varnumber_T flag = tv_get_number_chk(&argvars[0], &error);
+    if (error) {
+      return;
+    }
+    flag_list_all = flag != 0;
+  }
+
+  getdigraphlist_common(flag_list_all, rettv);
+}
+
+/// "setdigraph()" function
+void f_setdigraph(typval_T *argvars, typval_T *rettv, FunPtr fptr)
+{
+  rettv->v_type = VAR_BOOL;
+  rettv->vval.v_bool = kBoolVarFalse;
+
+  if (!setdigraph_common(&argvars[0], &argvars[1])) {
+    return;
+  }
+
+  rettv->vval.v_bool = kBoolVarTrue;
+}
+
+/// "setdigraphlist()" function
+void f_setdigraphlist(typval_T *argvars, typval_T *rettv, FunPtr fptr)
+{
+  rettv->v_type = VAR_BOOL;
+  rettv->vval.v_bool = kBoolVarFalse;
+
+  if (argvars[0].v_type != VAR_LIST) {
+    emsg(_(e_setdigraphlist_argument_must_be_list_of_lists_with_two_items));
+    return;
+  }
+
+  list_T *pl = argvars[0].vval.v_list;
+  if (pl == NULL) {
+    // Empty list always results in success.
+    rettv->vval.v_bool = kBoolVarTrue;
+    return;
+  }
+
+  TV_LIST_ITER_CONST(pl, pli, {
+    if (TV_LIST_ITEM_TV(pli)->v_type != VAR_LIST) {
+      emsg(_(e_setdigraphlist_argument_must_be_list_of_lists_with_two_items));
+      return;
+    }
+
+    list_T *l = TV_LIST_ITEM_TV(pli)->vval.v_list;
+    if (l == NULL || tv_list_len(l) != 2) {
+      emsg(_(e_setdigraphlist_argument_must_be_list_of_lists_with_two_items));
+      return;
+    }
+
+    if (!setdigraph_common(TV_LIST_ITEM_TV(tv_list_first(l)),
+                           TV_LIST_ITEM_TV(TV_LIST_ITEM_NEXT(l, tv_list_first(l))))) {
+      return;
+    }
+  });
+
+  rettv->vval.v_bool = kBoolVarTrue;
+}
+
 /// structure used for b_kmap_ga.ga_data
 typedef struct {
   char_u *from;
diff --git a/src/nvim/digraph.h b/src/nvim/digraph.h
index 71330ae9b1..80b4148213 100644
--- a/src/nvim/digraph.h
+++ b/src/nvim/digraph.h
@@ -2,6 +2,7 @@
 #define NVIM_DIGRAPH_H
 
 #include "nvim/ex_cmds_defs.h"
+#include "nvim/eval/funcs.h"
 #include "nvim/types.h"
 
 #ifdef INCLUDE_GENERATED_DECLARATIONS
diff --git a/src/nvim/eval.lua b/src/nvim/eval.lua
index 05e91a658f..0951cd239d 100644
--- a/src/nvim/eval.lua
+++ b/src/nvim/eval.lua
@@ -156,6 +156,8 @@ return {
     getcurpos={args={0, 1}, base=1},
     getcursorcharpos={args={0, 1}, base=1},
     getcwd={args={0, 2}, base=1},
+    getdigraph={args=1, base=1},
+    getdigraphlist={args={0, 1}, base=1},
     getenv={args=1, base=1},
     getfontname={args={0, 1}},
     getfperm={args=1, base=1},
@@ -322,6 +324,8 @@ return {
     setcharsearch={args=1, base=1},
     setcmdpos={args=1, base=1},
     setcursorcharpos={args={1, 3}, base=1},
+    setdigraph={args=2, base=1},
+    setdigraphlist={args=1, base=1},
     setenv={args=2, base=2},
     setfperm={args=2, base=1},
     setline={args=2, base=2},
diff --git a/src/nvim/eval/funcs.c b/src/nvim/eval/funcs.c
index d365e075e6..609db3990b 100644
--- a/src/nvim/eval/funcs.c
+++ b/src/nvim/eval/funcs.c
@@ -15,6 +15,7 @@
 #include "nvim/charset.h"
 #include "nvim/context.h"
 #include "nvim/cursor.h"
+#include "nvim/digraph.h"
 #include "nvim/diff.h"
 #include "nvim/edit.h"
 #include "nvim/eval.h"
diff --git a/src/nvim/testdir/test_digraph.vim b/src/nvim/testdir/test_digraph.vim
index 5965ee48ef..e9073db649 100644
--- a/src/nvim/testdir/test_digraph.vim
+++ b/src/nvim/testdir/test_digraph.vim
@@ -212,7 +212,7 @@ func Test_digraphs()
   call Put_Dig("el")
   call assert_equal(['␀', 'ü', '∞', 'l'], getline(line('.')-3,line('.')))
   call assert_fails('digraph xy z', 'E39:')
-  call assert_fails('digraph x', 'E474:')
+  call assert_fails('digraph x', 'E1214:')
   bw!
 endfunc
 
