|
Description
Francois Andriot
2012-10-11 12:35:17 CDT
Created attachment 863 [details]
libksquirrel: fix doc installation directory
libksquirel has several "/usr/share/doc" hardcoded path in the documentation directory. This patch fixes them.
Created attachment 864 [details]
kdebase: fix invalid [$e] in 'displayconfig.desktop'
The "update-desktop-database" command returns an error on 'displayconfig.desktop'. The [$e] syntax seems invalid here.
Created attachment 865 [details]
kdebindings: fix dcopc headers location
tdebinding's dcopc does not install correctly on some distro because of invalid syntax in Makefile.am.
Created attachment 866 [details]
kdemultimedia: fix XDG "music" menu (legacy RedHat patch)
Created attachment 867 [details]
kdemultimedia: fix linking xine_artsplugin with pthread (legacy RedHat patch)
Created attachment 868 [details]
kdeartwork: fix 'locolor' theme name
The 'locolor' theme name is incorrectly spelled by default. It has an uppercase "L".
Created attachment 869 [details]
kdegraphics: fix poppler-tqt include directory location
poppler-tqt includes files are installed in hardcoded in "kde" subfolder.
Created attachment 870 [details]
kdepim: fix various include directory location
Created attachment 871 [details]
python-trinity: fix 'pykdeconfig.py' installation directory
Created attachment 872 [details]
python-trinity: fix mandrake distribution detection
Created attachment 873 [details]
pykdeextensions: fix libgcc library detection, fix QT directory detection
Under RHEL/Fedora, the 'libgcc.so' file is located under "/usr/lib/xxx/" instead of "/usr/lib64/", even on x86_64 system.
Also, the QT include and QT library directories are not under standard $QTDIR location, but instead designated by $QTINC and $QTLIB variables.
Patch for kdebindings - attachment 865 [details] pushed into GIT in hash 4d7f4e86
and also for v3.5.13-sru branch - together with a few patches for Ruby.
Comment on attachment 868 [details]
kdeartwork: fix 'locolor' theme name
Pushed to GIT in hash e5a2da23 and also into v3.5.13-sru branch.
François, I believe that the patch for the poppler-tqt is not sufficient, it is also necessary to change the path in poppler-tqt.pc.cmake. Because otherwise it would be a real path and path in pkgconfig file in conflict. Please, can you confirm this? Comment on attachment 864 [details]
kdebase: fix invalid [$e] in 'displayconfig.desktop'
Pushed into GIT in hash 1f79e531 and also into v3.5.13-sru branch.
Created attachment 885 [details]
amarok: fix xine 1.2 library support
Created attachment 886 [details]
amarok: add conditionnal inotify support
It looks like amarok's inotify support is always enabled, which cause FTBFS on older distros that do not have inotify. This patch enables conditional build.
Digikam: requires commit #18ecd512 on some distributions Created attachment 887 [details]
Kaffeine: add support for xine 1.2 libraries
Created attachment 888 [details]
kchmviewer: add missing LDFLAGS causing FTBFS
Created attachment 889 [details]
kdesvn: fix "not a string literal" error
Created attachment 890 [details]
kmplayer: add support for xine 1.2 libraries
Created attachment 891 [details]
ktechlab: fix harcoded gpsim link library
Created attachment 892 [details]
kvirc: arts detection fails because of missing -lmcop flag
Created attachment 893 [details]
rosegarden: missing link to fontconfig causes FTBFS
Created attachment 894 [details]
smartcardauth: missing #include causes FTBFS
Created attachment 895 [details]
tellico: fix "not a string literal" error
Created attachment 896 [details]
tellico: adds missing LDFLAGS to avoid FTBFS
Created attachment 897 [details]
kdegraphics: fix poppler-tqt include dir in pkgconfig file
Created attachment 898 [details]
kdegraphics: fix poppler-tqt include and library dir in pkgconfig file
The poppler-tqt.pc file also contains "lib" hardcoded path, which is not suitable for distros using "lib64".
Comment on attachment 869 [details]
kdegraphics: fix poppler-tqt include directory location
Fixed by cherry-pick 7a275239 into v3.5.13-sru branch.
Comment on attachment 898 [details]
kdegraphics: fix poppler-tqt include and library dir in pkgconfig file
Pushed into GIT in hash 7ae3a9e8 and also into v3.5.13-sru branch.
