On Sat, Jul 27, 2024 at 09:23:54PM +0200, Daniel Sahlberg wrote:
> Do you a specific preference regarding the change (does it affect your
> ability to continue providing the Subversion package, do you have any
> timing requirements etc.)?
>
> Do you have experience from other projects switching to
We have been seeing intermittent test failures in the Fedora Subversion
build for a while. Here is an example build failure:
https://kojipkgs.fedoraproject.org/work/tasks/5854/89115854/build.log
I think this is caused by libmagic getting applied to small files. I am
not 100% confident in this
On Wed, Mar 02, 2022 at 08:28:51PM -0500, James McCoy wrote:
> On Wed, Mar 02, 2022 at 03:47:23PM +, Danilo Spinella wrote:
> > Hi,
> >
> > I am the openSUSE maintainer of subversion package. With the latest
> > version, the automated build of subversion 1.14.1 for Tumbleweed is
> > failing du
On Wed, Feb 12, 2020 at 02:28:54PM +, Daniel Shahaf wrote:
> jor...@apache.org wrote on Wed, 12 Feb 2020 13:41 -:
> > Author: jorton
> > Date: Wed Feb 12 13:41:25 2020
> > New Revision: 1873943
> >
> > URL: http://svn.apache.org/viewvc?rev=1873943&view=rev
> > Log:
> > Fix test failures se
On Fri, Nov 02, 2018 at 10:35:40AM +0100, Branko Čibej wrote:
> On 02.11.2018 10:23, jor...@apache.org wrote:
> > Author: jorton
> > Date: Fri Nov 2 09:23:25 2018
> > New Revision: 184
> >
> > URL: http://svn.apache.org/viewvc?rev=184&view=rev
> > Log:
> > * configure.ac: Use $PYTHON (alwa
On Fri, Nov 02, 2018 at 10:44:33AM +0100, Stefan Sperling wrote:
> I confirm that your patch is correct. I mis-indented these blocks.
Great, thank you for checking, Stefan. Committed in r1845559.
There are two gcc 8.x (I'm using 8.2.1) warnings from which catch those:
if (foo)
bar;
baz;
type of errors. I fixed one case which looks obviously like a false
positive in r1845556, but I'm not sure about the other one, in this
code:
http://svn.apache.org/viewvc/subversion/trunk/su
On Tue, May 08, 2018 at 12:31:55PM +0200, Branko Čibej wrote:
> On 08.05.2018 12:19, Joe Orton wrote:
> > [[[
> > * subversion/tests/cmdline/svntest/main.py
> > (open_pipe): Run Python tests under the same Python executable for
> > all platforms rather than just Win
As part of the Python 2->3 migration effort in Fedora we've been trying
to get to the point where setting PYTHON=/usr/bin/python3 allows us to
ensure that /usr/bin/python is never used in the build environment.
One thing that doesn't work right with svn is that the test scripts are
run by execu
On Thu, May 03, 2018 at 08:32:56PM +0100, Philip Martin wrote:
> Your patch is an improvement so I would be happy to see it on
> trunk/1.10. Going further and removing some .pc files is probably also
> suitable for 1.10. Changing the DSO names and installation location is
> probably not suitable
On Wed, May 02, 2018 at 04:32:26PM +0100, Philip Martin wrote:
> Philip Martin writes:
>
> >> Fixes libsvn_auth_gnome_keyring.pc when configured with libsecret, and
> >> make dependencies private.
> >
> > Looks OK to me.
>
> Perhaps we should delete the .pc files for gnome/kwallet auth libraries
libsvn_auth_gnome_keyring.pc is wrong in 1.10.0; when built against
libsecret it still references the gnome-keyring library.
...
Requires: apr-1 gnome-keyring-1
Requires.private: libsvn_subr
The dep should also be private; fix attached OK?
[[[
Fixes libsvn_auth_gnome_keyring.pc when configure
On Mon, Aug 21, 2017 at 03:28:35PM +0200, Johan Corveleyn wrote:
> Maybe one of the downstreamers that happens to read this could chime
> in with their thoughts? Right now we're mostly hypothesizing from our
> end ...
As a downstream package owner (Fedora/RHEL) I'd be happy to subscribe to
a low
On Fri, Aug 19, 2016 at 04:44:21PM +0300, Evgeny Kotkov wrote:
...
> The problem is caused by how mod_dav passes the output filter list to its
> providers. Please see the deliver() hook definition in mod_dav.h:1948 and
> its usage in mod_dav.c:888:
>
> /* Repository provider hooks */
> st
On Mon, Jul 13, 2015 at 11:41:23AM +0200, Branko Čibej wrote:
> https://ci.apache.org/builders/svn-x64-ubuntu-gcc/builds/317
>
> Apparently this breaks older versions of Swig ...
Ah, ugh... yes I should have realised this. Sorry.
I think the choices are:
a) keep comments in that code & use {},
On Mon, Jul 13, 2015 at 11:07:57AM +0200, Branko Čibej wrote:
> On 13.07.2015 11:04, jor...@apache.org wrote:
> > Author: jorton
> > Date: Mon Jul 13 09:04:13 2015
> > New Revision: 1690591
> >
> > URL: http://svn.apache.org/r1690591
> > Log:
> > * subversion/bindings/swig/include/proxy.swg:
> >
On Tue, Jun 18, 2013 at 06:32:13PM +0200, Lieven Govaerts wrote:
> This 204 response is not a problem, there's a special case for
> Content-Length == 0 && code == 204 in serf (response_buckets.c).
Looking at the serf (1.2.1) code, that logic isn't right, you need to
follow the precedence in the R
We've seen a few cases where the test suite for the ruby bindings fails
spuriously. There are two (separate) issues:
a) there is a race between svnserve failing in bind() because the chosen
port is already in use, and the test suite checking that the forked
svnserve process is still running.
On Sun, Feb 26, 2012 at 11:57:11AM -, svn-r...@apache.org wrote:
> Author: svn-role
> Date: Sun Feb 26 11:57:11 2012
> New Revision: 1293811
>
> URL: http://svn.apache.org/viewvc?rev=1293811&view=rev
> Log:
> Merge r1292255 from trunk:
>
> * r1292255 r1292260 r1292296 r1292322
>APR hash
On Wed, Mar 09, 2011 at 08:50:37AM -0800, Justin Erenkrantz wrote:
> On Wed, Mar 9, 2011 at 8:28 AM, Philip Martin
> wrote:
> > 0.54 epoll_wait(3, {{EPOLLOUT, {u32=30583944, u64=30583944}}}, 16,
> > 36) = 1 <0.11>
>
> As best as I can tell, that strace isn't matching your ear
20 matches
Mail list logo
