RFA: add_setshow_cmd_full

RFA: add_setshow_cmd_full

[Tom Tromey]

In some situations we need to get both the set and show command
structure back when creating a set/show pair.  I propose a new
add_setshow_cmd_full function that returns both.

This patch includes my previous patch to make add_setshow_cmd public.

Ok?  Or is there another way you'd prefer I approach this problem?

I've converted all the add_show_from_set calls to use add_setshow_cmd.
This patch or something like it has to go in before I can completely
finish that patch.  At this point I'm going to stop working on
gettextization until the patch backlog clears up a bit.  It gets hard
to manage the patches when many files have already been touched.

Tom

2002-06-25  Tom Tromey  <tromey@redhat.com>

	* cli/cli-decode.c (add_setshow_cmd_full): New function.
	(add_setshow_cmd): Use it.
	* command.h (add_setshow_cmd_full): Declare.

Index: cli/cli-decode.c
===================================================================
RCS file: /cvs/src/src/gdb/cli/cli-decode.c,v
retrieving revision 1.25
diff -u -r1.25 cli-decode.c
--- cli/cli-decode.c 25 Jun 2002 05:39:18 -0000 1.25
+++ cli/cli-decode.c 25 Jun 2002 20:37:04 -0000
@@ -331,14 +331,16 @@
    command.  SET_FUNC and SHOW_FUNC are the callback functions (if
    non-NULL).  SET_DOC and SHOW_DOC are the documentation strings.  */
 
-static struct cmd_list_element *
-add_setshow_cmd (char *name,
-		 enum command_class class,
-		 var_types var_type, void *var,
-		 char *set_doc, char *show_doc,
-		 cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
-		 struct cmd_list_element **set_list,
-		 struct cmd_list_element **show_list)
+void
+add_setshow_cmd_full (char *name,
+		      enum command_class class,
+		      var_types var_type, void *var,
+		      char *set_doc, char *show_doc,
+		      cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
+		      struct cmd_list_element **set_list,
+		      struct cmd_list_element **show_list,
+		      struct cmd_list_element **set_result,
+		      struct cmd_list_element **show_result)
 {
   struct cmd_list_element *set;
   struct cmd_list_element *show;
@@ -350,9 +352,35 @@
 			      show_doc, show_list);
   if (show_func != NULL)
     set_cmd_sfunc (show, show_func);
+
+  if (set_result != NULL)
+    *set_result = set;
+  if (show_result != NULL)
+    *show_result = show;
+}
+
+/* Add element named NAME to both the command SET_LIST and SHOW_LIST.
+   CLASS is as in add_cmd.  VAR_TYPE is the kind of thing we are
+   setting.  VAR is address of the variable being controlled by this
+   command.  SET_FUNC and SHOW_FUNC are the callback functions (if
+   non-NULL).  SET_DOC and SHOW_DOC are the documentation strings.  */
+
+struct cmd_list_element *
+add_setshow_cmd (char *name,
+		 enum command_class class,
+		 var_types var_type, void *var,
+		 char *set_doc, char *show_doc,
+		 cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
+		 struct cmd_list_element **set_list,
+		 struct cmd_list_element **show_list)
+{
+  cmd_list_element *result;
   /* The caller often wants to modify set to include info like an
      enumeration.  */
-  return set;
+  add_setshow_cmd_full (name, class, var_type, var, set_doc, show_doc,
+			set_func, show_func, set_list, show_list,
+			&result, NULL);
+  return result;
 }
 
 struct cmd_list_element *
Index: command.h
===================================================================
RCS file: /cvs/src/src/gdb/command.h,v
retrieving revision 1.33
diff -u -r1.33 command.h
--- command.h 15 Jun 2002 22:05:32 -0000 1.33
+++ command.h 25 Jun 2002 20:37:04 -0000
@@ -210,6 +210,28 @@
 extern void help_cmd_list (struct cmd_list_element *, enum command_class,
 			   char *, int, struct ui_file *);
 
+extern struct cmd_list_element *add_setshow_cmd (char *name,
+						 enum command_class class,
+						 var_types var_type, void *var,
+						 char *set_doc, char *show_doc,
+						 cmd_sfunc_ftype *set_func,
+						 cmd_sfunc_ftype *show_func,
+						 struct cmd_list_element **set_list,
+						 struct cmd_list_element **show_list);
+
+extern struct cmd_list_element *add_setshow_cmd_full (char *name,
+						      enum command_class class,
+						      var_types var_type,
+						      void *var,
+						      char *set_doc,
+						      char *show_doc,
+						      cmd_sfunc_ftype *set_func,
+						      cmd_sfunc_ftype *show_func,
+						      struct cmd_list_element **set_list,
+						      struct cmd_list_element **show_list,
+						      struct cmd_list_element **set_result,
+						      struct cmd_list_element **show_result);
+
 extern struct cmd_list_element *add_set_cmd (char *name, enum
 					     command_class class,
 					     var_types var_type, void *var,

