problem building qt5

Alien Bob · 01-06-2017, 07:43 AM

Quote:

Originally Posted by kjhambrick

Alien Bob --

I use more than a few of your SlackBuilds, including qt5 ( for Calibre ).

I don't see qt5-5.7.1 in the ChangeLog.txt nor is it online either.

All I do see is qt5-5.7.0 ...

Is this a pending / imminent update to qt5 ?

Thanks.

-- kjh

I think I was not thinking right when I wrote that.
I still have 5.7.1 on my TODO list, and it will be compiled against yet another new dependency: libwacom. That will get even more people complaining that their qt5 package upgrade broke their system ;-) That is also why I am delaying this package build.

willysr · 01-06-2017, 07:44 AM

x86_64 will build just fine. It happened on x86 only

kjhambrick · 01-06-2017, 08:47 AM

Quote:

Originally Posted by Alien Bob

I think I was not thinking right when I wrote that.
I still have 5.7.1 on my TODO list, and it will be compiled against yet another new dependency: libwacom. That will get even more people complaining that their qt5 package upgrade broke their system ;-) That is also why I am delaying this package build.

Thanks Alien Bob.

I was afraid that perhaps my home-brewed `get-alien-ChangeLog` script was buggy.

Thanks again for all you're doing !

-- kjh

USUARIONUEVO · 01-06-2017, 01:15 PM

im pretty sure , something broken in python 2.7.13 package.

read de errors log , i get same error in multiprocess module for meld3 (sbo)

Quote:

"/usr/lib64/python2.7/multiprocessing/queues.py", línea 48, en <módulo>

and sorry but repeat , people testing this in virtual machines ...please , in the hardware section asign more than 1 cpu or no get the error.

eror is in multiprocessing python module , only get this errors on "stable" , but no problem in "current".

lscpu

Quote:

Arquitectura: x86_64
modo(s) de operación de las CPUs:32-bit, 64-bit
Orden de los bytes: Little Endian
CPU(s): 4
Lista de la(s) CPU(s) en línea:0-3
Hilo(s) de procesamiento por núcleo:2
Núcleo(s) por «socket»:2
«Socket(s)» 1
ID de fabricante: GenuineIntel
Familia de CPU: 6
Modelo: 42
Nombre del modelo: Intel(R) Core(TM) i3-2350M CPU @ 2.30GHz
Revisión: 7
CPU MHz: 799.890
CPU MHz máx.: 2300,0000
CPU MHz mín.: 800,0000
BogoMIPS: 4589.46
Virtualización: VT-x
Caché L1d: 32K
Caché L1i: 32K
Caché L2: 256K
Caché L3: 3072K
Indicadores: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer xsave avx lahf_lm epb tpr_shadow vnmi flexpriority ept vpid xsaveopt dtherm arat pln pts

BratPit · 01-06-2017, 03:18 PM

Quote:

Originally Posted by USUARIONUEVO

i suspect is same error i get on meld3 build.
old but interesting to read --> https://bugs.python.org/issue3770

I have that bug when I try build ffmpeg library from chromium

https://gist.github.com/ruario/bec42d156d30affef655

Recompile latest stable python from 14.2 source and everything goes well.

larryhaja · 01-08-2017, 03:20 PM

I'm getting this error as well compiling qt5 with the new python 2.7.13 on Slackware64 14.2.

Code:

ImportError: This platform lacks a functioning sem_open implementation, therefore, the required synchronization primitives needed will not function, see issue 3770.

This relates to issue 3770 here. Looking at the original python 2.7.11 and the new 2.7.13 I see that _multiprocessing.so doesn't have HAVE_SEM_OPEN defined in the library as mentioned here.

Code:

bash-4.3$ strings python_2.7.11/usr/lib64/python2.7/lib-dynload/_multiprocessing.so | grep HAVE_SEM_OPEN
HAVE_SEM_OPEN
bash-4.3$ strings python_2.7.13/usr/lib64/python2.7/lib-dynload/_multiprocessing.so | grep HAVE_SEM_OPEN
bash-4.3$

Rebuilding python 2.7.13 in a VM for me with only 1 core fixed the issue.

chris.willing · 01-08-2017, 10:39 PM

Hey FTIO,

Your compilation may appear to have finished because (probably) multiple parallel make processes were being used and the error appeared in one of them some time ago while 'good' processes continued without error; you saw the end of those 'good' processes. If you kept a log of the build, you can see the error by grepping for ImportError. Otherwise you could cd to the build directory (maybe /tmp/SBo/qt-everywhere-opensource-5.6.1) and run 'make -j1' i.e. build with a single make process so that the build will halt immediately an error occurs.

chris

FTIO · 01-09-2017, 07:15 AM

Quote:

Originally Posted by chris.willing

Hey FTIO,

Your compilation may appear to have finished because (probably) multiple parallel make processes were being used and the error appeared in one of them some time ago while 'good' processes continued without error; you saw the end of those 'good' processes. If you kept a log of the build, you can see the error by grepping for ImportError. Otherwise you could cd to the build directory (maybe /tmp/SBo/qt-everywhere-opensource-5.6.1) and run 'make -j1' i.e. build with a single make process so that the build will halt immediately an error occurs.

chris

Thanks chris.willing, I'll give it a try and see how it goes.

chris.willing · 01-09-2017, 07:41 AM

Quote:

Originally Posted by USUARIONUEVO

STABLE --> Problems in some python related builds/packages arround semaphore
CURRENT --> Working as expected

Furthermore, /usr/lib64/python2.7/_sysconfigdata.py from the 14.2's previous python (2.7.11) also had

Code:

'POSIX_SEMAPHORES_NOT_ENABLED': 0,

and building qt5 with that previous version of python was fine. It sure looks like the 2.7.13 binary is the odd man out with no semaphores. Rebuilding it in a multiprocessor environment, as suggested by USUARIONUEVO, has fixed the qt5 build issue for me.

chris

kjhambrick · 01-09-2017, 09:04 AM

All --

I reran Pat's python.SlackBuild ( version 2.7.13 ) with no changes other than to change BUILD="-1_kjh" and ran `upgradepkg python-2.7.13-x86_64-1_kjh.txz`

I modified the SBo qt5.SlackBuild for qt5 version 5.7.1 and let-r-rip.

