Re: GDB 6.1.1 2004-05-21-gmt

Re: GDB 6.1.1 2004-05-21-gmt

[Michael Elizabeth Chastain]

As far as test results go:

i686-pc-linux-gnu:
  gdb_6_1-branch has no improvements or regressions from gdb 6.1

hppa2.0w-hp-hpux11.11:
  I haven't run this for a while so I can't say
  The last time I tried gdb HEAD versus gdb 6.1, though,
    there were some improvements, but also some regressions
  I need to sit down with that ol' hp test drive machine for a few hours

Re: GDB 6.1.1 2004-05-21-gmt

[Michael Elizabeth Chastain]

manoj> I have similar issues with GDB 6.1 wrt ppc64, there are many
manoj> changes in mainline that I would like to have in 6.1, so my question
manoj> is... When will 6.2 be released?

Our estimate is August 2004.

  http://www.gnu.org/software/gdb/schedule/

Of course, that is only an estimate.

Michael C

GDB 6.1.1 2004-05-21-gmt

[Andrew Cagney]

It's time for that re-spin.  It's ment to happen in may so I'd better do 
it before the comming weekend :-/

Randolph, Joel, for HP/PA, should the architecture specific changes 
(i.e., hp*-{tdep,nat}) be pulled into the branch?

Andrew

Re: GDB 6.1.1 2004-05-21-gmt

[Randolph Chung]

In reference to a message from Andrew Cagney, dated May 25:
> It's time for that re-spin.  It's ment to happen in may so I'd better do 
> it before the comming weekend :-/
> 
> Randolph, Joel, for HP/PA, should the architecture specific changes 
> (i.e., hp*-{tdep,nat}) be pulled into the branch?

i'm ok with it either way. i also asked drow about this earlier since
debian already has hppa* latest in the debian gdb package; he's ok with
the current arrangement for now.

i still have a few patches that are not yet approved/committed. i'd like
to at least get the one related to global pointers checked in if we are
going to put hppa support into the respin. i believe the hpux test
results are comparible to what we had in the 6.1 timeframe (maybe a
little bit better). hppa-linux support is of course new.... but it is a
lot better than hpux.

randolph
-- 
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/

Re: GDB 6.1.1 2004-05-21-gmt

[Joel Brobecker]

> Randolph, Joel, for HP/PA, should the architecture specific changes 
> (i.e., hp*-{tdep,nat}) be pulled into the branch?

Thanks to the changes Randolph made, the HP/UX port is in better shape.
I wouldn't argue against pulling the pa-specific files in, except that
I am not sure whether some of them rely on some changes that were made
only in mainline... Randolph?

-- 
Joel

Re: GDB 6.1.1 2004-05-21-gmt

[Randolph Chung]

> Thanks to the changes Randolph made, the HP/UX port is in better shape.
> I wouldn't argue against pulling the pa-specific files in, except that
> I am not sure whether some of them rely on some changes that were made
> only in mainline... Randolph?

there are a few arch-indep changes, most of it are just cleanups. the
other ones that come to mind are new and not yet checked in, so we can
discuss what we want to do with them:

- gdbarch change for passing function object into push_dummy_call
- infrun.c changes for trampoline handling
- use inferior_created event for updating dyncall, etc values

if we want to update the hppa support in 6.1 i'd like to see the first
one (gdbarch) go in.... the other ones can wait for 6.2

randolph
-- 
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/

Re: GDB 6.1.1 2004-05-21-gmt

[Andrew Cagney]

>>Thanks to the changes Randolph made, the HP/UX port is in better shape.
>>> I wouldn't argue against pulling the pa-specific files in, except that
>>> I am not sure whether some of them rely on some changes that were made
>>> only in mainline... Randolph?
> 
> 
> there are a few arch-indep changes, most of it are just cleanups. the
> other ones that come to mind are new and not yet checked in, so we can
> discuss what we want to do with them:
> 
> - gdbarch change for passing function object into push_dummy_call
> - infrun.c changes for trampoline handling
> - use inferior_created event for updating dyncall, etc values
> 
> if we want to update the hppa support in 6.1 i'd like to see the first
> one (gdbarch) go in.... the other ones can wait for 6.2

There seem to be a significant number of fixes even without this - 
basics such as value return and backtraces.

