On Sun, Mar 20, 2016 at 8:16 PM, Ingo Liebhardt <ingo.liebha...@ziggo.nl> wrote:
> Hi Roman, > > Thanks a lot! > I’m also now pretty sure it’s not darktable’s fault. ;-) > > Is OpenMP an absolute requirement to build? Or can it be inactivated? (not > talking about performance at this stage) > No, OpenMP is absolutely not a requirement. You can disable it either by passing --disable-openmp to ./build.sh (anyone uses that script?) Or by passing -DUSE_OPENMP=OFF to cmake. In any case thanks so much for your help so far. > > Another maybe relevant side-info: > I followed the instructions in BUILD.txt, in particular: > To have working OpenMP (at least partly, only for C source files) you will > also need gcc49: > > sudo port install gcc49 > So I did this and interestingly the commit from mid-February worked. > > in any case, I’ll have to have a closer look at my toolchain, that’s clear > to me. > > Thanks again & cheers, > Ingo > Roman. > > > Am 20.03.2016 um 18:04 schrieb Roman Lebedev <lebedev...@gmail.com>: > > On Sun, Mar 20, 2016 at 7:58 PM, Ingo Liebhardt <ingo.liebha...@ziggo.nl> > wrote: > > Aha, here we seem to have something else now. > > Finally :) > > What I get this time: > [ 66%] Building CXX object > src/iop/CMakeFiles/demosaic.dir/amaze_demosaic_RT.cc.o > [ 67%] Linking CXX shared module libdemosaic.so > clang: warning: argument unused during compilation: '-pthread' > > For clang, openmp support is _really_ experimental. > So i would guess you either need newer compiler > (no less than clang-3.7, clang-3.8+ highly preferable) > > Or your toolchain is broken, e.g. it does not link to the openmp library/ > does not detect the right library to link to. > > In either case, i'm somewhat sure this is not darktable fault. > With clang, i would suggest to avoid using openmp for a few clang > releases. (until 3.9 i'd say) > > Undefined symbols for architecture x86_64: > "_GOMP_atomic_end", referenced from: > _green_equilibration_favg._omp_fn.1 in introspection_demosaic.c.o > "_GOMP_atomic_start", referenced from: > _green_equilibration_favg._omp_fn.1 in introspection_demosaic.c.o > "_GOMP_loop_dynamic_next", referenced from: > _xtrans_markesteijn_interpolate._omp_fn.0 in > introspection_demosaic.c.o > "_GOMP_loop_dynamic_start", referenced from: > _xtrans_markesteijn_interpolate._omp_fn.0 in > introspection_demosaic.c.o > "_GOMP_loop_end_nowait", referenced from: > _xtrans_markesteijn_interpolate._omp_fn.0 in > introspection_demosaic.c.o > "_GOMP_parallel", referenced from: > _lin_interpolate in introspection_demosaic.c.o > _green_equilibration_favg in introspection_demosaic.c.o > _green_equilibration_lavg in introspection_demosaic.c.o > _vng_interpolate in introspection_demosaic.c.o > _process in introspection_demosaic.c.o > "_omp_get_num_procs", referenced from: > _process in introspection_demosaic.c.o > "_omp_get_num_threads", referenced from: > _green_equilibration_favg._omp_fn.2 in introspection_demosaic.c.o > _green_equilibration_favg._omp_fn.1 in introspection_demosaic.c.o > _green_equilibration_lavg._omp_fn.3 in introspection_demosaic.c.o > _passthrough_monochrome._omp_fn.4 in introspection_demosaic.c.o > _vng_interpolate._omp_fn.5 in introspection_demosaic.c.o > _vng_interpolate._omp_fn.6 in introspection_demosaic.c.o > _lin_interpolate._omp_fn.8 in introspection_demosaic.c.o > ... > "_omp_get_thread_num", referenced from: > _xtrans_markesteijn_interpolate._omp_fn.0 in > introspection_demosaic.c.o > _green_equilibration_favg._omp_fn.2 in introspection_demosaic.c.o > _green_equilibration_favg._omp_fn.1 in introspection_demosaic.c.o > _green_equilibration_lavg._omp_fn.3 in introspection_demosaic.c.o > _passthrough_monochrome._omp_fn.4 in introspection_demosaic.c.o > _vng_interpolate._omp_fn.5 in introspection_demosaic.c.o > _vng_interpolate._omp_fn.6 in introspection_demosaic.c.o > ... > ld: symbol(s) not found for architecture x86_64 > clang: error: linker command failed with exit code 1 (use -v to see > invocation) > make[2]: *** [src/iop/libdemosaic.so] Error 1 > make[1]: *** [src/iop/CMakeFiles/demosaic.dir/all] Error 2 > make: *** [all] Error 2 > > Cheers, > Ingo > > Roman. > > > > Am 20.03.2016 um 17:46 schrieb Roman Lebedev <lebedev...@gmail.com>: > > On Sun, Mar 20, 2016 at 7:35 PM, Ingo Liebhardt <ingo.liebha...