Created attachment 901 [details]
ktechlab: fix harcoded gpsim link library (2)
Comment on attachment 866 [details]
kdemultimedia: fix XDG "music" menu (legacy RedHat patch)
Pushed into GIT in hash 56474575 and also into v3.5.13-sru branch.
Comment on attachment 867 [details]
kdemultimedia: fix linking xine_artsplugin with pthread (legacy RedHat patch)
Pushed into GIT in hash 1973b023 and also into v3.5.13-sru branch.
Created attachment 904 [details]
amarok: add conditionnal inotify support (2)
Comment on attachment 870 [details]
kdepim: fix various include directory location
Fixed by cherry-pick a34b0b48 into v3.5.13-sru branch.
Created attachment 926 [details]
kdenetworks: fix various cmake problems
I tried for tdenetworks on Debian and Ubuntu change from autotools to cmake and I found some problems and differences:
changed names (autotools × cmake):
+ libkntsrcfilepropsdlg × kntsrcfilepropsdlg
+ kopete_meanwhile × new_target0
+ kopete-account-kconf_update × kopete_account_kconf_update
+ kopete-pluginloader2-kconf_update × kopete_pluginloader2_kconf_update
+ kopete-nameTracking-kconf_update × kopete_nameTracking_kconf_update
changed version number for libkopete_videodevice: 0.0.0 × 1.0.0
missing install kopeteblacklister.h
wrong install path for kppp checkrules: ${DATA_INSTALL_DIR}/kppp/Rules × ${DATA_INSTALL_DIR}
changed install path for libkwireless: ${LIB_INSTALL_DIR} × ${PLUGIN_INSTALL_DIR}
files not installed into 'doc' with cmake:
+ knewsticker: TODO
+ kppp: costsgraphs.fig, kppp.faq.question, ttyS-cua.txt
+ krdc: authentication.eps, krdc_window.eps, preferences_profilestab.eps, preferences_rdpdefaultstab.eps, preferences_vncdefaultstab.eps, snapshot.eps, snapshot_connectionspeed.eps, snapshot_nobrowse.eps, snapshot_vncentry.eps
+ krfb: configuration_access.eps, configuration_network.eps, configuration_session.eps, connection.eps, email_invitation.eps, invitation_management.eps, personal_invitation.eps, screenshot.eps
Attached patch solves all above mentioned problems. However, for not installed files into doc I not know which files are really undesirable? Are all changed names and paths mistake or a intentional?
Created attachment 927 [details] tdenetworks: fix various cmake problems Same as attachment 926 [details], but for R14. Note: attachment 926 [details] as well as attachment 927 [details] also contains patch from bug 1274 (change position for 'idn') and both patches from bug 1242 (wireless-tools 30 and msn/libmimic). Slavek, I submitted several tdenetwork patches in bug report 1040.
I will test the patch in attachment 927 [details].
Thanks for the note about the patch for 1274. :)
(Odpověď na komentář #41) > Slavek, I submitted several tdenetwork patches in bug report 1040. > > I will test the patch in attachment 927 [details]. > > Thanks for the note about the patch for 1274. :) During write informations about kdenetworks cmake patch, I noticed bug 1278 - I'll look at it later. Because now I solved cmake build, so I've not tried patches for autotools from bug 1040. I'll look at it later. Stay tuned :) There are three cmake patches in bug report 1040. :) The patch in attachment 927 [details] allows me to build tdenetwork with -DBUILD_WIFI=ON, without using my patch in bug report 598.
Without either patch I can't build tdenetwork with -DBUILD_WIFI=ON. I have to use one or the other with that configure option.
As your patch is simpler, when you push to GIT I'll close bug report 598 as resolved with that patch.
The patch in attachment 926 [details] / 927 solves differences between autotools × cmake build. But the question is whether it is desirable to install added files into 'doc' folder? Who knows it?
As mentioned in the developer's mail list regarding attachment 926 [details] and 927, the extra doc files are unnecessary. Delete the three "doc" diffs in the patch and we will be okay. :)
Created attachment 933 [details]
kdenetworks: fix various cmake problems (2)
Deleted unwanted "doc" diffs.
Created attachment 934 [details]
tdenetworks: fix various cmake problems
Created attachment 935 [details]
kdenetwork: fix unwanted rename icon with autotools
When testing cmake build, I noticed that with autotools is one icon incorrectly renamed: jabber_gateway_http/ws.png × jabber_gateway_http-ws.png
Fix attached - please test it with autotools build.
No problems building with the latest patch in automake. The icon image was part of the package. Comment on attachment 935 [details]
kdenetwork: fix unwanted rename icon with autotools
Pushed to GIT in hash f425336e
Comment on attachment 934 [details]
tdenetworks: fix various cmake problems
Pushed to GIT in hash 93dc8723
Comment on attachment 933 [details]
kdenetworks: fix various cmake problems (2)
Cherry-picked into v3.5.13-sru branch
Comment on attachment 863 [details]
libksquirrel: fix doc installation directory
Pushed to GIT in hash cfc258b0 and also into v3.5.13-sru branch.
Comment on attachment 885 [details]
amarok: fix xine 1.2 library support
Cherry-picked from 8d45a64d into v3.5.13-sru branch.
Comment on attachment 904 [details]
amarok: add conditionnal inotify support (2)
Pushed to GIT in hash af8d937e and also into v3.5.13-sru branch.
(Odpověď na komentář #18)
> Digikam: requires commit #18ecd512 on some distributions
Thank you for pointing. Apparently I overlooked this patch.
Cherry-picked into v3.5.13-sru branch.
Comment on attachment 887 [details]
Kaffeine: add support for xine 1.2 libraries
Pushed to GIT in hash 222c53bf and also into v3.5.13-sru branch.
Comment on attachment 888 [details]
kchmviewer: add missing LDFLAGS causing FTBFS
Pushed to GIT in hash ec566dbe and also into v3.5.13-sru branch.
Created attachment 943 [details]
kmplayer: FTBFS with xine 1.2 libraries
François,
I tried to add to KMPlayer your patch for xine 1.2 libraries (in buildlog listed as bp900.diff), but the build fails. Buildlog attached (gzip).
Please, can you see it?
François,
I looked at the patch in attachment 901 [details] and is strange to me that for one linking contained 2× identical $(LIB_GPSIM). I traced that LIB_GPSIM contains:
$LIBREADLINE -lgpsim -lgpsim_eXdbm -lgpsimcli
So is not included gpsimgui.
Is 'gui' library need or not?
Hello, about ktechlab and libgpsim, I was so hurry to build all packages that I did not even saw that $(LIB_GPSIM) was already right here. Obviously it doesn't need to be there twice. My actual problem is that I build ktechlab without gpsim at all. In that case, the $(LIB_GPSIM) variable is empty (which is OK), but the hardcoded -lgpsimgui causes FTBFS with "cannot find -lgpsimgui". So the correct solution would be to add -lgpsimgui to $(LIB_GPSIM) only if gpsim support is requested. About kmplayer and xine 1.2 support, I see you have the following error:
error: invalid conversion from 'int' to 'xine_stream_t* {aka xine_stream_s*}'
And I've just found that I had left a "-fpermissive" in my build script, so that this error becomes a warning ! When removing -fpermissive it FTBFS as you describe ...
There was a mail from David C. Rankin about this issue:
http://trinity-devel.pearsoncomputing.net/?0::7002
But since he gets "warning invalid conversion" and not "error invalid conversion", I guess he's using -fpermissive too.
I have no solution yet.
Comment on attachment 901 [details]
ktechlab: fix harcoded gpsim link library (2)
Build without gpsim fixed in GIT hash 3fc5e3c3 and also in v3.5.13-sru branch.
(Odpověď na komentář #63)
> But since he gets "warning invalid conversion" and not "error invalid
> conversion", I guess he's using -fpermissive too.
>
> I have no solution yet.
Yes, I tested that adding -fpermissive 'solves' FTBFS with xine 1.2. Build with xine 1.1 is okay without -fpermissive. Question is whether integrate the existing patch, or wait for another solution?
Comment on attachment 892 [details]
kvirc: arts detection fails because of missing -lmcop flag
Pushed to GIT in hash 5f918c68 and also into v3.5.13-sru branch.
Comment on attachment 893 [details]
rosegarden: missing link to fontconfig causes FTBFS
Because CMake for R14 and 3.5.13-sru are significantly different, were incorporated as separate patches.
