Обсуждение: CVS - psql segfault

The CVS version of psql was segfaulting on exit yesterday, and today's
version segfaults when the program starts.  Is this happening to anyone
else?

I'm on Red Hat ES4.

My understanding is that the latest code is rarely broken.  Is is not
unusual for a problem like this to persist for a couple days?  Should I
even bother reporting this?  If so, should I use "hackers" or a
different list?

On Fri, Jul 15, 2005 at 03:47:20PM +0000, Matt Miller wrote:
> The CVS version of psql was segfaulting on exit yesterday, and today's
> version segfaults when the program starts.  Is this happening to anyone
> else?

Did you try "make distclean" before rebuilding?

--
Alvaro Herrera (<alvherre[a]alvh.no-ip.org>)
Si no sabes adonde vas, es muy probable que acabes en otra parte.

On Fri, 2005-07-15 at 12:27 -0400, Alvaro Herrera wrote:
> On Fri, Jul 15, 2005 at 03:47:20PM +0000, Matt Miller wrote:
> > The CVS version of psql was segfaulting
>
> Did you try "make distclean" before rebuilding?
>

I had not done "make distclean."  After doing this all is well.

Matt Miller <mattm@epx.com> writes:
> The CVS version of psql was segfaulting on exit yesterday, and today's
> version segfaults when the program starts.  Is this happening to anyone
> else?

The only work done recently in psql is Bruce's stuff for locale-specific
numeric formatting, so I'm betting that's at fault.  What are your
locale settings?

> My understanding is that the latest code is rarely broken.  Is is not
> unusual for a problem like this to persist for a couple days?  Should I
> even bother reporting this?  If so, should I use "hackers" or a
> different list?

It's worth mentioning, but hackers is definitely the place; development
tip issues are off-topic for -general or -bugs.

            regards, tom lane

On Fri, 2005-07-15 at 13:22 -0400, Tom Lane wrote:
> Matt Miller <mattm@epx.com> writes:
> > The CVS version of psql was segfaulting
>
>  What are your locale settings?

[mmiller@dbamm01-linux ~]$ locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=

However, my problem was solved doing a "make distclean" and then
rebuilding.  Does this mean that there is a dependency missing from a
makefile somewhere?  Or, as a matter of policy, should I just always
clean after updating?

On Fri, Jul 15, 2005 at 05:30:52PM +0000, Matt Miller wrote:

> However, my problem was solved doing a "make distclean" and then
> rebuilding.  Does this mean that there is a dependency missing from a
> makefile somewhere?  Or, as a matter of policy, should I just always
> clean after updating?

You should be using --enable-depend in configure if you want GCC
to track dependencies; they are not tracked manually in the makefiles.

I do that and I very rarely have problems of that nature.

--
Alvaro Herrera (<alvherre[a]alvh.no-ip.org>)
"La rebeldía es la virtud original del hombre" (Arthur Schopenhauer)