On 23/03/2019 21:22, Bruce Dubbs via blfs-dev wrote:
> On 3/23/19 12:13 PM, Pierre Labastie via blfs-dev wrote:
>>
>> On 3/23/19 5:36 PM, Pierre Labastie via blfs-dev wrote:
>>> When building plasma 5.15, everything goes well until "plasma-workspace". 
>>> Then
>>> I get the following:
>>> ---------------
>>> [...]
>>> Scanning dependencies of target appmenu
>>> [  7%] Building CXX object
>>> appmenu/CMakeFiles/appmenu.dir/appmenu_autogen/mocs_c
>>> ompilation.cpp.o
>>> [  7%] Building CXX object appmenu/CMakeFiles/appmenu.dir/appmenu.cpp.o
>>> In file included from
>>> /sources/plasma/plasma-workspace-5.15.0/appmenu/appmenu.cp
>>> p:27:
>>> /sources/plasma/plasma-workspace-5.15.0/build/config-X11.h:44:18: warning:
>>> "TRUE
>>> " is not defined, evaluates to 0 [-Wundef]
>>>   #define HAVE_X11 TRUE
>>>                    ^~~~
>>> /sources/plasma/plasma-workspace-5.15.0/appmenu/appmenu.cpp:47:5: note: in
>>> expan
>>> sion of macro ‘HAVE_X11’
>>>   #if HAVE_X11
>>>       ^~~~~~~~
>>> /sources/plasma/plasma-workspace-5.15.0/appmenu/appmenu.cpp: In member
>>> function ‘void AppMenuModule::slotWindowRegistered(WId, const QString&, 
>>> const
>>> QDBusObjectPath&)’:
>>> /sources/plasma/plasma-workspace-5.15.0/appmenu/appmenu.cpp:111:19: error:
>>> ‘QX11Info’ has not been declared
>>>           auto *c = QX11Info::connection();
>>>                     ^~~~~~~~
>>> /sources/plasma/plasma-workspace-5.15.0/appmenu/appmenu.cpp:113:16: error:
>>> ‘xcb_atom_t’ does not name a type
>>>           static xcb_atom_t s_serviceNameAtom = XCB_ATOM_NONE;
>>> [...]
>>> make[2]: *** [appmenu/CMakeFiles/appmenu.dir/build.make:98:
>>> appmenu/CMakeFiles/appmenu.dir/appmenu.cpp.o] Error 1
>>> make[1]: *** [CMakeFiles/Makefile2:1793: appmenu/CMakeFiles/appmenu.dir/all]
>>> Error 2
>>> make: *** [Makefile:130: all] Error 2
>>> ---------------
>>>
>>> The reason for all the errors is in the first warning: TRUE is not defined, 
>>> so
>>> it evaluate to 0. Then the appmenu.cpp file has:
>>> ----------
>>> #if HAVE_X11
>>> #include <QX11Info>
>>> #include <xcb/xcb.h>
>>> #endif
>>> [...]
>>> #ifdef HAVE_X11
>>>      if (KWindowSystem::isPlatformX11()) {
>>>          auto *c = QX11Info::connection();
>>> [...]
>>> ----------
>>> The first #if HAVE_X11 evaluates to false, so the header file is not 
>>> included.
>>> Then the second #ifdef HAVE_X11 evaluates to true, since HAVE_X11 _is_
>>> defined, so that "auto *c = QX11Info::connection();" is compiled, but of
>>> course, QX11Info has not been defined.
>>>
>>> There is a fix in 5.15.3.
>>>
>>> Now, how is it that we have not seen that before? I've built plasma-5.15.0
>>> already once, without error... The only difference I see is that I had
>>> gcc-8.2.0. But I tried to reinstall gcc-8.2.0, and: same error!
>>>
>>> I may well be out of my depth here.
>>
>>
>> OK, the other difference is cmake version... The above is
>> with 3.14. There is no error with 3.13.4: what is generated
>> is:
>>
>> #define HAVE_X11 1
>>
>> with 3.13.4
>>
>> while it is
>>
>> #define HAVE_X11 TRUE
>>
>> with 3.14.
> 
> I'm surprised that the problem didn't show up in kf5 or the earlier plasma
> packages.  I have not built plasma since tests for the 8.4 release and wasn't
> planning on rebuilding the next version until May.
> 
> The current version of plasma is 5.15.3.  Can you check if the problem is
> fixed there?
> 

Cmake-3.14 is a new version, and has been put recently into the book. I had
already built kf5-5.55 with the previous version of cmake, together with apps,
and for some reason, I had not built plasma. So I cannot tell whether the
problem would have shown up with KF5... I have decided to build plasma today,
but I have installed the new cmake version in the meantime.

Ok, I can try 5.15.3. I already have part of plasma installed, but I think I
can remove it with porg. If so, I'll be able to measure disk space. Hmm,
thinking about that, I'm not sure how to proceed: take the maximum of the disk
spaces before each source removal?

It's late here. And I won't be able to work much tomorrow (Family...).

So there could be no result until Monday :)

Pierre
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to