> a) precache the results of these checks in cmake
>   (so we don't run pkgconfig tests at all!)
>    - tricky, I'm not too familiar with cmake in this regard but would assist.

sorry for my unfamiliarity with cmake.
you mean, "add an option (such as -C) to cmake in the webkit2-efl.spec"?
how can I make the 'precached result'?

> b) fix qemu - can be hard, but useful. see the launchpad entry.
> c) make a pkgconfig-x86-arm package as transparent replacement to the arm
>    pkgconfig for usage in obs and gbs (so we don't execute qemu)
> d) same for cmake

I cannot understand, could you explain how to fix qemu, how to make a
pkgconfig-x86-arm as you mentioned. (step by step command line
descriptions would be better)

> If you just need a patched webkit-efl, I could produce one for you quickly.

I don't know exactly what you mean the patched webkit-efl (binary or
source code for webkit-efl?).
Everytime, I need to update and build only webkit-efl source code but
not other packages.
If the patched webkit-efl is helpful to solve above build problem, It
will be appreciated.

Thank you again so much, Jan-Simon.

- JY


On Tue, Apr 2, 2013 at 11:31 PM, Jan-Simon Möller <[email protected]> wrote:
> On Tuesday 02 April 2013 22:03:50 Jeong-Joon Yoo wrote:
>> Thank you for your help again, Jan-Simon,
>>
>> After patching the qemu with attached 'cmake.patch'
>> I did configure, make, make install the qemu
>> here is the result, for the following command (hang problem is occurred)
>> $ gbs build -A armv7l --include-all --clean --clean-repos
>>
>> [  308s] + cmake /home/abuild/rpmbuild/BUILD/webkit2-efl-123997_0.9.84
>> -DPORT=Efl -DCMAKE_INSTALL_PREFIX=/usr -DCMAKE_BUILD_TYPE=Release
>> -DENABLE_WEBKIT=Off -DENABLE_WEBKIT2=On -DENABLE_ALLINONE=On
>> -DEFL_TARGET=arm
>> [  311s] -- The C compiler identification is GNU
>> [  313s] -- The CXX compiler identification is GNU
>> [  313s] -- Check for working C compiler: /usr/bin/gcc
>> [  315s] -- Check for working C compiler: /usr/bin/gcc -- works
>> [  315s] -- Detecting C compiler ABI info
>> [  317s] -- Detecting C compiler ABI info - done
>> [  318s] -- Check for working CXX compiler: /usr/bin/c++
>> [  320s] -- Check for working CXX compiler: /usr/bin/c++ -- works
>> [  320s] -- Detecting CXX compiler ABI info
>> [  322s] -- Detecting CXX compiler ABI info - done
>> [  323s] -- Found BISON: /usr/bin/bison
>> [  323s] -- Found FLEX: /usr/bin/flex
>> [  323s] -- Found Gperf: /usr/bin/gperf
>> [  323s] -- Found Perl: /usr/bin/perl
>> [  324s] -- Found PythonInterp: /usr/bin/python2.7
>> [  324s] -- Could NOT find Ruby (missing:  RUBY_EXECUTABLE)
>> [  324s] -- checking for module 'cairo'
>> [  324s] --   found cairo, version 1.12.8
>>
>> hang at this point. (this hang point is different whenever I run gbs build)
>
> Oh, bad. For me it works 90% of the time (still hangs sometimes). Seems like
> you still trigger the aforementioned race condition in qemu. Maybe follow-up
> on the bugreport to bump it again.
>
>
> So lets consider a few ways out here in order of what to try first:
> a) precache the results of these checks in cmake
>   (so we don't run pkgconfig tests at all!)
>    - tricky, I'm not too familiar with cmake in this regard but would assist.
> b) fix qemu - can be hard, but useful. see the launchpad entry.
> c) make a pkgconfig-x86-arm package as transparent replacement to the arm
>    pkgconfig for usage in obs and gbs (so we don't execute qemu)
> d) same for cmake
>
> There's unfortunately no simple fix atm.
>
> If you just need a patched webkit-efl, I could produce one for you quickly.
>
>
> Best,
> JS
_______________________________________________
General mailing list
[email protected]
https://lists.tizen.org/listinfo/general

Reply via email to