This is the mail archive of the
mailing list for the Cygwin project.
Re: python aborts
- From: Christopher Faylor <cgf-use-the-mailinglist-please at cygwin dot com>
- To: cygwin at cygwin dot com
- Date: Thu, 25 Jul 2013 11:14:10 -0400
- Subject: Re: python aborts
- References: <d0191e464b44f772f988b2050cda4ac8 at denis-excoffier dot org> <20130522123148 dot GW2406 at calimero dot vinschen dot de> <51F0CBD2 dot 2030502 at dancol dot org> <51F0CF92 dot 60009 at dancol dot org> <51F0E39C dot 1090500 at dancol dot org> <20130725091741 dot GA20682 at calimero dot vinschen dot de> <51F0FB68 dot 90904 at gmail dot com> <51F103CB dot 20904 at gmail dot com>
- Reply-to: cygwin at cygwin dot com
On Thu, Jul 25, 2013 at 06:54:03PM +0800, JonY wrote:
>On 7/25/2013 18:18, JonY wrote:
>> On 7/25/2013 17:17, Corinna Vinschen wrote:
>>> On Jul 25 01:36, Daniel Colascione wrote:
>>>> On 7/25/2013 12:11 AM, Daniel Colascione wrote:
>>>>> On 7/24/2013 11:55 PM, Daniel Colascione wrote:
>>>>>> Does that help at all? I only started seeing this problem after I recompiled
>>>>>> _wp.dll using gcc 4.7.3.
>>>>> Actually, this problem looks a lot like
>>>>> http://email@example.com/msg68316.html: neither Python nor
>>>>> _wp links dynamically to libgcc, but cygsqlite3-0.dll does.
>>>> And this is a very nasty bug; Eli's analysis is correct. Say we have modules Foo
>>>> and Bar. Foo links against shared libgcc, but Bar does not. Now, if we load Foo,
>>>> load Bar, unload Foo, then unload Bar, then Foo's initialization code finds
>>>> libgcc and registers itself with it, but Foo's deinitializaton code doesn't find
>>>> libgcc, tries to instead unregister with Foo's internal data structures, finds
>>>> them uninitialized, and aborts. No wonder changing Python module order around
>>>> makes the problem go away for a little while.
>>>> The right fix for libgcc looks something like this:
>>> Good catch! Any chance you could send this upstream?
>>> JonY, do you have any spare cycles to create new 32 and 64 bit gcc
>>> packages with this fix?
>> Sure, should be done during the weekends, uploads and all. Kai seems to
>> be on holiday, so getting it accepted upstream might take a while.
>Daniel, please apply for FSF copyright assignment if you have not
>already done so, if not, this patch is not going to be accepted upstream.
I'm still a (silent) gcc maintainer so I can presumably accept this patch.
Are there outstanding patches that need attention? If so, please send URLs
from the gcc-patches archive here so that I can attend to them.
Problem reports: http://cygwin.com/problems.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple