[RFA] File-name completion improvements

[Eli Zaretskii <eliz@is.elta.co.il>]

I was never happy with GDB's completion: I think it doesn't quite help as 
much as it could.  For example, there are commands which mostly operate 
on files, but insist to complete on symbols instead.

So I will be submitting a series of patches to make this better.  But 
first, I need to make the file-name completion DTRT; currently, it 
doesn't.  For example, try "dir >foo TAB" or "dir foo:bar TAB" (assuming 
you have a directory or a file which begin with `foo' and `bar').

(Yes, I know "dir >foo" is nonsense, but GDB shouldn't prevent me from 
saying that; there are GDB commands which can use redirection near file 
names.)

The following patches correct several minor but annoying problems in 
file-name completion:

  - completion on a file name in a list of file names didn't work;
  - GDB would not always append a slash if the completion is a directory;
  - completion failed when the file name had non-file-name characters,
    such as redirection, around it;
  - on DOS/Windows, completion would fail with files with a drive letter.

Okay to commit?

2001-02-12  Eli Zaretskii  <eliz@is.elta.co.il>

	* completer.c (gdb_completer_file_name_break_characters): Remove
	slash from file-name break characters.
	[__MSDOS__]: Special definition for DOS/Windows file names.
	(line_completion_function): When completing on file names, bump
	`p' to the first file-name constituent character of `word', before
	invoking the completer.

--- gdb/completer.c~0	Fri Dec 15 03:01:46 2000
+++ gdb/completer.c	Mon Feb 12 00:28:42 2001
@@ -64,7 +64,13 @@ static char *gdb_completer_command_word_
    break characters any characters that are commonly used in file
    names, such as '-', '+', '~', etc.  Otherwise, readline displays
    incorrect completion candidates.  */
-static char *gdb_completer_file_name_break_characters = " \t\n*|\"';:?/><";
+#ifdef __MSDOS__
+/* MS-DOS and MS-Windows use colon as part of the drive spec, and most
+   programs support @foo style response files.  */
+static char *gdb_completer_file_name_break_characters = " \t\n*|\"';?><@";
+#else
+static char *gdb_completer_file_name_break_characters = " \t\n*|\"';:?><";
+#endif
 
 /* Characters that can be used to quote completion strings.  Note that we
    can't include '"' because the gdb C parser treats such quoted sequences
@@ -348,10 +354,25 @@ line_completion_function (char *text, in
 		    {
 		      /* It is a normal command; what comes after it is
 		         completed by the command's completer function.  */
-		      list = (*c->completer) (p, word);
 		      if (c->completer == filename_completer)
-			rl_completer_word_break_characters =
-			  gdb_completer_file_name_break_characters;
+			{
+			  char *fbc = gdb_completer_file_name_break_characters;
+
+			  /* Many commands which want to complete on
+			     file names accept several file names, as
+			     in "run foo bar >>baz".  So we don't want
+			     to complete the entire text after the
+			     command, just the last word.  To this
+			     end, we need to find the beginning of the
+			     file name starting at word and going
+			     backwards.  */
+			  for (p = word;
+			       p > tmp_command && strchr (fbc, p[-1]) == NULL;
+			       p--)
+			    ;
+			  rl_completer_word_break_characters = fbc;
+			}
+		      list = (*c->completer) (p, word);
 		    }
 		}
 	      else
@@ -397,10 +418,19 @@ line_completion_function (char *text, in
 	      else
 		{
 		  /* It is a normal command.  */
-		  list = (*c->completer) (p, word);
 		  if (c->completer == filename_completer)
-		    rl_completer_word_break_characters =
-		      gdb_completer_file_name_break_characters;
+		    {
+		      char *fbc = gdb_completer_file_name_break_characters;
+
+		      /* See the commentary above about the specifics
+			 of file-name completion.  */
+		      for (p = word;
+			   p > tmp_command && strchr (fbc, p[-1]) == NULL;
+			   p--)
+			;
+		      rl_completer_word_break_characters = fbc;
+		    }
+		  list = (*c->completer) (p, word);
 		}
 	    }
 	}