Re: RFA: add_setshow_cmd_full

[Andrew Cagney]

[fernando asked me to pick up a bit of this]


> 2002-06-25  Tom Tromey  <tromey@redhat.com>
> 
> 	* cli/cli-decode.c (add_setshow_cmd_full): New function.
> 	(add_setshow_cmd): Use it.
> 	* command.h (add_setshow_cmd_full): Declare.

Please be careful to configure GDB with 
--enable-gdb-build-warnings=,-Werror so that it is compiled with the 
-Werror flag.

Anyway, this is wrong ...

 > +extern struct cmd_list_element *add_setshow_cmd_full (char *name,
 > +		 
			      enum command_class class,

But this is approved ....

> +void
> +add_setshow_cmd_full (char *name,
> +		      enum command_class class,
> +		      var_types var_type, void *var,
> +		      char *set_doc, char *show_doc,
> +		      cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
> +		      struct cmd_list_element **set_list,
> +		      struct cmd_list_element **show_list,
> +		      struct cmd_list_element **set_result,
> +		      struct cmd_list_element **show_result)

Don't forget to remove they stray declaration:

> +extern struct cmd_list_element *add_setshow_cmd (char *name,
> +						 enum command_class class,
> +						 var_types var_type, void *var,
> +						 char *set_doc, char *show_doc,

from the change.

Andrew

Re: RFA: add_setshow_cmd_full

[Eli Zaretskii]

On 25 Jun 2002, Tom Tromey wrote:

> This patch includes my previous patch to make add_setshow_cmd public.

Sounds good, but could you please add this function to gdbint.texinfo?

TIA

Re: RFA: add_setshow_cmd_full

[Tom Tromey]

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:

Andrew> Please be careful to configure GDB with
Andrew> --enable-gdb-build-warnings=,-Werror so that it is compiled
Andrew> with the -Werror flag.

Thanks for catching this.  I'm rebuilding with that right now; I wired
it into my standard build script so I won't make this mistake again.

Andrew> Don't forget to remove they stray declaration:

>> +extern struct cmd_list_element *add_setshow_cmd (char *name,
>> +						 enum command_class class,
>> +						 var_types var_type, void *var,
>> +						 char *set_doc, char *show_doc,

Andrew> from the change.

That declaration is part of the change, since I wanted to make
add_setshow_cmd public, and add add_setshow_cmd_full as well.  The
latter is only used in a few places.

Would you prefer I use add_setshow_cmd_full everywhere, and just have
90% of the calls have `NULL, NULL' at the end?  I can do that...

Tom

Re: RFA: add_setshow_cmd_full

[Tom Tromey]

>>>>> "Eli" == Eli Zaretskii <eliz@is.elta.co.il> writes:

>> This patch includes my previous patch to make add_setshow_cmd public.

Eli> Sounds good, but could you please add this function to
Eli> gdbint.texinfo?

Will do.  Thanks.

Tom

Re: RFA: add_setshow_cmd_full

[Andrew Cagney]

> That declaration is part of the change, since I wanted to make
> add_setshow_cmd public, and add add_setshow_cmd_full as well.  The
> latter is only used in a few places.
> 
> Would you prefer I use add_setshow_cmd_full everywhere, and just have
> 90% of the calls have `NULL, NULL' at the end?  I can do that...

No, I was looking at them separatly.

If I understand right, with the above applied, a new 
add_setshow_cmd(..), that returns void, can be implemented using 
add_setshow_cmd_full().

Andrew

Re: RFA: add_setshow_cmd_full

[Tom Tromey]

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:

Andrew> If I understand right, with the above applied, a new
Andrew> add_setshow_cmd(..), that returns void, can be implemented using
Andrew> add_setshow_cmd_full().

Yes.  I went ahead and made that change.  The new patch is appended.
It subsumes the last two add_setshow_cmd() patches.

This time built with -Werror :-).

Ok to commit?

After this I'll submit my jumbo "use add_setshow_cmd instead of
add_show_from_set" patch.

