This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH/RFA] Introduce new $_isvoid() convenience function
- From: Eli Zaretskii <eliz at gnu dot org>
- To: Sergio Durigan Junior <sergiodj at redhat dot com>
- Cc: dje at google dot com, gdb-patches at sourceware dot org
- Date: Fri, 13 Sep 2013 10:29:47 +0300
- Subject: Re: [PATCH/RFA] Introduce new $_isvoid() convenience function
- Authentication-results: sourceware.org; auth=none
- References: <m3a9jh6bs0 dot fsf at redhat dot com> <CADPb22Q-66NK8u3DLTSG0K7QSYNaxVa0SRBQoBCXoLy7OHvZ9w at mail dot gmail dot com> <m3txhp2pes dot fsf at redhat dot com>
- Reply-to: Eli Zaretskii <eliz at gnu dot org>
> From: Sergio Durigan Junior <sergiodj@redhat.com>
> Cc: GDB Patches <gdb-patches@sourceware.org>
> Date: Fri, 13 Sep 2013 03:42:35 -0300
>
> On Thursday, September 12 2013, Doug Evans wrote:
>
> > Hi. Thanks for persevering!
>
> Thank you for reviewing :-).
>
> > While I often follow "it's easier to relax restrictions than impose
> > them after the fact",
> > in this case I wonder if there's any reason to restrict $_isvoid() to
> > convenience variables.
> >
> > Thoughts?
> >
> > For "void foo(){}", is there any real difference to:
> > (gdb) set $foo = foo()
> > (gdb) p $_isvoid($foo)
> > (gdb) p $_isvoid(foo())
> >
> > [Haven't tried it, but I'm guessing removing the internalva4 check in
> > your patch would make that work.]
>
> Yes, I agree. I have changed the code to accept not only convenience
> variables, but other kinds of expressions (it was just a matter of
> removing the check for internalvar, really). BTW, I have reworded the
> help message and the documentation and did s/convenience
> variable/expression/g.
>
> > Also, you've put a fair bit of time into this, and I'm happy with
> > what you've got. And thanks!
> > The high order bit here for me is being able to script $_exitcode + $exitsignal.
> > If we're going with checking for a void value, one could just leave it
> > to be handled in Python.
>
> Yeah, I thought about it too, but then I thought that maybe we shouldn't
> always depend on Python. This was easy to do in C, and has the
> advantage that it will be available everywhere. But I guess you said
> something along those lines below...
>
> > OTOH, this patch is really simple, and IMO it is *nice* to be able to
> > do this from just a plain gdb script,
> > so I'm advocating accepting this patch (modulo maybe handling any void
> > value, not just conv vars).
>
> Nice. I hope you like the new version of the patch.
>
> > btw, I noticed that if I run a program to exit, and then rerun to
> > main, $_exitcode has the last exit code - I would have expected it to
> > get reset to void upon rerun. After all, if I stop at main the first
> > time through, $_exitcode is void.
> > [I'll file a bug, no need to address this in your patch series.]
>
> Yes, this is a bug indeed. Thanks for filing it. I may even have a fix
> hanging around, will look.
>
> --
> Sergio
>
> gdb/
> 2013-09-13 Sergio Durigan Junior <sergiodj@redhat.com>
>
> * NEWS: Mention new convenience function $_isvoid.
> * value.c (isvoid_internal_fn): New function.
> (_initialize_values): Add new convenience function $_isvoid.
>
> gdb/doc/
> 2013-09-13 Sergio Durigan Junior <sergiodj@redhat.com>
>
> * gdb.texinfo (Convenience Functions): Mention new convenience
> function $_isvoid.
>
> gdb/testsuite/
> 2013-09-13 Sergio Durigan Junior <sergiodj@redhat.com>
>
> * gdb.base/gdbvars.c (foo_void): New function.
> (foo_int): Likewise.
> * gdb.base/gdbvars.exp (test_convenience_functions): New
> function. Call it.
>
> diff --git a/gdb/NEWS b/gdb/NEWS
> index ad97f6f..e58d6ce 100644
> --- a/gdb/NEWS
> +++ b/gdb/NEWS
> @@ -3,6 +3,9 @@
>
> *** Changes since GDB 7.6
>
> +* New convenience function "$_isvoid", to check whether an expression
> + is void.
This is OK, but please add a sentence explaining what is a "void
expression". This is not something GDB users will automatically know
or understand.
> +
> * The "maintenance print objfiles" command now takes an optional regexp.
>
> * The "catch syscall" command now works on arm*-linux* targets.
> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
> index b6ba239..70d580f 100644
> --- a/gdb/doc/gdb.texinfo
> +++ b/gdb/doc/gdb.texinfo
> @@ -9829,6 +9829,18 @@ Returns the length of string @var{str}.
>
> @end table
>
> +These functions do not require @value{GDBN} to be configured with
> +@code{Python} support, which means that they are always available.
I'd prefer to describe functions that don't need Python before those
that do, not after.
> +@table @code
> +
> +@item $_isvoid (@var{expr})
> +@findex $_isvoid@r{, convenience function}
> +Return one if the expression @code{expr} is void. Otherwise it
^^^^^^^^^^^
@var{expr}.
And please put here an explanation of what is a void expression,
preferably more detailed than what you will have in NEWS. AFAICS, we
never explain that anywhere in the manual.
And now a question about this feature: what will use of this function
do in versions of GDB before 7.7?
Thanks.