@@ -505,4 +505,82 @@ func Test_entering_digraph()
   call StopVimInTerminal(buf)
 endfunc
 
+func Test_setdigraph_function()
+  new
+  call setdigraph('aa', 'あ')
+  call Put_Dig('aa')
+  call assert_equal('あ', getline('$'))
+  call setdigraph(' i', 'い')
+  call Put_Dig(' i')
+  call assert_equal('い', getline('$'))
+  call setdigraph('  ', 'う')
+  call Put_Dig('  ')
+  call assert_equal('う', getline('$'))
+
+  eval 'aa'->setdigraph('え')
+  call Put_Dig('aa')
+  call assert_equal('え', getline('$'))
+
+  call assert_fails('call setdigraph("aaa", "あ")', 'E1214: Digraph must be just two characters: aaa')
+  call assert_fails('call setdigraph("b", "あ")', 'E1214: Digraph must be just two characters: b')
+  call assert_fails('call setdigraph("あ", "あ")', 'E1214: Digraph must be just two characters: あ')
+  call assert_fails('call setdigraph("aa", "ああ")', 'E1215: Digraph must be one character: ああ')
+  call assert_fails('call setdigraph("aa", "か" .. nr2char(0x3099))',  'E1215: Digraph must be one character: か' .. nr2char(0x3099))
+  bwipe!
+endfunc
+
+func Test_getdigraph_function()
+  " Built-in digraphs
+  call assert_equal('∞', getdigraph('00'))
+
+  " User-defined digraphs
+  call setdigraph('aa', 'あ')
+  call setdigraph(' i', 'い')
+  call setdigraph('  ', 'う')
+  call assert_equal('あ', getdigraph('aa'))
+  call assert_equal('あ', 'aa'->getdigraph())
+  call assert_equal('い', getdigraph(' i'))
+  call assert_equal('う', getdigraph('  '))
+  call assert_fails('call getdigraph("aaa")', 'E1214: Digraph must be just two characters: aaa')
+  call assert_fails('call getdigraph("b")', 'E1214: Digraph must be just two characters: b')
+endfunc
+
+func Test_getdigraph_function_encode()
+  throw 'Skipped: Nvim does not support setting encoding=japan'
+  CheckFeature iconv
+  let testcases = {
+        \'00': '∞',
+        \'aa': 'あ',
+        \}
+  for [key, ch] in items(testcases)
+    call setdigraph(key, ch)
+    set encoding=japan
+    call assert_equal(iconv(ch, 'utf-8', 'japan'), getdigraph(key))
+    set encoding&
+  endfor
+endfunc
+
+func Test_setdigraphlist_function()
+  call setdigraphlist([['aa', 'き'], ['bb', 'く']])
+  call assert_equal('き', getdigraph('aa'))
+  call assert_equal('く', getdigraph('bb'))
+
+  call assert_fails('call setdigraphlist([[]])', 'E1216:')
+  call assert_fails('call setdigraphlist([["aa", "b", "cc"]])', '1216:')
+  call assert_fails('call setdigraphlist([["あ", "あ"]])', 'E1214: Digraph must be just two characters: あ')
+endfunc
+
+func Test_getdigraphlist_function()
+  " Make sure user-defined digraphs are defined
+  call setdigraphlist([['aa', 'き'], ['bb', 'く']])
+
+  for pair in getdigraphlist(1)
+    call assert_equal(getdigraph(pair[0]), pair[1])
+  endfor
+
+  " We don't know how many digraphs are registered before, so check the number
+  " of digraphs returned.
+  call assert_equal(getdigraphlist()->len(), getdigraphlist(0)->len())
+  call assert_notequal((getdigraphlist()->len()), getdigraphlist(1)->len())
+endfunc
 " vim: shiftwidth=2 sts=2 expandtab