[RFA] Make word-break-characters language independent (1/3)

[RFA] Make word-break-characters language independent (1/3)

[Joel Brobecker]

[-- Attachment #1: Type: text/plain, Size: 1382 bytes --]

This is the first patch of a series of 3 that follow up on a discussion
about making the word_break_characters language dependent. See:
http://sources.redhat.com/ml/gdb-patches/2003-09/msg00399.html.

The plan is as follow:
  1. Add a new function in language: default_word_break_characters()
  2. Add a new field la_word_break_characters in struct language_defn
     Update all instances of this structure to set this new field
     to default_word_break_characters.
  3. Remove gdb_completer_word_break_characters and
     get_gdb_completer_word_break_characters. Replace them with
     either default_word_break_characters or the current_language
     la_word_break_characters.

In the end, GDB's behavior is entirely identical. But it will allow
Ada to define it's own set of word-break-characters, allowing it to
fine-tune the completion mechanism.

I divided this change in 3 patches to show the process I used, and
to show that the changes are mostly mechanical (and hopefully correct
:-). If necessary, I can break them up further (for instance patch3
can be broken up in two steps), or submit all 3 of them in one patch.
Let me know.

Attached is the first patch.

2003-10-01  J. Brobecker  <brobecker@gnat.com>

        * language.h (default_word_break_characters): Add prototype.
        * language.c (default_word_break_characters): New function.

Ok to apply?
-- 
Joel

[-- Attachment #2: patch1.diff --]
[-- Type: text/plain, Size: 1261 bytes --]

Index: language.c
===================================================================
RCS file: /cvs/src/src/gdb/language.c,v
retrieving revision 1.38
diff -u -p -r1.38 language.c
--- language.c	25 Sep 2003 08:40:45 -0000	1.38
+++ language.c	2 Oct 2003 04:59:56 -0000
@@ -1176,6 +1176,15 @@ language_demangle (const struct language
   return NULL;
 }
 
+/* Return the default string containing the list of characters
+   delimiting words.  This is a reasonable default value that
+   most languages should be able to use.  */
+
+char *
+default_word_break_characters (void)
+{
+  return " \t\n!@#$%^&*()+=|~`}{[]\"';:?/>.<,-";
+}
 
 /* Define the language that is no language.  */
 
Index: language.h
===================================================================
RCS file: /cvs/src/src/gdb/language.h,v
retrieving revision 1.22
diff -u -p -r1.22 language.h
--- language.h	25 Sep 2003 08:40:45 -0000	1.22
+++ language.h	2 Oct 2003 04:59:56 -0000
@@ -502,4 +502,7 @@ extern CORE_ADDR skip_language_trampolin
 extern char *language_demangle (const struct language_defn *current_language, 
 				const char *mangled, int options);
 
+/* Splitting strings into words.  */
+extern char *default_word_break_characters (void);
+
 #endif /* defined (LANGUAGE_H) */

Re: [RFA] Make word-break-characters language independent (1/3)

[Andrew Cagney]

> This is the first patch of a series of 3 that follow up on a discussion
> about making the word_break_characters language dependent. See:
> http://sources.redhat.com/ml/gdb-patches/2003-09/msg00399.html.
> 
> The plan is as follow:
>   1. Add a new function in language: default_word_break_characters()
>   2. Add a new field la_word_break_characters in struct language_defn
>      Update all instances of this structure to set this new field
>      to default_word_break_characters.
>   3. Remove gdb_completer_word_break_characters and
>      get_gdb_completer_word_break_characters. Replace them with
>      either default_word_break_characters or the current_language
>      la_word_break_characters.
> 
> In the end, GDB's behavior is entirely identical. But it will allow
> Ada to define it's own set of word-break-characters, allowing it to
> fine-tune the completion mechanism.
> 
> I divided this change in 3 patches to show the process I used, and
> to show that the changes are mostly mechanical (and hopefully correct
> :-). If necessary, I can break them up further (for instance patch3
> can be broken up in two steps), or submit all 3 of them in one patch.
> Let me know.

Eli, I think this is really completer stuff.  Ok?

(the language vector stuff works fine).

Andrew

Re: [RFA] Make word-break-characters language independent (1/3)

[Eli Zaretskii]

> Date: Thu, 02 Oct 2003 01:50:52 -0400
> From: Andrew Cagney <ac131313@redhat.com>
> 
> Eli, I think this is really completer stuff.  Ok?

The completion-related patches are okay with me.  Thanks, Joel.

Re: [RFA] Make word-break-characters language independent (1/3)

[Joel Brobecker]

On Fri, Oct 03, 2003 at 09:02:28PM +0200, Eli Zaretskii wrote:
> > Date: Thu, 02 Oct 2003 01:50:52 -0400
> > From: Andrew Cagney <ac131313@redhat.com>
> > 
> > Eli, I think this is really completer stuff.  Ok?
> 
> The completion-related patches are okay with me.  Thanks, Joel.

Thank you Eli. The patches have been checked in.

-- 
Joel