Mirror of the gdb mailing list
 help / color / mirror / Atom feed
From: Paul Koning <paul_koning@dell.com>
To: Joel Brobecker <brobecker@adacore.com>
Cc: <gdb@sourceware.org>
Subject: Re: Undefined symbol while executing Python
Date: Tue, 05 Jul 2011 13:57:00 -0000	[thread overview]
Message-ID: <AD631A14-F043-4190-9B42-8CF067E661A9@dell.com> (raw)
In-Reply-To: <20110702020837.GG2421@adacore.com>


On Jul 1, 2011, at 10:08 PM, Joel Brobecker wrote:

>>> (gdb) python import itertools
>>> Traceback (most recent call last):
>>> File "<string>", line 1, in ?
>>> ImportError: /usr/lib/python2.4/lib-dynload/itertoolsmodule.so: undefined symbol: PyObject_SelfIter
>>> Error while executing Python code.
> [...]
>> More...  The symbol is defined in libpython2.4.so.  The native 7.0.1
>> gdb has that library in its required libraries list.  The cross 7.2
>> gdb I built does not.  That makes sense, it alllows that gdb to be
>> executed on another machine that might not have the same python
>> installed.  
> 
> Is GDB linked against the shared libpython? From what you are saying,
> it seems like it is.  But I know that there are problems when linking
> against the static version of libpython.  I have the following change
> in python-config.py to deal with the same sort of issue after I linked
> GDB with the static libpython.
> 
> +                if platform.system() == 'Linux':
> +                    # Make sure that the loader can resolve symbols from
> +                    # the libpython archive when loading modules implemented
> +                    # as DSOs (Eg: "import time").  This is to work around
> +                    # a side-effect of linking against the static version
> +                    # of libpython.
> +                    libs.insert(0, '-Xlinker -export-dynamic')

Thanks, that cured the problem (once I figured out where to insert it).  It turns out that, while there is a /usr/lib/libpython2.4, the gdb link is done against /usr/lib/python2.4/config/libpython2.4.a.  I'm not sure why.  It's probably actually a good thing because that binary is being built for a group of people, some of whom have a different version of python installed.

Anyway, it would be good if gdb built correctly (i.e., with that switch) out of the box rather than requiring manual non-obvious hacking.

	paul


  reply	other threads:[~2011-07-05 13:57 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-07-01 19:38 Paul Koning
2011-07-01 20:56 ` Paul Koning
2011-07-02  2:08   ` Joel Brobecker
2011-07-05 13:57     ` Paul Koning [this message]
2011-07-05 15:01       ` Joel Brobecker

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=AD631A14-F043-4190-9B42-8CF067E661A9@dell.com \
    --to=paul_koning@dell.com \
    --cc=brobecker@adacore.com \
    --cc=gdb@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox