[Digikam-devel] digikam5 make error

[Simon Frei]

Ok thanks a lot for your patience, now I got digikam5 and libkipi to 
compile for the first time :)
One thing that might again be something I did wrong, as I am new to 
cmake/make: Strangely the -DENABLE_AKONADICONTACTSUPPORT=off does not 
seem to have an effect. As long as the akonadi libraries are found it 
gets compiled with support for akonadi resulting in the same error as 
before. This can be solved by deinstalling those libraries. For 
reference here the command I used:
cmake -DCMAKE_INSTALL_PREFIX=/usr/local/digikam/ 
-DCMAKE_BUILD_TYPE=debug -DENABLE_INTERNALMYSQL=on 
-DENABLE_MYSQLSUPPORT=on -DENABLE_AKONADICONTACTSUPPORT=off .

On 02/03/16 05:12, Gilles Caulier wrote:
> The test code do not need to be compiled in client computer. It's for 
> developper only... There is a cmake flag to disable it 
> : -DBUILD_TESTING=OFF
>
> Gilles Caulier
>
> 2016-03-01 20:00 GMT+01:00 Simon Frei <freisim93 at gmail.com 
> <mailto:freisim93 at gmail.com>>:
>
>     Thanks for taking the time to look at my problems, unfortunately I
>     need to ask again:
>     Cmake aborts when libkipi is not found. So I tried to build
>     libkipi from the most recent package 15.12.2 and I run into the
>     following error:
>
>     [ 58%] Linking CXX executable kipicmd
>     /usr/bin/ld: cannot open output file kipicmd: Is a directory
>     collect2: error: ld returned 1 exit status
>     tests/CMakeFiles/kipicmd.dir/build.make:250: recipe for target
>     'tests/kipicmd' failed
>     make[2]: *** [tests/kipicmd] Error 1
>     CMakeFiles/Makefile2:230: recipe for target
>     'tests/CMakeFiles/kipicmd.dir/all' failed
>     make[1]: *** [tests/CMakeFiles/kipicmd.dir/all] Error 2
>     Makefile:138: recipe for target 'all' failed
>     make: *** [all] Error 2
>
>     By googling I determined that the problem is, that ld tries to
>     create an executable named kipicmd in tests/ but there is already
>     a directory of that name in there with the "main.cpp" in it.
>     Changing the folder name doesn't work as make needs the main.cpp.
>     A workaround is to rename the output name in
>     /tests/CMakeFiles/kipicmd.dir/link.txt to kipicmd2 (same for
>     kxmlkipicmd), then it builds to 100% but fails installing.
>
>     Again any hints appreciated.
>
>
>
>     On 01/03/16 15:49, Gilles Caulier wrote:
>>     See this report :
>>
>>     https://bugs.kde.org/show_bug.cgi?id=359897
>>
>>     libkipi is optional. If library is not found, the kipi support
>>     will be disabled.
>>
>>     Gilles Caulier
>>
>>     2016-02-29 22:41 GMT+01:00 Simon Frei <freisim93 at gmail.com
>>     <mailto:freisim93 at gmail.com>>:
>>
>>         That is weird, the README states that the akonadi contacts
>>         part (which I do not need) is disabled by default and I used
>>         "cmake -DENABLE_MYSQLSUPPORT=on .". Anyway, I now explicitly
>>         disabled it by "make -DENABLE_MYSQLSUPPORT=on
>>         -DENABLE_AKONADICONTACTSUPPORT=off ." and now the error is
>>         with kipiinterface.cpp (see new attachment), which I guess is
>>         part of libkipi and thus non-optional?
>>
>>
>>         On 29/02/16 20:09, Gilles Caulier wrote:
>>>         Problem is with you libKF5AkonadiContact.so.
>>>         Why do you want to enable this feature which is a very small
>>>         part of digiKam : the way to get contacts list from Kontact
>>>         application to populate Tags... It still experimental. You
>>>         can forget it for the moment. Turn of the option with cmake
>>>         at configuration time. Look README for details.
>>>         Gilles Caulier
>>>
>>>         2016-02-29 7:40 GMT+01:00 Simon Frei <freisim93 at gmail.com
>>>         <mailto:freisim93 at gmail.com>>:
>>>
>>>             As all dependencies are available in the debian repos, I
>>>             tried to build digikam5. Cmake runs smoothly and finds
>>>             everything, but make aborts at various stages, curiously
>>>             not the same ones every run. I first thougth that using
>>>             many cores (-j 4) is problematic, but the problem
>>>             remains without it. The output of make in the latter
>>>             case is attached. I appreciate any hints about where to
>>>             investigate and please ask for what additional
>>>             information you need.
>>>
>>>             Cheers,
>>>             Simon
>>>
>>>             _______________________________________________
>>>             Digikam-devel mailing list
>>>             Digikam-devel at kde.org <mailto:Digikam-devel at kde.org>
>>>             https://mail.kde.org/mailman/listinfo/digikam-devel
>>>
>>>
>>>
>>>
>>>         _______________________________________________
>>>         Digikam-devel mailing list
>>>         Digikam-devel at kde.org <mailto:Digikam-devel at kde.org>
>>>         https://mail.kde.org/mailman/listinfo/digikam-devel
>>
>>
>>         _______________________________________________
>>         Digikam-devel mailing list
>>         Digikam-devel at kde.org <mailto:Digikam-devel at kde.org>
>>         https://mail.kde.org/mailman/listinfo/digikam-devel
>>
>>
>>
>>
>>     _______________________________________________
>>     Digikam-devel mailing list
>>     Digikam-devel at kde.org <mailto:Digikam-devel at kde.org>
>>     https://mail.kde.org/mailman/listinfo/digikam-devel
>
>
>     _______________________________________________
>     Digikam-devel mailing list
>     Digikam-devel at kde.org <mailto:Digikam-devel at kde.org>
>     https://mail.kde.org/mailman/listinfo/digikam-devel
>
>
>
>
> _______________________________________________
> Digikam-devel mailing list
> Digikam-devel at kde.org
> https://mail.kde.org/mailman/listinfo/digikam-devel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/digikam-devel/attachments/20160302/1aa92aeb/attachment.html>