Pushed to GIT in hash af88555e (master) and 5b92ab8c (v3.5.13-sru).
Is the rosegarden patch possibly related to bug report 1076? I am following the conversation about kmplayer and xine 1.2 as a spectator. I'm just thinking out loud here in the hopes of helping. I'm curious that kmplayer fails to build against xine 1.2 yet amarok, kaffeine, and tdemultimedia/xine_artsplugin do not. The leads me to think the code somewhere in kmplayer is bad or that the other three modules have a similar but yet undiscovered bugs. As Slavek and I discovered with some other kmplayer bugs, the code in kmplayer is a tad different. I'm inclined to think the problem with xine 1.2 is kmplayer. A quick grep of the entire GIT source tree for occurrences of xine_stream_t show only a few hits: kmplayer/kmplayer/src/xineplayer.cpp:68:static xine_stream_t *stream; kmplayer/kmplayer/src/xineplayer.cpp:69:static xine_stream_t *sub_stream; kmplayer/src/xineplayer.cpp:68:static xine_stream_t *stream; kmplayer/src/xineplayer.cpp:69:static xine_stream_t *sub_stream; amarok/amarok/src/engine/xine/xine-scope.c:32:scope_port_open( xine_audio_port_t *port_gen, xine_stream_t *stream, uint32_t bits, uint32_t rate, int mode ) amarok/amarok/src/engine/xine/xine-scope.c:51:scope_port_close( xine_audio_port_t *port_gen, xine_stream_t *stream ) amarok/amarok/src/engine/xine/xine-scope.c:66:scope_port_put_buffer( xine_audio_port_t *port_gen, audio_buffer_t *buf, xine_stream_t *stream ) amarok/amarok/src/engine/xine/xine-engine.cpp:1078: xine_stream_t* tmpstream = xine_stream_new(m_xine, NULL, NULL); amarok/amarok/src/engine/xine/xine-engine.h:75: xine_stream_t *m_stream; amarok/amarok/src/engine/xine/xine-engine.h:109: xine_stream_t *m_decrease; amarok/amarok/src/engine/xine/xine-engine.h:110: xine_stream_t *m_increase; kaffeine/kaffeine/src/player-parts/xine-part/xine_part.cpp:1616: xine_stream_t* xineStreamForMeta = xine_stream_new((xine_t*)m_xine->getXineEngine(), NULL, NULL); kaffeine/kaffeine/src/player-parts/xine-part/kxinewidget.h:399: xine_stream_t* m_xineStream; tdemultimedia/xine_artsplugin/tools/thumbnail/videocreator.cpp:309: xine_stream_t *stream = xine_stream_new( xine, ao_port, vo_port ); tdemultimedia/xine_artsplugin/xinePlayObject_impl.h:118: xine_stream_t *stream; Searching the web reveals the error message in Comment 63 probably means trying to store a pointer value with a variable declared to hold an integer. As I'm very much a C++ newbie, I don't kow how to fix that. However, in the grep search above I notice the "static" data declarations in kmplayer whereas the other modules do not have that declaration. I don't know whether than makes any difference. Again, just thinking out loud and trying to help. :) Comment on attachment 894 [details]
smartcardauth: missing #include causes FTBFS
Pushed to GIT in hash c3b6defe and also into v3.5.13-sru branch.
Comment on attachment 895 [details]
tellico: fix "not a string literal" error
Pushed to GIT in hash 592d3b63 and also into v3.5.13-sru branch.
Comment on attachment 896 [details]
tellico: adds missing LDFLAGS to avoid FTBFS
Pushed to GIT in hash 04171359 and also into v3.5.13-sru branch.
Comment on attachment 871 [details]
python-trinity: fix 'pykdeconfig.py' installation directory
Pushed to GIT in hash c3a0fc05 and also into v3.5.13-sru branch.
Comment on attachment 872 [details]
python-trinity: fix mandrake distribution detection
Pushed to GIT in hash 45328775 and also into v3.5.13-sru branch.
Comment on attachment 873 [details]
pykdeextensions: fix libgcc library detection, fix QT directory detection
Pushed to GIT in hash 6221fac2 and also into v3.5.13-sru branch.
