Review Request 109503: Compile with current qt dev branch

[Ben Cooksley]

> On March 15, 2013, 5:07 p.m., Stephen Kelly wrote:
> > The problem is that build.kde.org uses qt5.git, which does not yet have that commit. Actually qt5.git hasn't been updated in 6 weeks: https://codereview.qt-project.org/#change,46445
> > 
> > If b.k.o can be updated to use the dev branches of the qt 5 submodules, I think that's a good idea. In general, I think it's a bad idea to use qt5.git or to recommend anyone else use it (as is currently recommended on the wiki).
> > 
> > At any rate, this patch shouldn't go in until b.k.o can build it.
> 
> Ben Cooksley wrote:
>     Easier said than done. build.kde.org uses a separate install prefix for each project it builds.
>     Somehow I suspect the Qt5 CMake system would shatter if it was asked to install each Qt5 submodule into it's own prefix.
>     
>     Not to mention that it would make updating Qt5 on build.kde.org a royal PITA because you would have to run the module builds in their precise upstream dependency order (in case of incompatibilites) which Jenkins isn't really designed for (from what I have seen in it's interface so far at least). It would also mean maintaining a list of their modules on build.kde.org.
>     
>     Is there any good reason why the Qt developers are dragging the chain and not updating qt5.git? Obviously the individual modules build fine...
> 
> David Faure wrote:
>     qt5.git is behind because of unstability of the Qt CI. This is being worked on, I don't think we should change anything on our side.
>     You can track progress at https://codereview.qt-project.org/46445
> 
> Albert Astals Cid wrote:
>     Ok, this means that our code is uncompilable if one follows the instructions on how to compile it, but i guess we are ok with that.
> 
> David Faure wrote:
>     This is a slight simplification of the problem... the instructions do say "use qt5.git". The other solution (use qtbase+qtsvg+qtx11extras only) is already documented as suboptimal (not enough for plasma-framework, which we should at least check compiles when making API changes in kdelibs).
>     
>     Fixed: added a note to the instructions ("git checkout 90361fd in qtbase").
> 
> Stephen Kelly wrote:
>     Ben, would it be possible to run 
>     
>     git submodule foreach 'git pull || true'
>     
>     on qt5.git when updating it? That would isolate us from the need for qt5.git to get through the CI.

In theory, however Jenkins itself does Git handling for part of the process and that may very well override our invocation of that.

In any case, git spits out the following if I run that for each submodule:

You are not currently on a branch, so I cannot use any
'branch.<branchname>.merge' in your configuration file.
Please specify which branch you want to merge with on the command
line and try again (e.g. 'git pull <repository> <refspec>').
See git-pull(1) for details.

So we would need something more advanced than that.


- Ben


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/109503/#review29284
-----------------------------------------------------------


On March 15, 2013, 8:42 p.m., Albert Astals Cid wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/109503/
> -----------------------------------------------------------
> 
> (Updated March 15, 2013, 8:42 p.m.)
> 
> 
> Review request for KDE Frameworks.
> 
> 
> Description
> -------
> 
> 34a3f2cd442ff1396e508f7a8890b968f1bb3179 has been merged to qtbase and now frameworks contains definitions of functions that clash with the ones from Qt itself
> 
> 
> Diffs
> -----
> 
>   kdecore/tests/kdatetimetest.cpp d64492b 
>   libkdeqt5staging/src/CMakeLists.txt 6572b92 
>   libkdeqt5staging/src/qtest_staging.h c2b081e 
>   staging/kde4support/src/kdecore/qtest_kde.h 7174893 
> 
> Diff: http://git.reviewboard.kde.org/r/109503/diff/
> 
> 
> Testing
> -------
> 
> Compiles
> 
> 
> Thanks,
> 
> Albert Astals Cid
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20130316/434a1fd7/attachment.html>