Tom

2002-06-26  Tom Tromey  <tromey@redhat.com>

	* command.h (add_setshow_cmd): Declare.
	(add_setshow_cmd_full): Declare.
	* cli/cli-decode.c (add_setshow_cmd): No longer static.  Now
	returns void.  Use add_setshow_cmd_full.
	(add_setshow_cmd_full): New function.
	(add_setshow_auto_boolean_cmd): Use add_setshow_cmd_full.
	(add_setshow_boolean_cmd): Likewise.

Index: command.h
===================================================================
RCS file: /cvs/src/src/gdb/command.h,v
retrieving revision 1.33
diff -u -r1.33 command.h
--- command.h 15 Jun 2002 22:05:32 -0000 1.33
+++ command.h 26 Jun 2002 17:54:08 -0000
@@ -210,6 +210,26 @@
 extern void help_cmd_list (struct cmd_list_element *, enum command_class,
 			   char *, int, struct ui_file *);
 
+extern void add_setshow_cmd (char *name,
+			     enum command_class class,
+			     var_types var_type, void *var,
+			     char *set_doc, char *show_doc,
+			     cmd_sfunc_ftype *set_func,
+			     cmd_sfunc_ftype *show_func,
+			     struct cmd_list_element **set_list,
+			     struct cmd_list_element **show_list);
+
+extern void add_setshow_cmd_full (char *name,
+				  enum command_class class,
+				  var_types var_type, void *var,
+				  char *set_doc, char *show_doc,
+				  cmd_sfunc_ftype *set_func,
+				  cmd_sfunc_ftype *show_func,
+				  struct cmd_list_element **set_list,
+				  struct cmd_list_element **show_list,
+				  struct cmd_list_element **set_result,
+				  struct cmd_list_element **show_result);
+
 extern struct cmd_list_element *add_set_cmd (char *name, enum
 					     command_class class,
 					     var_types var_type, void *var,
Index: cli/cli-decode.c
===================================================================
RCS file: /cvs/src/src/gdb/cli/cli-decode.c,v
retrieving revision 1.25
diff -u -r1.25 cli-decode.c
--- cli/cli-decode.c 25 Jun 2002 05:39:18 -0000 1.25
+++ cli/cli-decode.c 26 Jun 2002 17:54:11 -0000
@@ -331,14 +331,16 @@
    command.  SET_FUNC and SHOW_FUNC are the callback functions (if
    non-NULL).  SET_DOC and SHOW_DOC are the documentation strings.  */
 
-static struct cmd_list_element *
-add_setshow_cmd (char *name,
-		 enum command_class class,
-		 var_types var_type, void *var,
-		 char *set_doc, char *show_doc,
-		 cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
-		 struct cmd_list_element **set_list,
-		 struct cmd_list_element **show_list)
+void
+add_setshow_cmd_full (char *name,
+		      enum command_class class,
+		      var_types var_type, void *var,
+		      char *set_doc, char *show_doc,
+		      cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
+		      struct cmd_list_element **set_list,
+		      struct cmd_list_element **show_list,
+		      struct cmd_list_element **set_result,
+		      struct cmd_list_element **show_result)
 {
   struct cmd_list_element *set;
   struct cmd_list_element *show;
@@ -350,9 +352,31 @@
 			      show_doc, show_list);
   if (show_func != NULL)
     set_cmd_sfunc (show, show_func);
-  /* The caller often wants to modify set to include info like an
-     enumeration.  */
-  return set;
+
+  if (set_result != NULL)
+    *set_result = set;
+  if (show_result != NULL)
+    *show_result = show;
+}
+
+/* Add element named NAME to both the command SET_LIST and SHOW_LIST.
+   CLASS is as in add_cmd.  VAR_TYPE is the kind of thing we are
+   setting.  VAR is address of the variable being controlled by this
+   command.  SET_FUNC and SHOW_FUNC are the callback functions (if
+   non-NULL).  SET_DOC and SHOW_DOC are the documentation strings.  */
+
+void
+add_setshow_cmd (char *name,
+		 enum command_class class,
+		 var_types var_type, void *var,
+		 char *set_doc, char *show_doc,
+		 cmd_sfunc_ftype *set_func, cmd_sfunc_ftype *show_func,
+		 struct cmd_list_element **set_list,
+		 struct cmd_list_element **show_list)
+{
+  add_setshow_cmd_full (name, class, var_type, var, set_doc, show_doc,
+			set_func, show_func, set_list, show_list,
+			NULL, NULL);
 }
 
 struct cmd_list_element *
@@ -405,9 +429,10 @@
 {
   static const char *auto_boolean_enums[] = { "on", "off", "auto", NULL };
   struct cmd_list_element *c;
-  c = add_setshow_cmd (name, class, var_auto_boolean, var,
-		       set_doc, show_doc, set_func, show_func,
-		       set_list, show_list);
+  add_setshow_cmd_full (name, class, var_auto_boolean, var,
+			set_doc, show_doc, set_func, show_func,
+			set_list, show_list,
+			&c, NULL);
   c->enums = auto_boolean_enums;
 }
 