@ziggo.nl> > wrote: > > Hi Roman, > > Tried it. > > Unfortunately still there :-( > [ 66%] Building CXX object > src/iop/CMakeFiles/demosaic.dir/amaze_demosaic_RT.cc.o > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:42:11: error: > expected > unqualified-id > if(std::isinf(x)) > ^ > /usr/include/math.h:174:5: note: expanded from macro 'isinf' > ( sizeof(x) == sizeof(float) ? __inline_isinff((float)(x)) \ > ^ > > Err, right, that is the reverse issue from what i have fixed. > Try again, maybe this time? :) > > Cheers, > Ingo > > Roman. > > Am 20.03.2016 um 17:26 schrieb Roman Lebedev <lebedev...@gmail.com>: > > On Sun, Mar 20, 2016 at 7:20 PM, Ingo Liebhardt <ingo.liebha...@ziggo.nl> > wrote: > > Hi Roman, > > I get it with this commit: > commit 8de505ce8b41251442704b9567f73db9fb7703ce > Author: Ulrich Pegelow <ulrich.pege...@tongareva.de> > Date: Sun Mar 20 13:11:25 2016 +0100 > > tonecurve: fix long standing bug in opencl codepath; fixes #10957 > > Hm, please execute: > $ cd darktable/ && git clean -d -f -x > > and re-run cmake and make > > Cheers, > Ingo > > Roman. > > > Am 20.03.2016 um 17:13 schrieb Roman Lebedev <lebedev...@gmail.com>: > > On Sun, Mar 20, 2016 at 7:08 PM, Ingo Liebhardt <ingo.liebha...@ziggo.nl> > wrote: > > Dear all, > > I am taking my first prudent steps into the darktable code, trying to get a > bit accustomed to it. > Past makes used to work just fine. > I’m using OS X + MacPorts. > I’ve not been pulling for a while. > > I pulled today, the build failed (also after make cleans) on > amaze_demosaic_RT.cc. > > To help you possibly locating something, the latest commit that was > building > perfectly for was this one: > > —— > git log -1 > commit 3e8b92fb7d080feb2bdd6b0b92dfd7cfe28949fa > Author: Roman Lebedev <lebedev...@gmail.com> > Date: Sun Feb 21 00:23:53 2016 +0300 > > Graduatednd IOP: move gradient RGB color computation into > commit_params() > —— > > The error I get now is this one: > > With which commit? I'm pretty sure i fixed that already. > > —— > Scanning dependencies of target demosaic > [ 66%] Building C object > src/iop/CMakeFiles/demosaic.dir/introspection_demosaic.c.o > [ 66%] Building CXX object > src/iop/CMakeFiles/demosaic.dir/amaze_demosaic_RT.cc.o > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:42:11: error: > expected unqualified-id > if(std::isinf(x)) > ^ > /usr/include/math.h:174:5: note: expanded from macro 'isinf' > ( sizeof(x) == sizeof(float) ? __inline_isinff((float)(x)) \ > ^ > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:43:10: error: no > member named '__builtin_isless' > in namespace 'std'; did you mean simply '__builtin_isless'? > r = (std::isless(x, m) ? m : (std::isgreater(x, M) ? M : x)); > ^~~~~ > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:43:15: note: > '__builtin_isless' declared here > r = (std::isless(x, m) ? m : (std::isgreater(x, M) ? M : x)); > ^ > /usr/include/math.h:559:22: note: expanded from macro 'isless' > #define isless(x, y) __builtin_isless((x),(y)) > ^ > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:43:35: error: no > member named > '__builtin_isgreater' in namespace 'std'; did you mean simply > '__builtin_isgreater'? > r = (std::isless(x, m) ? m : (std::isgreater(x, M) ? M : x)); > ^~~~~ > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:43:40: note: > '__builtin_isgreater' declared here > r = (std::isless(x, m) ? m : (std::isgreater(x, M) ? M : x)); > ^ > /usr/include/math.h:557:25: note: expanded from macro 'isgreater' > #define isgreater(x, y) __builtin_isgreater((x),(y)) > ^ > /Users/ingo/src/darktable/src/iop/amaze_demosaic_RT.cc:44:16: error: > expected unqualified-id > else if(std::isnan(x)) > ^ > /usr/include/math.h:179:5: note: expanded from macro 'isnan' > ( sizeof(x) == sizeof(float) ? __inline_isnanf((float)(x)) \ > ^ > 4 errors generated. > —— > > With full build messages as annex below. > > Hope it helps you. > > I’d be glad for any hint. > > Cheers, > Ingo > > > Annex: > ... > > > > Roman. > > ___________________________________________________________________________ > darktable developer mailing list to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > > > ___________________________________________________________________________ > darktable developer mailing list > to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > > > ___________________________________________________________________________ > darktable developer mailing list > to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > > > ___________________________________________________________________________ > darktable developer mailing list > to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > > > ___________________________________________________________________________ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
