[RFA] Fix bug in add_alias_cmd

[RFA] Fix bug in add_alias_cmd

[Pierre Muller]

  Out of curiosity, I tried to start
gdb with --xdb options to see the special
options that this provides.

  I got this:
$ ./gdb -xdb
../../purecvs/gdb/cli/cli-decode.c:255: internal-error: add_alias_cmd:
Assertion
 `!aliases && !prehook && prehookee && !posthook && ! posthookee' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n) [answered Y; input not from terminal]
../../purecvs/gdb/cli/cli-decode.c:255: internal-error: add_alias_cmd:
Assertion
 `!aliases && !prehook && prehookee && !posthook && ! posthookee' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Create a core file of GDB? (y or n) [answered Y; input not from terminal]
Aborted (core dumped)

  This is due to "bu" being an alias command of "ubreak"
for xdb options, but "ubreak" doesn't seem to exist anywhere
and I found no trace of it anywhere is the sources :(

I fact after looking at the sources, it appears to me
that the assertion itself is wrong.
prehookee should also be NULL, as the other pointers tested.


This small patch fixes it.

Tested on complier farm without regression.

OK to commit?


Pierre Muller
Pascal language support maintainer for GDB

PS: What should we do about 
Could I also remove that line 
 if (xdb_commands)
    {
      add_com_alias ("ba", "break", class_breakpoint, 1);
      add_com_alias ("bu", "ubreak", class_breakpoint, 1);
    }


2009-06-06  Pierre Muller  <muller@ics.u-strasbg.fr>

	* cli/cli-decode.c (add_alias_cmd): Correct assertion.

Index: src/gdb/cli/cli-decode.c
===================================================================
RCS file: /cvs/src/src/gdb/cli/cli-decode.c,v
retrieving revision 1.79
diff -u -p -r1.79 cli-decode.c
--- src/gdb/cli/cli-decode.c	12 May 2009 16:51:13 -0000	1.79
+++ src/gdb/cli/cli-decode.c	6 Jun 2009 11:18:08 -0000
@@ -251,7 +251,7 @@ add_alias_cmd (char *name, char *oldname
 						     &prehook, &prehookee,
 						     &posthook,
&posthookee);
       /* If this happens, it means a programmer error somewhere.  */
-      gdb_assert (!aliases && !prehook && prehookee
+      gdb_assert (!aliases && !prehook && !prehookee
 		  && !posthook && ! posthookee);
       return 0;
     }

Re: [RFA] Fix bug in add_alias_cmd

[Tom Tromey]

>>>>> "Pierre" == Pierre Muller <muller@ics.u-strasbg.fr> writes:

Pierre>   This is due to "bu" being an alias command of "ubreak"
Pierre> for xdb options, but "ubreak" doesn't seem to exist anywhere
Pierre> and I found no trace of it anywhere is the sources :(

Pierre> I fact after looking at the sources, it appears to me
Pierre> that the assertion itself is wrong.
Pierre> prehookee should also be NULL, as the other pointers tested.

Pierre> This small patch fixes it.

Pierre> OK to commit?

Yes, thanks.
IMO, you could have put this patch in under the "obvious" rule.

Pierre> PS: What should we do about 
Pierre> Could I also remove that line 
Pierre>  if (xdb_commands)
Pierre>     {
Pierre>       add_com_alias ("ba", "break", class_breakpoint, 1);
Pierre>       add_com_alias ("bu", "ubreak", class_breakpoint, 1);
Pierre>     }

You can just delete the "bu" line.

If you want you could do some archaeology to see whether ubreak ever
existed and, if so, when and why it was removed.

Tom

RE: [RFA] Fix bug in add_alias_cmd

[Pierre Muller]

> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Tom Tromey
> Envoyé : Monday, June 08, 2009 5:37 PM
> À : Pierre Muller
> Cc : gdb-patches@sourceware.org
> Objet : Re: [RFA] Fix bug in add_alias_cmd
> 
> >>>>> "Pierre" == Pierre Muller <muller@ics.u-strasbg.fr> writes:
> 
> Pierre>   This is due to "bu" being an alias command of "ubreak"
> Pierre> for xdb options, but "ubreak" doesn't seem to exist anywhere
> Pierre> and I found no trace of it anywhere is the sources :(
> 
> Pierre> I fact after looking at the sources, it appears to me
> Pierre> that the assertion itself is wrong.
> Pierre> prehookee should also be NULL, as the other pointers tested.
> 
> Pierre> This small patch fixes it.
> 
> Pierre> OK to commit?
> 
> Yes, thanks.
> IMO, you could have put this patch in under the "obvious" rule.

  Thanks, I committed the patch.
 
> Pierre> PS: What should we do about
> Pierre> Could I also remove that line
> Pierre>  if (xdb_commands)
> Pierre>     {
> Pierre>       add_com_alias ("ba", "break", class_breakpoint, 1);
> Pierre>       add_com_alias ("bu", "ubreak", class_breakpoint, 1);
> Pierre>     }
> 
> You can just delete the "bu" line.
Patch also committed.
 
> If you want you could do some archaeology to see whether ubreak ever
> existed and, if so, when and why it was removed.

  I already did a 
grep -R -n -i ubreak * 
at src/gdb level, it returned only that
single line, where else could I find something?
Google on 'gdb ubreak' didn't find anything interesting either.


Pierre

Re: [RFA] Fix bug in add_alias_cmd

[Tom Tromey]

>>>>> "Pierre" == Pierre Muller <muller@ics.u-strasbg.fr> writes:

Pierre>   I already did a 
Pierre> grep -R -n -i ubreak * 
Pierre> at src/gdb level, it returned only that
Pierre> single line, where else could I find something?
Pierre> Google on 'gdb ubreak' didn't find anything interesting either.

You could try looking to see when the alias was added.  I also look at
the ChangeLogs for things like this, but I didn't see any mention of
"ubreak".

Tom

Re: [RFA] Fix bug in add_alias_cmd

[Daniel Jacobowitz]

On Sat, Jun 06, 2009 at 05:49:53PM +0200, Pierre Muller wrote:
> 
>   Out of curiosity, I tried to start
> gdb with --xdb options to see the special
> options that this provides.
> 
>   I got this:
> $ ./gdb -xdb
> ../../purecvs/gdb/cli/cli-decode.c:255: internal-error: add_alias_cmd:
> Assertion
>  `!aliases && !prehook && prehookee && !posthook && ! posthookee' failed.

Any idea how long this has been broken?

If it's been several releases, it's likely we can remove gdb -xdb.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [RFA] Fix bug in add_alias_cmd

[Pedro Alves]

On Monday 08 June 2009 17:26:09, Daniel Jacobowitz wrote:

> Any idea how long this has been broken?
> 
> If it's been several releases, it's likely we can remove gdb -xdb.

I thought the same.  But, this is post-6.8 breakage.

I also wondered if "ubreak" was a typo for something else,
like "tbreak".  Does anyone actually know what "ub" was
supposed to do in xdb?  Maybe "unbreak" -- remove
breakpoint?  Just curiosity, I've never used xdb, or
known someone that actually used the -xdb mode in gdb.

-- 
Pedro Alves