Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Michael Elizabeth Chastain]

I think it's good to send the patch to gdb-patches as usual,
and then mail to gdb-gnats with a URL that points to the patch.
That means you have to for the patch to show up in the gdb-patches
archive, but that takes just a few seconds.

Michael C

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Christopher Faylor]

On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
>I think it's good to send the patch to gdb-patches as usual,
>and then mail to gdb-gnats with a URL that points to the patch.
>That means you have to for the patch to show up in the gdb-patches
>archive, but that takes just a few seconds.

Just as an aside, I'm wondering if it is time to consider switching
to bugzilla.  I think bugzilla handles this issue a little better.

cgf

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Jim Blandy]

Christopher Faylor <cgf@redhat.com> writes:

> On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
> >I think it's good to send the patch to gdb-patches as usual,
> >and then mail to gdb-gnats with a URL that points to the patch.
> >That means you have to for the patch to show up in the gdb-patches
> >archive, but that takes just a few seconds.
> 
> Just as an aside, I'm wondering if it is time to consider switching
> to bugzilla.  I think bugzilla handles this issue a little better.

Does bugzilla allow you to CC the bug on E-mail messages, the way
GNATS does?  I love that, and I haven't seen bugzilla do it.

bugzilla (was Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3)

[Christopher Faylor]

On Fri, Sep 19, 2003 at 07:01:44PM -0500, Jim Blandy wrote:
>Christopher Faylor <cgf@redhat.com> writes:
>
>> On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
>> >I think it's good to send the patch to gdb-patches as usual,
>> >and then mail to gdb-gnats with a URL that points to the patch.
>> >That means you have to for the patch to show up in the gdb-patches
>> >archive, but that takes just a few seconds.
>> 
>> Just as an aside, I'm wondering if it is time to consider switching
>> to bugzilla.  I think bugzilla handles this issue a little better.
>
>Does bugzilla allow you to CC the bug on E-mail messages, the way
>GNATS does?  I love that, and I haven't seen bugzilla do it.

You mean can you send email to bugzilla and have it file the bug?
I think Daniel Berlin has that set up for gcc, yes.

cgf

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Daniel Berlin]

On Fri, 19 Sep 2003, Jim Blandy wrote:

> Christopher Faylor <cgf@redhat.com> writes:
>
> > On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
> > >I think it's good to send the patch to gdb-patches as usual,
> > >and then mail to gdb-gnats with a URL that points to the patch.
> > >That means you have to for the patch to show up in the gdb-patches
> > >archive, but that takes just a few seconds.
> >
> > Just as an aside, I'm wondering if it is time to consider switching
> > to bugzilla.  I think bugzilla handles this issue a little better.
>
> Does bugzilla allow you to CC the bug on E-mail messages, the way
> GNATS does?  I love that, and I haven't seen bugzilla do it.

Uhh, of course.
It's even much simpler than the gnats handling code, because it's a simple
database operation.

The bulk of the code (a couple hundred lines of perl) is devoted to
parsing out and handling attachments in the email (while a perl module
handles most of the details, we still have to do stuff with the data
itself, make up a filename for stupid mailers, etc), as well as verifying
permissions and issuing new accounts for new users.

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Daniel Jacobowitz]

On Fri, Sep 19, 2003 at 07:58:18PM -0400, Christopher Faylor wrote:
> On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
> >I think it's good to send the patch to gdb-patches as usual,
> >and then mail to gdb-gnats with a URL that points to the patch.
> >That means you have to for the patch to show up in the gdb-patches
> >archive, but that takes just a few seconds.
> 
> Just as an aside, I'm wondering if it is time to consider switching
> to bugzilla.  I think bugzilla handles this issue a little better.

I'd say, absolutely.  I've been meaning to ask Daniel B. about setting
one up for binutils too.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Daniel Berlin]

On Sat, 20 Sep 2003, Daniel Jacobowitz wrote:

> On Fri, Sep 19, 2003 at 07:58:18PM -0400, Christopher Faylor wrote:
> > On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
> > >I think it's good to send the patch to gdb-patches as usual,
> > >and then mail to gdb-gnats with a URL that points to the patch.
> > >That means you have to for the patch to show up in the gdb-patches
> > >archive, but that takes just a few seconds.
> >
> > Just as an aside, I'm wondering if it is time to consider switching
> > to bugzilla.  I think bugzilla handles this issue a little better.
>
> I'd say, absolutely.  I've been meaning to ask Daniel B. about setting
> one up for binutils too.

binutils is easier than gdb, because binutils has nothing right now.
Thus, there's nothing to convert :P.