After 3-hours, 50 minutes and 43 MB of log-file entries I have an SBo version of qt5:

Code:

#ls -la /tmp/qt5-5.7.1-x86_64-1_SBo.tgz

-rw-r--r--   1 root   root  448233765 Jan  9 08:49 /tmp/qt5-5.7.1-x86_64-1_SBo.tgz

USUARIONUEVO nailed it ... Python 2.7.13 was the culprit.

-- kjh

FTIO · 01-09-2017, 10:16 AM

And after doing as chris.willing suggested, I did:

Code:

root@veggies:/home/me/built/qt5# NUMJOBS=-j1 ./qt5.SlackBuild

and got this/these errors at the end after a couple hours I think it was...

Code:

Updating projects from gyp files...
Traceback (most recent call last):
  File "/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/tools/buildscripts/gyp_qtwebengine", line 178, in <module>
    sys.exit(gyp.main(args))
  File "/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src/3rdparty/chromium/tools/gyp/pylib/gyp/__init__.py", line 538, in main
    return gyp_main(args)
  File "/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src/3rdparty/chromium/tools/gyp/pylib/gyp/__init__.py", line 514, in gyp_main
    options.duplicate_basename_check)
  File "/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src/3rdparty/chromium/tools/gyp/pylib/gyp/__init__.py", line 130, in Load
    params['parallel'], params['root_targets'])
  File "/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src/3rdparty/chromium/tools/gyp/pylib/gyp/input.py", line 2759, in Load
    check, generator_input_info)
  File "/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src/3rdparty/chromium/tools/gyp/pylib/gyp/input.py", line 596, in LoadTargetBuildFilesParallel
    parallel_state.pool = multiprocessing.Pool(multiprocessing.cpu_count())
  File "/usr/lib64/python2.7/multiprocessing/__init__.py", line 232, in Pool
    return Pool(processes, initializer, initargs, maxtasksperchild)
  File "/usr/lib64/python2.7/multiprocessing/pool.py", line 138, in __init__
    self._setup_queues()
  File "/usr/lib64/python2.7/multiprocessing/pool.py", line 233, in _setup_queues
    from .queues import SimpleQueue
  File "/usr/lib64/python2.7/multiprocessing/queues.py", line 48, in <module>
    from .synchronize import Lock, BoundedSemaphore, Semaphore, Condition
  File "/usr/lib64/python2.7/multiprocessing/synchronize.py", line 59, in <module>
    " function, see issue 3770.")
ImportError: This platform lacks a functioning sem_open implementation, therefore, the required synchronization primitives needed will not function, see issue 3770.
Project ERROR: -- running gyp_qtwebengine failed --
Makefile:116: recipe for target 'sub-gyp_run-pro-make_first' failed
make[3]: *** [sub-gyp_run-pro-make_first] Error 3
make[3]: Leaving directory '/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src/core'
Makefile:49: recipe for target 'sub-core-make_first' failed
make[2]: *** [sub-core-make_first] Error 2
make[2]: Leaving directory '/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine/src'
Makefile:44: recipe for target 'sub-src-make_first' failed
make[1]: *** [sub-src-make_first] Error 2
make[1]: Leaving directory '/tmp/SBo/qt-everywhere-opensource-src-5.6.1/qtwebengine'
Makefile:716: recipe for target 'module-qtwebengine-make_first' failed
make: *** [module-qtwebengine-make_first] Error 2

Wasn't sure if there was a log file of this build, or if there is/was where to look for it to see what or if there was an 'error 1'.

Going to try what kjh did and then try to build qt5 again but with my regular '-j7' and see what happens.

willysr · 01-09-2017, 11:29 AM

You need to rebuilt python 2.7.13 first

bassmadrigal · 01-09-2017, 12:43 PM

Quote:

Originally Posted by kjhambrick

After 3-hours, 50 minutes and 43 MB of log-file entries I have an SBo version of qt5:

It took you less than 4 hours to build qt5? Dang, I really need to upgrade my computer. Mine would take about 8-10x that. I just need to wait a little bit longer for the release of AMD Zen (although, I guess it is now called Ryzen).

FTIO · 01-09-2017, 03:41 PM

@willysr - Yep, as you can see in my post that was my plan, heh.

I did the python rebuild (I have an AMD FX-6300) and then started the slackbuilds build of qt5.6.1 using

Code:

PROPRIETARY_CODECS=yes ./qt5.SlackBuild

just for the heck of it.

Started at 10:37AM and finished at 3:23PM and it gave me a package to install. I deleted that package and I'm going to do the 5.7.1 just to have the latest and greatest now that I know it will build.

Thanks to kjhambrick and everyone who helped figure this out for the rest of us!!

kjhambrick · 01-09-2017, 04:43 PM

Thanks FITO but credit where credit is due ...

USUARIONUEVO nailed this one.

And another thing about this exercise is that it sure makes me appreciate Alien Bob's Slackware Packages even more than I already did

-- kjh