FYI: Ubuntu 10.04 lucid strange segfault

Started by Yeb Havingaover 15 years ago5 messages
#1Yeb Havinga
yebhavinga@gmail.com

Hello list,

Due to dependency requirements my development machine has ubuntu
repositories jaunty, lucid and intrepid.

Today I went testing a patch on a recent PostgreSQL. I had to install
autoconf to configure the patched source. While seeing some other
packages getting installed on the screen, I remember thinking 'should've
installed with aptitude to see change set instead of apt-get install'.

I did not get past make install due to a segfault on ./zic in the
src/timezone directory. The strange thing is that another source tree I
was working on was intact and could configure and make install/run zic
without problems. Circumventing the zic segfault showed that after
installation the postgres binary had no problems, but psql had similar
problems as zic. (0x0000000000659bd0 in strncpy@@GLIBC_2.2.5 () where
zic went wrong in strchr).

<long story of dead ends in problem tree traversal removed>

At some point read reading http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43556

Changing versions to gcc 4.4.3-4ubuntu5 and binutils 2.20.1-3ubuntu5
made the strange segfaults go away.

I don't know which versions I had installed when having the segfault, sorry.

regards,
Yeb Havinga

#2Joshua D. Drake
jd@commandprompt.com
In reply to: Yeb Havinga (#1)
Re: FYI: Ubuntu 10.04 lucid strange segfault

On Fri, 2010-07-09 at 16:57 +0200, Yeb Havinga wrote:

Hello list,

Due to dependency requirements my development machine has ubuntu
repositories jaunty, lucid and intrepid.

I would not expect anything to work in an environment that is that
misconfigured. I would suggest running VirtualBox for your different
requirements.

Sincerely,

Joshua D. Drake

#3Yeb Havinga
yebhavinga@gmail.com
In reply to: Joshua D. Drake (#2)
Re: FYI: Ubuntu 10.04 lucid strange segfault

Joshua D. Drake wrote:

On Fri, 2010-07-09 at 16:57 +0200, Yeb Havinga wrote:

Hello list,

Due to dependency requirements my development machine has ubuntu
repositories jaunty, lucid and intrepid.

I would not expect anything to work in an environment that is that
misconfigured. I would suggest running VirtualBox for your different
requirements.

Thanks for your advice but my message was a FYI. The link I provided
suggests it can happen at systems having only lucid as well.

regards,
Yeb Havinga

#4Robert Haas
robertmhaas@gmail.com
In reply to: Yeb Havinga (#3)
Re: FYI: Ubuntu 10.04 lucid strange segfault

On Fri, Jul 9, 2010 at 4:10 PM, Yeb Havinga <yebhavinga@gmail.com> wrote:

Joshua D. Drake wrote:

On Fri, 2010-07-09 at 16:57 +0200, Yeb Havinga wrote:

Hello list,

Due to dependency requirements my development machine has ubuntu
repositories jaunty, lucid and intrepid.

I would not expect anything to work in an environment that is that
misconfigured. I would suggest running VirtualBox for your different
requirements.

Thanks for your advice but my message was a FYI. The link I provided
suggests it can happen at systems having only lucid as well.

It sounds like it's an Ubuntu linker bug, though - not really anything
to do with us.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company

#5Yeb Havinga
yebhavinga@gmail.com
In reply to: Robert Haas (#4)
Re: FYI: Ubuntu 10.04 lucid strange segfault

On Fri, Jul 9, 2010 at 10:18 PM, Robert Haas <robertmhaas@gmail.com> wrote:

On Fri, Jul 9, 2010 at 4:10 PM, Yeb Havinga <yebhavinga@gmail.com> wrote:

Joshua D. Drake wrote:

On Fri, 2010-07-09 at 16:57 +0200, Yeb Havinga wrote:

Hello list,

Due to dependency requirements my development machine has ubuntu
repositories jaunty, lucid and intrepid.

I would not expect anything to work in an environment that is that
misconfigured. I would suggest running VirtualBox for your different
requirements.

Thanks for your advice but my message was a FYI. The link I provided
suggests it can happen at systems having only lucid as well.

It sounds like it's an Ubuntu linker bug, though - not really anything
to do with us.

I realize the pg lists are not the place to discuss particular distribution
bugs. However, because the bug occurred only in a particular pg source tree
and not another, I spent a lot of time looking in the wrong place. This
could happen to somebody else as well. Since searching for some keywords in
the pg lists returned 0 results I thought lets change that, to possibly safe
somebody else some time.

Sorry to cause confusion. I'll try to be more prudent in the future.

regards,
Yeb Havinga