persistent module does not use Py_ssize_t

The following diff (from svn) fixes the bug for me.

HTH,
Gerald

The same problem is also present in the BTrees module. It manifests itself in various ways, like slicing not working correctly.

I have attached a patch for the BTrees module. It needs the "#if PY_VERSION_HEX" like the other patch and typedefs for the other new types (lenfunc, ssizeargfunc, etc.) to work with older python versions. I felt that this code is better put in some common place instead of being copy & pasted into each file (but not sure where that is).

All tests pass with both patches applied.

Shane

Can you please explain a bit more when this happens? Best would be to provide a test case that provokes the error.

The problem here is that existing tests fail, but only on (some) 64-
bit systems. :(

No *new* test case is needed.

Jim

On Nov 9, 2007, at 8:14 AM, Christian Theune wrote:

> Can you please explain a bit more when this happens? Best would be to
> provide a test case that provokes the error.
>
> ** Changed in: zodb
> Importance: Undecided => Medium
> Status: New => Incomplete
>
> --
> persistent module does not use Py_ssize_t
> https://bugs.launchpad.net/bugs/153316
> You received this bug notification because you are a member of Zodb-
> developers, which is the registrant for ZODB.

--
Jim Fulton
Zope Corporation

Am Freitag, den 09.11.2007, 13:28 +0000 schrieb Jim Fulton:
> The problem here is that existing tests fail, but only on (some) 64-
> bit systems. :(

Do we know which? I'll try some of the 64-bit systems later but it would
be good to know where they fail.

--
gocept gmbh & co. kg - forsterstrasse 29 - 06112 halle (saale) - germany
www.gocept.com - <email address hidden> - phone +49 345 122 9889 7 -
fax +49 345 122 9889 1 - zope and plone consulting and development

On Nov 9, 2007, at 9:25 AM, Christian Theune wrote:

>
> Am Freitag, den 09.11.2007, 13:28 +0000 schrieb Jim Fulton:
>> The problem here is that existing tests fail, but only on (some) 64-
>> bit systems. :(
>
> Do we know which? I'll try some of the 64-bit systems later but it
> would
> be good to know where they fail.

No. I asked the original poster on zodb-dev but I don't remember
seeing a reply, although I might have missed it. BTW, I'm assuming
that this is a 64-bit issue because the tests passed for me on the 32-
bit systems I tested against.

Jim

--
Jim Fulton
Zope Corporation

The patches look ok. I'm reorganising the Python 2.4 compatibility (thanks to Gregoire Weber for some C knowledge).

Btw: It's reproducable on this machine for me:

Linux uter 2.6.12-gentoo-r6 #1 SMP Sun Aug 21 19:27:51 CEST 2005 x86_64 Intel(R) Xeon(TM) CPU 2.80GHz GenuineIntel GNU/Linux

Fixed on the trunk using a cleaned-up version of the patches. Passes on 32-bit and 64-bit Xeon.

Fixed in r81758.

Released in 3.8.0c1

I'm seeing something simular to this on Ubuntu 64-bit machines against the current 3.8.9b egg. will drive to derive more a replicable test