@@ -426,10 +451,11 @@
 {
   static const char *boolean_enums[] = { "on", "off", NULL };
   struct cmd_list_element *c;
-  c = add_setshow_cmd (name, class, var_boolean, var,
-		       set_doc, show_doc,
-		       set_func, show_func,
-		       set_list, show_list);
+  add_setshow_cmd_full (name, class, var_boolean, var,
+			set_doc, show_doc,
+			set_func, show_func,
+			set_list, show_list,
+			&c, NULL);
   c->enums = boolean_enums;
 }
 

Re: RFA: add_setshow_cmd_full

[Tom Tromey]

>>>>> "Eli" == Eli Zaretskii <eliz@is.elta.co.il> writes:

Eli> Sounds good, but could you please add this function to
Eli> gdbint.texinfo?

Assuming Andrew approves my most recent patch in this area, what do
you think of the appended documentation patch?  I didn't list the
arguments as the other entries don't seem to, and anyway they are
documented in the source.

In Java it is standard to extract special comments in the source and
use those to construct API documentation.  I know a long time ago gdb
did this, but it was abandoned.  Anyway, wouldn't it be simpler to
keep the docs up to date with something like this in place?

Tom

Index: ChangeLog
from  Tom Tromey  <tromey@redhat.com>

	* gdbint.texinfo (User Interface): Mention add_setshow_cmd and
	add_setshow_cmd_full.

Index: gdbint.texinfo
===================================================================
RCS file: /cvs/src/src/gdb/doc/gdbint.texinfo,v
retrieving revision 1.91
diff -u -r1.91 gdbint.texinfo
--- gdbint.texinfo 9 Jun 2002 17:15:40 -0000 1.91
+++ gdbint.texinfo 26 Jun 2002 18:13:36 -0000
@@ -50,7 +50,7 @@
 @end tex
 
 @vskip 0pt plus 1filll
-Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001
+Copyright @copyright{} 1990,1991,1992,1993,1994,1996,1998,1999,2000,2001, 2002
    Free Software Foundation, Inc.
 
 Permission is granted to copy, distribute and/or modify this document
@@ -721,6 +721,14 @@
 the main command list, and should be used for those commands.  The usual
 place to add commands is in the @code{_initialize_@var{xyz}} routines at
 the ends of most source files.
+
+@findex add_setshow_cmd
+@findex add_setshow_cmd_full
+To add paired @samp{set} and @samp{show} commands, use
+@code{add_setshow_cmd} or @code{add_setshow_cmd_full}.  The former is
+a slightly simpler interface which is useful when you don't need to
+further modify the new command structures, while the latter returns
+the new command structures for manipulation.
 
 @cindex deprecating commands
 @findex deprecate_cmd

Re: RFA: add_setshow_cmd_full

[Eli Zaretskii]

> From: Tom Tromey <tromey@redhat.com>
> Date: 26 Jun 2002 12:36:13 -0600
> 
> Assuming Andrew approves my most recent patch in this area, what do
> you think of the appended documentation patch?

It's okay, please commit it with the source patches.

> In Java it is standard to extract special comments in the source and
> use those to construct API documentation.  I know a long time ago gdb
> did this, but it was abandoned.  Anyway, wouldn't it be simpler to
> keep the docs up to date with something like this in place?

I don't mind going back to keeping the docs with the sources.

Re: RFA: add_setshow_cmd_full

[Andrew Cagney]

[heads up, my e-mail address is changing]

> +extern void add_setshow_cmd (char *name,
> +			     enum command_class class,
> +			     var_types var_type, void *var,
> +			     char *set_doc, char *show_doc,
> +			     cmd_sfunc_ftype *set_func,
> +			     cmd_sfunc_ftype *show_func,
> +			     struct cmd_list_element **set_list,
> +			     struct cmd_list_element **show_list);
> +
> +extern void add_setshow_cmd_full (char *name,
> +				  enum command_class class,
> +				  var_types var_type, void *var,
> +				  char *set_doc, char *show_doc,
> +				  cmd_sfunc_ftype *set_func,
> +				  cmd_sfunc_ftype *show_func,
> +				  struct cmd_list_element **set_list,
> +				  struct cmd_list_element **show_list,
> +				  struct cmd_list_element **set_result,
> +				  struct cmd_list_element **show_result);
> +

Yep, fine.

Andrew