[RFA] target record_core -> target record-core?

[RFA] target record_core -> target record-core?

[Doug Evans]

I realize this is a nit but is it possible to rename
target record_core to target record-core?
Or is it too late ...

All other targets (that I know of) use dashes in shortname, and having
to remember which use dashes and which use underscores is, umm, blechy.
[Unless we have and enforce a convention, it can only get worse,
so might as well establish one now - if it doesn't exist already.]

2010-04-18  Doug Evans  <dje@sebabeach.org>

	* record.c (init_record_core_ops): Rename record_core to record_core.

Index: record.c
===================================================================
RCS file: /cvs/src/src/gdb/record.c,v
retrieving revision 1.44
diff -u -p -r1.44 record.c
--- record.c	8 Mar 2010 13:34:38 -0000	1.44
+++ record.c	19 Apr 2010 06:44:51 -0000
@@ -1863,7 +1863,7 @@ record_core_has_execution (struct target
 static void
 init_record_core_ops (void)
 {
-  record_core_ops.to_shortname = "record_core";
+  record_core_ops.to_shortname = "record-core";
   record_core_ops.to_longname = "Process record and replay target";
   record_core_ops.to_doc =
     "Log program while executing and replay execution from log.";

Re: [RFA] target record_core -> target record-core?

[Michael Snyder]

Doug Evans wrote:
> I realize this is a nit but is it possible to rename
> target record_core to target record-core?
> Or is it too late ...
> 
> All other targets (that I know of) use dashes in shortname, and having
> to remember which use dashes and which use underscores is, umm, blechy.
> [Unless we have and enforce a convention, it can only get worse,
> so might as well establish one now - if it doesn't exist already.]

I don't think anybody uses it, so this should be ok.

Hui?

> 
> 2010-04-18  Doug Evans  <dje@sebabeach.org>
> 
> 	* record.c (init_record_core_ops): Rename record_core to record_core.
> 
> Index: record.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/record.c,v
> retrieving revision 1.44
> diff -u -p -r1.44 record.c
> --- record.c	8 Mar 2010 13:34:38 -0000	1.44
> +++ record.c	19 Apr 2010 06:44:51 -0000
> @@ -1863,7 +1863,7 @@ record_core_has_execution (struct target
>  static void
>  init_record_core_ops (void)
>  {
> -  record_core_ops.to_shortname = "record_core";
> +  record_core_ops.to_shortname = "record-core";
>    record_core_ops.to_longname = "Process record and replay target";
>    record_core_ops.to_doc =
>      "Log program while executing and replay execution from log.";

Re: [RFA] target record_core -> target record-core?

[Pedro Alves]

FWIW, I agree with this,

On Monday 19 April 2010 07:47:17, Doug Evans wrote:
> 2010-04-18  Doug Evans  <dje@sebabeach.org>
> 
>         * record.c (init_record_core_ops): Rename record_core to record_core.

though the log entry mentions a nop.  :-)

-- 
Pedro Alves

Re: [RFA] target record_core -> target record-core?

[Hui Zhu]

On Tue, Apr 20, 2010 at 17:56, Pedro Alves <pedro@codesourcery.com> wrote:
>
> FWIW, I agree with this,
>
> On Monday 19 April 2010 07:47:17, Doug Evans wrote:
> > 2010-04-18  Doug Evans  <dje@sebabeach.org>
> >
> >         * record.c (init_record_core_ops): Rename record_core to record_core.
>
> though the log entry mentions a nop.  :-)
>
> --
> Pedro Alves

Thanks guys,  If it doesn't affect the testsuite (Sorry I didn't have
gdb close to me), I agree with this patch.

Regards,
Hui

Re: [RFA] target record_core -> target record-core?

[Tom Tromey]

>>>>> "Doug" == Doug Evans <dje@sebabeach.org> writes:

Doug> I realize this is a nit but is it possible to rename
Doug> target record_core to target record-core?
Doug> Or is it too late ...

Thanks for catching this.

Doug> All other targets (that I know of) use dashes in shortname, and having
Doug> to remember which use dashes and which use underscores is, umm, blechy.

Additionally, I prefer dashes because they are simpler to type.

I think we can retroactively fix mistakes here by adding deprecated
aliases.

One more thing I've wanted is also making it so deprecated commands
don't show up in completion when there is a non-deprecated completion
available.  Any comments on that?

Tom

Re: [RFA] target record_core -> target record-core?

[Joel Brobecker]

> One more thing I've wanted is also making it so deprecated commands
> don't show up in completion when there is a non-deprecated completion
> available.  Any comments on that?

I think that would be useful indeed.

-- 
Joel

Patch: completion -vs- deprecated commands (Was: [RFA] target record_core -> target record-core?)

[Tom Tromey]

Tom> One more thing I've wanted is also making it so deprecated commands
Tom> don't show up in completion when there is a non-deprecated completion
Tom> available.  Any comments on that?

Joel> I think that would be useful indeed.

Here's the patch.  It looks big because I had to reindent a loop.

I plan to commit this early next week, barring objections or comments.

Built and regtested on x86-64 (compile farm).
New tests included.

Tom

2010-04-23  Tom Tromey  <tromey@redhat.com>

	* cli/cli-decode.c (complete_on_cmdlist): Make two passes over the
	command list.

2010-04-23  Tom Tromey  <tromey@redhat.com>

	* gdb.base/completion.exp: Add tests for completion and deprecated
	commands.

Index: cli/cli-decode.c
===================================================================
RCS file: /cvs/src/src/gdb/cli/cli-decode.c,v
retrieving revision 1.86
diff -u -r1.86 cli-decode.c
--- cli/cli-decode.c	1 Jan 2010 07:31:46 -0000	1.86
+++ cli/cli-decode.c	23 Apr 2010 20:17:41 -0000
@@ -1597,43 +1597,65 @@
   int sizeof_matchlist;
   int matches;
   int textlen = strlen (text);
+  int pass;
+  int saw_deprecated_match = 0;
 
   sizeof_matchlist = 10;
   matchlist = (char **) xmalloc (sizeof_matchlist * sizeof (char *));
   matches = 0;
 
-  for (ptr = list; ptr; ptr = ptr->next)
-    if (!strncmp (ptr->name, text, textlen)
-	&& !ptr->abbrev_flag
-	&& (ptr->func
-	    || ptr->prefixlist))
-      {
-	if (matches == sizeof_matchlist)
-	  {
-	    sizeof_matchlist *= 2;
-	    matchlist = (char **) xrealloc ((char *) matchlist,
-					    (sizeof_matchlist
-					     * sizeof (char *)));
-	  }
-
-	matchlist[matches] = (char *)
-	  xmalloc (strlen (word) + strlen (ptr->name) + 1);
-	if (word == text)
-	  strcpy (matchlist[matches], ptr->name);
-	else if (word > text)
-	  {
-	    /* Return some portion of ptr->name.  */
-	    strcpy (matchlist[matches], ptr->name + (word - text));
-	  }
-	else
-	  {
-	    /* Return some of text plus ptr->name.  */
-	    strncpy (matchlist[matches], word, text - word);
-	    matchlist[matches][text - word] = '\0';
-	    strcat (matchlist[matches], ptr->name);
-	  }
-	++matches;
-      }
+  /* We do one or two passes.  In the first pass, we skip deprecated
+     commands.  If we see no matching commands in the first pass, and
+     if we did happen to see a matching deprecated command, we do
+     another loop to collect those.  */
+  for (pass = 0; matches == 0 && pass < 2; ++pass)
+    {
+      for (ptr = list; ptr; ptr = ptr->next)
+	if (!strncmp (ptr->name, text, textlen)
+	    && !ptr->abbrev_flag
+	    && (ptr->func
+		|| ptr->prefixlist))
+	  {
+	    if (pass == 0)
+	      {
+		if ((ptr->flags & CMD_DEPRECATED) != 0)
+		  {
+		    saw_deprecated_match = 1;
+		    continue;
+		  }
+	      }
+
+	    if (matches == sizeof_matchlist)
+	      {
+		sizeof_matchlist *= 2;
+		matchlist = (char **) xrealloc ((char *) matchlist,
+						(sizeof_matchlist
+						 * sizeof (char *)));
+	      }
+
+	    matchlist[matches] = (char *)
+	      xmalloc (strlen (word) + strlen (ptr->name) + 1);
+	    if (word == text)
+	      strcpy (matchlist[matches], ptr->name);
+	    else if (word > text)
+	      {
+		/* Return some portion of ptr->name.  */
+		strcpy (matchlist[matches], ptr->name + (word - text));
+	      }
+	    else
+	      {
+		/* Return some of text plus ptr->name.  */
+		strncpy (matchlist[matches], word, text - word);
+		matchlist[matches][text - word] = '\0';
+		strcat (matchlist[matches], ptr->name);
+	      }
+	    ++matches;
+	  }
+      /* If we saw no matching deprecated commands in the first pass,
+	 just bail out.  */
+      if (!saw_deprecated_match)
+	break;
+    }
 
   if (matches == 0)
     {
Index: testsuite/gdb.base/completion.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/completion.exp,v
retrieving revision 1.44
diff -u -r1.44 completion.exp
--- testsuite/gdb.base/completion.exp	4 Apr 2010 23:47:16 -0000	1.44
+++ testsuite/gdb.base/completion.exp	23 Apr 2010 20:17:43 -0000
@@ -847,6 +847,12 @@
         eof             { fail "(eof) Completing non-existing component #2" }
         }
 
+# If there is a non-deprecated completion, it should be returned.
+gdb_test "complete sav" "save" "test non-deprecated completion"
+# If there is only a deprecated completion, then it should be returned.
+gdb_test "complete save-t" "save-tracepoints" "test deprecated completion"
+
+
 # Restore globals modified in this test...
 set timeout $oldtimeout1
 

Re: Patch: completion -vs- deprecated commands

[Tom Tromey]

>>>>> "Tom" == Tom Tromey <tromey@redhat.com> writes:

Tom> Here's the patch.  It looks big because I had to reindent a loop.

Tom> I plan to commit this early next week, barring objections or comments.

I've checked this in.

Tom