Comment on attachment 889 [details]
kdesvn: fix "not a string literal" error
Pushed to GIT in hash 104a8598. Into v3.5.13-sru branch pushed in combination with part of commit 4ce4baa8.
Created attachment 953 [details]
kmplayer: add support for xine 1.2 libraries (2)
I looked to the error message closely and found that the structure xine_event_t probably have changed the order of items stream and type. I modified the code so that the filling structure was independent of the order of items.
Please test attached patch.
For me, build is successfull on both libxine1 and libxine2.
I confirm you patch allows building on both xine 1.1 and xine 1.2. Well done ! The patch built without failure for me with xine 1.1. I haven't done any usability testing. Comment on attachment 953 [details]
kmplayer: add support for xine 1.2 libraries (2)
Pushed to GIT in hash 824f886d and also into v3.5.13-sru branch.
Created attachment 958 [details] kdesdk: fix various cmake issues During the preparation of the patch for bug 1041 I noticed a few issues with build kdesdk/tdesdk using cmake. All are addressed in the attached patch. Please test it. Created attachment 959 [details] tdesdk: fix various cmake issues Same as attachment 958 [details], but for R14. Created attachment 960 [details] tdesdk: fix various cmake issues (2) Same as attachment 958 [details], but for R14. Added one more fix for R14. Created attachment 961 [details]
kdesdk: fix various cmake issues (for TDE 3.5.13.1)
Small modification to your patch to make it build on 3.5.13.1.
Works for me now.
(Odpověď na komentář #84)
> Vytvořena příloha 961
> kdesdk: fix various cmake issues (for TDE 3.5.13.1)
>
> Small modification to your patch to make it build on 3.5.13.1.
> Works for me now.
François, it is possible that ${BIN_INSTALL_DIR}/kapptemplate/bin is your local modification? The GIT version is in the master and also in v3.5.13-sru branch ${DATA_INSTALL_DIR}/kapptemplate/bin - as foreseen by the original patch.
You are right, this path modification was one of my local patch. I have no idea why I changed this path... Anyway your patch is OK finally. Comment on attachment 960 [details]
tdesdk: fix various cmake issues (2)
Pushed to GIT in hash 3dd9fca2 and also into v3.5.13-sru branch.
Created attachment 967 [details]
kdeartwork: fix various cmake and automake issues
I tried to compare automake and cmake build kdeartwork. I found some differences. Patch attached.
I'm not sure how it is with krootbacking. With automake is installed always, for cmake was missing. With automake install path is hard coded to /usr/local/bin on Debian / Ubuntu is changed to /opt/trinity/bin. Because /opt/trinity/bin seems to me as a good default path, the patch changes this path. What is your opinion?
Comment on attachment 967 [details]
kdeartwork: fix various cmake and automake issues
Note: This patch is valid also for R14
Created attachment 974 [details]
kdevelop: fix various cmake issues
I went hunting for differences automake × cmake build in tdevelop.
There were a lot. The attached patch solves this.
Please test it.
Note: Patch for v3.5.13-sru contain parts from commits 998ebd2e and 159cec9e.
Created attachment 975 [details]
tdevelop: fix various cmake issues
Same as attachmend 974, but for R14.
I tested the patch in attachment 975 [details]. The package fails to build with the following error:
/usr/lib/gcc/i486-slackware-linux/4.4.4/../../../../i486-slackware-linux/bin/ld: cannot find -lkdeui-shared
Created attachment 983 [details]
tdevelop: fix various cmake issues (2)
Fixed overlooked renaming.
The updated patch is working here. :) No problems with the patch in attachment 967 [details] with GIT R14.
(Odpověď na komentář #95)
> No problems with the patch in attachment 967 [details] with GIT R14.
The proposal to change the default path for xscreensaver-getimage* from /usr/local/bin to $(kde_bindir) without objections?
No objections. :) Comment on attachment 967 [details]
kdeartwork: fix various cmake and automake issues
Pushed to GIT in hash 54c21d38 and also into v3.5.13-sru branch.
Comment on attachment 974 [details]
kdevelop: fix various cmake issues
Pushed to GIT in hash 96f0726b and also into v3.5.13-sru branch.
Other issues relating to the forthcoming 3.5.13.2 (and also R14) will be addressed in a new bug 1393. |