As for the arch and other core changes, they can wait until 6.2.

Andrew

Re: GDB 6.1.1 2004-05-21-gmt

[Randolph Chung]

> There seem to be a significant number of fixes even without this - 
> basics such as value return and backtraces.
> 
> As for the arch and other core changes, they can wait until 6.2.

ok... it's just that the gdbarch change will fix MANY problems in the
testsuite. :)

what is the process for getting patches into 6.1? Are they resubmitted
one-by-one, or do i do a wholesale diff of hppa-specific files? There
are about 30 patches from me that has gone into CVS since 6.1.

thanks
randolph
-- 
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/

Re: GDB 6.1.1 2004-05-21-gmt

[Andrew Cagney]

>>There seem to be a significant number of fixes even without this - 
>>> basics such as value return and backtraces.
>>> 
>>> As for the arch and other core changes, they can wait until 6.2.
> 
> 
> ok... it's just that the gdbarch change will fix MANY problems in the
> testsuite. :)
> 
> what is the process for getting patches into 6.1? Are they resubmitted
> one-by-one, or do i do a wholesale diff of hppa-specific files? There
> are about 30 patches from me that has gone into CVS since 6.1.

Something like:

$ cp src/gdb/hppa*.[hc] 61/src/gdb ; emacs 61/src/gdb/ChangeLog

For the changelog suggest both mentioning that the stuff is an import 
from the mainline and list a concat of the original changelogs (minus 
dates et.al.).

Most important though is:

$ emacs 61/src/gdb/NEWS 61/src/gdb/ChangeLog

where there should be a brief of how the update affects (benefits) the 
user - mention a few key bug fixes for instance.

When it comes to architecture specific changes GDB is very liberal - the 
worst they can do is break that architecture and presumably the person 
pulling in the changes is testing it :-)

enjoy,
Andrew

PS: I've some incompatible frame changes on hold waiting for 6.1.1 to 
ship - help make these imports easier.

Re: GDB 6.1.1 2004-05-21-gmt

[Randolph Chung]

> Something like:
> 
> $ cp src/gdb/hppa*.[hc] 61/src/gdb ; emacs 61/src/gdb/ChangeLog

ok, i'll try to come up with something.

fwiw, the current status of hppa test results, as far i know, is:

hppa-linux: ~70 FAILs, but depends on a few patches that are not yet
approved

hppa2.0w-hp-hpux11.11 with gcc: ~700 FAILs based on CVS HEAD
hppa2.0w-hp-hpux11.00 with HP ANSI C compiler: ~750 FAILs based on CVS
HEAD

hppa64-hp-hpux11.11 is still very broken.

thanks,
randolph
-- 
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/

Re: GDB 6.1.1 2004-05-21-gmt

[Manoj Iyer]

I have similar issues with GDB 6.1 wrt ppc64, there are many changes in
mainline that I would like to have in 6.1, so my question is... When will
6.2 be released?

--
Manoj

On Tue, 25 May 2004, Randolph Chung wrote:

> > There seem to be a significant number of fixes even without this -
> > basics such as value return and backtraces.
> >
> > As for the arch and other core changes, they can wait until 6.2.
>
> ok... it's just that the gdbarch change will fix MANY problems in the
> testsuite. :)
>
> what is the process for getting patches into 6.1? Are they resubmitted
> one-by-one, or do i do a wholesale diff of hppa-specific files? There
> are about 30 patches from me that has gone into CVS since 6.1.
>
> thanks
> randolph
> --
> Randolph Chung
> Debian GNU/Linux Developer, hppa/ia64 ports
> http://www.tausq.org/
>

Re: GDB 6.1.1 2004-05-21-gmt

[Andrew Cagney]

> I have similar issues with GDB 6.1 wrt ppc64, there are many changes in
> mainline that I would like to have in 6.1, so my question is... When will
> 6.2 be released?

Per my earlier post, the current ``in theory'' schedule can be found at 
http://www.gnu.org/software/gdb/schedule/

Andrew

Re: GDB 6.1.1 2004-05-21-gmt

[Andrew Cagney]

Just FYI, GDB's ``schedule'' is found at:

http://www.gnu.org/software/gdb/schedule/

of course it is open to negotiation :-)

Andrew