Bug#562159: swig change affects python-opensync
On Wed, 2010-05-26 at 13:56 +0200, Michael Banck wrote: > Hi, > > On Tue, May 25, 2010 at 09:41:44AM +0100, Mark Ellis wrote: > > On Tue, 2010-05-25 at 10:34 +0200, Michael Banck wrote: > > > Ah, ok. I am afraid 0.22 will be removed from testing soonish, though. > > > It is basically unmaintained upstream for years now and maintaining it > > > over the squeeze lifetime seems unrealistic. > > > > > > > Ouch, this is very very very bad news. I know 0.22 is basically > > unmaintained, but from my experience with 0.39, this will mean that > > syncing with synce (Windows Mobile) will be impossible in testing. Apart > > from the general problems with the python wrapper, 0.39 fails with > > plugins that don't require a configuration, and there is a bizarre > > g_slice error that I have yet to identify. > > First off, I think svn HEAD has seen some python fixes, so it might be > worthwhile to upload a snapshot (because I don't see another development > release coming up soon). However, that will not magically fix all the > issues in 0.3x. Yeah, that was me :) I'll take this up on the opensync lists soon, I'm trying to make a concerted effort to get it to work, but not really getting anywhere. It's not really relevant here though. > > The biggest problem with 0.22 is that its syncml plugin no longer works > with current (as in, last couple of years) libsyncml. > > Maybe a solution would be to drop most plugins, including the syncml > plugin, and focus on the python bindings, i.e. the synce and barry > plugins mostly (I don't think the moto plugin will be used much or is > remotely usable). > > In that case, we'd still need somebody to fix up issues with 0.22 and > python at least, if not in the 0.22 core. If you can commit to that > (and/or maybe we find somebody else as well), I can try to get this > approved by the release team and do the necessary testing uploads. > > > Michael Points all taken, I agree it's best to move on, even if it means moving the synce plugin away from python. Would it be possible to get this patch applied to testing until 0.22 is dropped ? It would just save me explaining over and again why it doesn't work and how to fix it... Mark signature.asc Description: This is a digitally signed message part
Bug#562159: swig change affects python-opensync
Hi, On Tue, May 25, 2010 at 09:41:44AM +0100, Mark Ellis wrote: > On Tue, 2010-05-25 at 10:34 +0200, Michael Banck wrote: > > Ah, ok. I am afraid 0.22 will be removed from testing soonish, though. > > It is basically unmaintained upstream for years now and maintaining it > > over the squeeze lifetime seems unrealistic. > > > > Ouch, this is very very very bad news. I know 0.22 is basically > unmaintained, but from my experience with 0.39, this will mean that > syncing with synce (Windows Mobile) will be impossible in testing. Apart > from the general problems with the python wrapper, 0.39 fails with > plugins that don't require a configuration, and there is a bizarre > g_slice error that I have yet to identify. First off, I think svn HEAD has seen some python fixes, so it might be worthwhile to upload a snapshot (because I don't see another development release coming up soon). However, that will not magically fix all the issues in 0.3x. The biggest problem with 0.22 is that its syncml plugin no longer works with current (as in, last couple of years) libsyncml. Maybe a solution would be to drop most plugins, including the syncml plugin, and focus on the python bindings, i.e. the synce and barry plugins mostly (I don't think the moto plugin will be used much or is remotely usable). In that case, we'd still need somebody to fix up issues with 0.22 and python at least, if not in the 0.22 core. If you can commit to that (and/or maybe we find somebody else as well), I can try to get this approved by the release team and do the necessary testing uploads. Michael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#562159: swig change affects python-opensync
