RE: debugging gcc 3.x c++

RE: debugging gcc 3.x c++

[Dan Conti]

> -----Original Message-----
> From: Daniel Jacobowitz [mailto:drow@mvista.com]
> Subject: Re: debugging gcc 3.x c++
>
> [snip]
>
> This is fixed in GDB CVS.  It has been for a bit over a month now. 
> Could you give it a try?
> 
> I see on the PR trail that I told you this at the time.  The
> misunderstanding may come from 5.1.1 being released afterwards; that
> does NOT contain the fix.  5.2 will.

Yes, i remember you telling me that at the time. I tried out a snap from
01/22/02 which seemed to fix it but showed other issues on my target.
When 5.1.1 came out, i figured that the existing bugfixes would be
incorporated into that release, and when i ran it initially it didn't
seem to exhibit this problem (although that could easily be attributed
to different test code). Out of curiosity, why wasn't this incorporated
into the 5.1.1 release?

In any case, when is gdb 5.2 scheduled to be released? If it was just
me, i'd use a cvs snapshot, but i have to set this up for a group of
people and i'd rather use something more stable. Alternatively, would it
be possible for me to patch my 5.1.1 with the appropriate fixes?

> Actually, it's for GNATS use only.  GNATS messages get copied there.

Then perhaps the web page should be updated to mention that it's for
GNATS use only? The description indicates that it can be used for
discussing bugs submitted to the database, so one would think that
discussion this bug (which was submitted to the database) should happen
on that list.

Re: debugging gcc 3.x c++

[Daniel Jacobowitz]

On Tue, Feb 12, 2002 at 03:02:55PM -0800, Dan Conti wrote:
> > -----Original Message-----
> > From: Daniel Jacobowitz [mailto:drow@mvista.com]
> > Subject: Re: debugging gcc 3.x c++
> >
> > [snip]
> >
> > This is fixed in GDB CVS.  It has been for a bit over a month now. 
> > Could you give it a try?
> > 
> > I see on the PR trail that I told you this at the time.  The
> > misunderstanding may come from 5.1.1 being released afterwards; that
> > does NOT contain the fix.  5.2 will.
> 
> Yes, i remember you telling me that at the time. I tried out a snap from
> 01/22/02 which seemed to fix it but showed other issues on my target.
> When 5.1.1 came out, i figured that the existing bugfixes would be
> incorporated into that release, and when i ran it initially it didn't
> seem to exhibit this problem (although that could easily be attributed
> to different test code). Out of curiosity, why wasn't this incorporated
> into the 5.1.1 release?

Because the 5.1.1 release was a maintenance release, not off the
development branch.  It mostly incorporated copyright and legal fixes,
and one DWARF-2 feature.  What other issues?

> In any case, when is gdb 5.2 scheduled to be released? If it was just
> me, i'd use a cvs snapshot, but i have to set this up for a group of
> people and i'd rather use something more stable. Alternatively, would it
> be possible for me to patch my 5.1.1 with the appropriate fixes?

In two months, I believe.  It'd be possible, but quite a lot of work.

> > Actually, it's for GNATS use only.  GNATS messages get copied there.
> 
> Then perhaps the web page should be updated to mention that it's for
> GNATS use only? The description indicates that it can be used for
> discussing bugs submitted to the database, so one would think that
> discussion this bug (which was submitted to the database) should happen
> on that list.

Yes, probably.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: debugging gcc 3.x c++

[Daniel Jacobowitz]

On Tue, Feb 12, 2002 at 02:00:12PM -0800, Dan Conti wrote:
> Hi, i had ran into this bug before with gcc 3.0.3 and gdb 5.1, and was
> told that it was properly fixed in the current versions of gdb (to be
> specific, i was told that it was fixed in cvs between the 5.1 and 5.1.1
> release). In the cvs snapshot i checked and the 5.1.1 release, it's
> fixed in some but not all instances.
> 
> Basically when i go to print some of my C++ classes in the debugger, i
> get a message saying "Value can't be converted to integer." There's a PR
> filed on this at:
> 
> http://sources.redhat.com/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&dat
> abase=gdb&pr=243
> 
> Personally i am using a arm-elf cross toolchain, gcc 3.0.3 and gdb 5.1.1
> on a cygwin host.
> 
> I have no problem fixing this myself if someone could point me in the
> right direction, since i'm not really familiar with gdb internals.

This is fixed in GDB CVS.  It has been for a bit over a month now. 
Could you give it a try?

I see on the PR trail that I told you this at the time.  The
misunderstanding may come from 5.1.1 being released afterwards; that
does NOT contain the fix.  5.2 will.

> Also, something i'm not clear on, according to the gdb mailing lists
> page, 'gdb-prs' is a "Mailing list for discussing bugs submitted to the
> bug reporting database", but for some reason i can't post to this list
> about a bug submitted to the bug reporting database. Perhaps the
> description should mention that it's for testers only (i'm assuming this
> is the case?) ?

Actually, it's for GNATS use only.  GNATS messages get copied there.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

debugging gcc 3.x c++

[Dan Conti]

Hi, i had ran into this bug before with gcc 3.0.3 and gdb 5.1, and was
told that it was properly fixed in the current versions of gdb (to be
specific, i was told that it was fixed in cvs between the 5.1 and 5.1.1
release). In the cvs snapshot i checked and the 5.1.1 release, it's
fixed in some but not all instances.

Basically when i go to print some of my C++ classes in the debugger, i
get a message saying "Value can't be converted to integer." There's a PR
filed on this at:

http://sources.redhat.com/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&dat
abase=gdb&pr=243

Personally i am using a arm-elf cross toolchain, gcc 3.0.3 and gdb 5.1.1
on a cygwin host.

I have no problem fixing this myself if someone could point me in the
right direction, since i'm not really familiar with gdb internals.

Also, something i'm not clear on, according to the gdb mailing lists
page, 'gdb-prs' is a "Mailing list for discussing bugs submitted to the
bug reporting database", but for some reason i can't post to this list
about a bug submitted to the bug reporting database. Perhaps the
description should mention that it's for testers only (i'm assuming this
is the case?) ?

--
Dan Conti
danc@fullplaymedia.com