GDB has GNATS, and in order to keep the bug numbers the same, we'd have
to set up a new install of bugzilla (since bugzilla bug ids are
"globally" unique to a given bugzilla instance, rather than unique to a
given product).


Does GDB have any different gnats fields than gcc's gnats used to?

If not, i can run the conversion script on the database, import the data
into my test database on dberlin.org, and you guys could see what the
converted db looks like (though i don't have inbound email handling set up
on my machine).
 --Dan

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Daniel Jacobowitz]

On Sun, Sep 21, 2003 at 11:05:05PM -0400, Daniel Berlin wrote:
> 
> 
> On Sat, 20 Sep 2003, Daniel Jacobowitz wrote:
> 
> > On Fri, Sep 19, 2003 at 07:58:18PM -0400, Christopher Faylor wrote:
> > > On Fri, Sep 19, 2003 at 06:29:09PM -0400, Michael Elizabeth Chastain wrote:
> > > >I think it's good to send the patch to gdb-patches as usual,
> > > >and then mail to gdb-gnats with a URL that points to the patch.
> > > >That means you have to for the patch to show up in the gdb-patches
> > > >archive, but that takes just a few seconds.
> > >
> > > Just as an aside, I'm wondering if it is time to consider switching
> > > to bugzilla.  I think bugzilla handles this issue a little better.
> >
> > I'd say, absolutely.  I've been meaning to ask Daniel B. about setting
> > one up for binutils too.
> 
> binutils is easier than gdb, because binutils has nothing right now.
> Thus, there's nothing to convert :P.
> 
> GDB has GNATS, and in order to keep the bug numbers the same, we'd have
> to set up a new install of bugzilla (since bugzilla bug ids are
> "globally" unique to a given bugzilla instance, rather than unique to a
> given product).

Yeah, at this point I guess we'd prefer to hold on to the PR ids. 
We've documented some of them...

> Does GDB have any different gnats fields than gcc's gnats used to?
> 
> If not, i can run the conversion script on the database, import the data
> into my test database on dberlin.org, and you guys could see what the
> converted db looks like (though i don't have inbound email handling set up
> on my machine).

I don't think it does.  Different things in the category lists et
cetera, of course.  But that's it.  Right now it's:

Reporter, CC, Number, Category, Synopsis, Confidential, Severity,
Priority, Responsible, State, Class, Submitter-Id, Arrival-Date,
Closed-Date, Last-Modified, Originator, Release, Organization,
Environment, Description, File Attachments, How-To-Repeat, Fix,
Release-Note, Unformatted, and the Audit Trail.

Usual caveats about how much some of those are worth.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Michael Elizabeth Chastain]

dberlin> GDB has GNATS, and in order to keep the bug numbers the same,
dberlin> we'd have to set up a new install of bugzilla (since bugzilla
dberlin> bug ids are "globally" unique to a given bugzilla instance,
dberlin> rather than unique to a given product).

I can work with all the bug numbers changing, as long as you
publish a table showing the conversion.  Also it would be
more convenient for me to do it a time when there is no live
gdb branch (so I don't have to change KFAIL numbers on the
branch as well as the mainline).

dberlin> Does GDB have any different gnats fields than gcc's gnats used to?

Dunno, off hand.  There are some new fields that I desperately
want, specifically "compiler version used to compile the program
the inferior program".

Michael C

Re: macros/726: Internal GDB errors with current GDB snapshots and -gdwarf2-3

[Jim Blandy]

This points out an interesting issue.  I can attach a patch to a PR in
two ways, neither of which are really satisfactory:

- I can attach the patch directly to the PR, which causes GNATS to
  generate E-mail messages that refer to "attachments" that are not
  present in the message.

- Or, if I'm replying to a PRMS message, I can attach the patch to the
  message, which results in the patch not appearing as an attachment
  to the PR itself.

Hmph.  I could always just do both, I suppose.