On Tue, 2010-05-25 at 10:34 +0200, Michael Banck wrote: > Hi, > > On Tue, May 25, 2010 at 09:17:35AM +0100, Mark Ellis wrote: > > On Mon, 2010-05-24 at 22:08 +0200, Michael Banck wrote: > > > tags 562159 -patch > > > thanks > > > > > > Hi, > > > > > > On Mon, May 24, 2010 at 03:46:26PM +0100, Mark Ellis wrote: > > > > This is caused by a change in swig from 1.3.37. The attached patch fixes > > > > this at source level, unlike the other patches mentioned. Please apply > > > > this, python-opensync is completely useless without it. > > > > > > That patch is for 0.22, unstable has 0.39 and it does not seem to apply > > > to that version. > > > > > > > > > Michael > > > > Sorry, did I attach it to the wrong section ? Yes, the patch is for > > 0.22, and needs to be applied to squeeze/testing. > > Ah, ok. I am afraid 0.22 will be removed from testing soonish, though. > It is basically unmaintained upstream for years now and maintaining it > over the squeeze lifetime seems unrealistic. > Ouch, this is very very very bad news. I know 0.22 is basically unmaintained, but from my experience with 0.39, this will mean that syncing with synce (Windows Mobile) will be impossible in testing. Apart from the general problems with the python wrapper, 0.39 fails with plugins that don't require a configuration, and there is a bizarre g_slice error that I have yet to identify. > > The python wrapper of 0.39 in unstable is a completely different story, > > and I don't believe has any chance of working properly at the moment. > > Yeah, possibly. > > > Michael signature.asc Description: This is a digitally signed message part
Bug#562159: swig change affects python-opensync
Hi, On Tue, May 25, 2010 at 09:17:35AM +0100, Mark Ellis wrote: > On Mon, 2010-05-24 at 22:08 +0200, Michael Banck wrote: > > tags 562159 -patch > > thanks > > > > Hi, > > > > On Mon, May 24, 2010 at 03:46:26PM +0100, Mark Ellis wrote: > > > This is caused by a change in swig from 1.3.37. The attached patch fixes > > > this at source level, unlike the other patches mentioned. Please apply > > > this, python-opensync is completely useless without it. > > > > That patch is for 0.22, unstable has 0.39 and it does not seem to apply > > to that version. > > > > > > Michael > > Sorry, did I attach it to the wrong section ? Yes, the patch is for > 0.22, and needs to be applied to squeeze/testing. Ah, ok. I am afraid 0.22 will be removed from testing soonish, though. It is basically unmaintained upstream for years now and maintaining it over the squeeze lifetime seems unrealistic. > The python wrapper of 0.39 in unstable is a completely different story, > and I don't believe has any chance of working properly at the moment. Yeah, possibly. Michael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#562159: swig change affects python-opensync
On Mon, 2010-05-24 at 22:08 +0200, Michael Banck wrote: > tags 562159 -patch > thanks > > Hi, > > On Mon, May 24, 2010 at 03:46:26PM +0100, Mark Ellis wrote: > > This is caused by a change in swig from 1.3.37. The attached patch fixes > > this at source level, unlike the other patches mentioned. Please apply > > this, python-opensync is completely useless without it. > > That patch is for 0.22, unstable has 0.39 and it does not seem to apply > to that version. > > > Michael Sorry, did I attach it to the wrong section ? Yes, the patch is for 0.22, and needs to be applied to squeeze/testing. The python wrapper of 0.39 in unstable is a completely different story, and I don't believe has any chance of working properly at the moment. Mark signature.asc Description: This is a digitally signed message part
Bug#562159: swig change affects python-opensync
tags 562159 -patch thanks Hi, On Mon, May 24, 2010 at 03:46:26PM +0100, Mark Ellis wrote: > This is caused by a change in swig from 1.3.37. The attached patch fixes > this at source level, unlike the other patches mentioned. Please apply > this, python-opensync is completely useless without it. That patch is for 0.22, unstable has 0.39 and it does not seem to apply to that version. Michael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#562159: swig change affects python-opensync
This is caused by a change in swig from 1.3.37. The attached patch fixes
this at source level, unlike the other patches mentioned. Please apply
this, python-opensync is completely useless without it.
Mark
Description: Fix for change in SWIG after 1.3.36
Upstream 0.2x branch is essentially unmaintained.
Forwarded: no
Author: Mark Ellis
Last-Update: 2010-05-24
diff -Nurp opensync-0.22.orig/wrapper/opensync.i opensync-0.22/wrapper/opensync.i
--- opensync-0.22.orig/wrapper/opensync.i 2007-03-27 12:49:09.0 +0100
+++ opensync-0.22/wrapper/opensync.i 2010-05-21 18:36:46.857447188 +0100
@@ -82,7 +82,7 @@ typedef struct {} OSyncHashTable;
%extend OSyncChange {
OSyncChange(PyObject *obj=NULL) {
OSyncChange *change = NULL;
- if (obj)
+ if ((obj) && (obj != Py_None))
change = (OSyncChange *)PyCObject_AsVoidPtr(obj);
else
change = osync_change_new();
signature.asc
Description: This is a digitally signed message part
