problem with compiling IMP after a commit "add adaptors for taking particles to reduce need for typemaps"
I have problems with compiling recent dev versions of IMP. The last version I can build successfully is 2.0.0-544-gb4dfefb.
The version 2.0.0-545-ga3aaf4a is the first that crashes (it is the version after commit a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd65 - "add adaptors for taking particles to reduce need for typemaps" https://github.com/salilab/imp/commit/a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd...).
The error is:
/home/kosinski/software/test_older_imp_a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd65/imp_dev/include/IMP/kernel/internal/../Particle.h: In constructor 'IMP::kernel::ParticleAdaptor::ParticleAdaptor(const IMP::kernel::Decorator&)': /home/kosinski/software/test_older_imp_a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd65/imp_dev/include/IMP/kernel/internal/../Particle.h:94: error: invalid use of incomplete type 'const struct IMP::kernel::Decorator' /struct/cmueller/kosinski/software/test_older_imp_a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd65/imp_dev/include/IMP/kernel/internal/../particle_index.h:18: error: forward declaration of 'const struct IMP::kernel::Decorator' /home/kosinski/software/test_older_imp_a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd65/imp_dev/include/IMP/kernel/internal/../Particle.h:95: error: invalid use of incomplete type 'const struct IMP::kernel::Decorator' /home/kosinski/software/test_older_imp_a3aaf4a5a9983d7b68ba3d5709b12ff3ad1bcd65/imp_dev/include/IMP/kernel/internal/../particle_index.h:18: error: forward declaration of 'const struct IMP::kernel::Decorator' make[2]: *** [modules/container/benchmark/CMakeFiles/container.benchmark_evaluate.dir/benchmark_evaluate.cpp.o] Error 1 make[1]: *** [modules/container/benchmark/CMakeFiles/container.benchmark_evaluate.dir/all] Error 2
My builds always worked with a commands like: $cmake_bin ../src/imp -DCMAKE_BUILD_TYPE=Release \ -DCGAL_DIR=$root_dir/lib/CGAL \ -DEIGEN3_INCLUDE_DIR=$src_dir/eigen-v3.1.3 \ -DCMAKE_INSTALL_PREFIX=$root_dir make -j 8
Is it a bug? The dev logs from http://salilab.org/imp/nightly/logs/develop/20130625/x86_64-intel8/ seem ok. So maybe I build IMP wrong way?
Thanks in advance, Jan
On 6/26/13 7:45 AM, Jan Kosinski wrote: > I have problems with compiling recent dev versions of IMP. The last > version I can build successfully is 2.0.0-544-gb4dfefb. > > The version 2.0.0-545-ga3aaf4a is the first that crashes ... > Is it a bug? The dev logs from > http://salilab.org/imp/nightly/logs/develop/20130625/x86_64-intel8/ seem > ok. So maybe I build IMP wrong way?
If you build from develop, you're going to hit compile problems from time to time, particularly when lots of things are going on, like right now. I'd suggest you check http://salilab.org/imp/nightly/results/ (as you seem to be doing already) and updating only when things look good there. The logs you point to are for 954614a11a, which is a few commits after yours, so most likely the decorator problems you ran into were fixed by that point.
Ben
It actually crashes on any version newer than 2.0.0-544-gb4dfefb
I just updated to 2.0.0-586-g29c49ec and crashes still. That pointed me to conclusion that maybe my building method is wrong ;-)
And I struggle to get the newest version to have this issue https://github.com/salilab/imp/issues/414 solved.
Cheers, Jan
On 06/26/2013 05:14 PM, Ben Webb wrote: > On 6/26/13 7:45 AM, Jan Kosinski wrote: >> I have problems with compiling recent dev versions of IMP. The last >> version I can build successfully is 2.0.0-544-gb4dfefb. >> >> The version 2.0.0-545-ga3aaf4a is the first that crashes > ... >> Is it a bug? The dev logs from >> http://salilab.org/imp/nightly/logs/develop/20130625/x86_64-intel8/ seem >> ok. So maybe I build IMP wrong way? > > If you build from develop, you're going to hit compile problems from > time to time, particularly when lots of things are going on, like > right now. I'd suggest you check > http://salilab.org/imp/nightly/results/ (as you seem to be doing > already) and updating only when things look good there. The logs you > point to are for 954614a11a, which is a few commits after yours, so > most likely the decorator problems you ran into were fixed by that point. > > Ben
OK, updated to 2.0.0-588-g04dc1f9 and I can enjoy functional IMP again :-)
I think the todays commits from drussel and maxbonomi made a difference, thanks a lot!
Jan
n 06/26/2013 05:14 PM, Ben Webb wrote: > On 6/26/13 7:45 AM, Jan Kosinski wrote: >> I have problems with compiling recent dev versions of IMP. The last >> version I can build successfully is 2.0.0-544-gb4dfefb. >> >> The version 2.0.0-545-ga3aaf4a is the first that crashes > ... >> Is it a bug? The dev logs from >> http://salilab.org/imp/nightly/logs/develop/20130625/x86_64-intel8/ seem >> ok. So maybe I build IMP wrong way? > > If you build from develop, you're going to hit compile problems from > time to time, particularly when lots of things are going on, like > right now. I'd suggest you check > http://salilab.org/imp/nightly/results/ (as you seem to be doing > already) and updating only when things look good there. The logs you > point to are for 954614a11a, which is a few commits after yours, so > most likely the decorator problems you ran into were fixed by that point. > > Ben
8am or 9am is often a good time to check out IMP as it is the last night tests plus fixes for anything that was discovered to be broken :-)
Would providing a tag/branch in git that is a more formal representation of that be useful to people? Specifically, have a "nightly" branch which is the version used in the last nightly test + any fixes for specific problems discovered.
On Wed, Jun 26, 2013 at 9:53 AM, Jan Kosinski kosinski@embl.de wrote:
> OK, updated to 2.0.0-588-g04dc1f9 and I can enjoy functional IMP again > :-) > > I think the todays commits from drussel and maxbonomi made a difference, > thanks a lot! > > Jan > > > n 06/26/2013 05:14 PM, Ben Webb wrote: > > On 6/26/13 7:45 AM, Jan Kosinski wrote: > > I have problems with compiling recent dev versions of IMP. The last > version I can build successfully is 2.0.0-544-gb4dfefb. > > The version 2.0.0-545-ga3aaf4a is the first that crashes > > ... > > Is it a bug? The dev logs from > http://salilab.org/imp/nightly/logs/develop/20130625/x86_64-intel8/ seem > ok. So maybe I build IMP wrong way? > > > If you build from develop, you're going to hit compile problems from time > to time, particularly when lots of things are going on, like right now. I'd > suggest you check http://salilab.org/imp/nightly/results/ (as you seem to > be doing already) and updating only when things look good there. The logs > you point to are for 954614a11a, which is a few commits after yours, so > most likely the decorator problems you ran into were fixed by that point. > > Ben > > > > -- > Jan Kosinski, PhD > Structural and Computational Biology Unit > European Molecular Biology Laboratory (EMBL) > Meyerhofstrasse 1 > 69117 Heidelberg > Germany > > > _______________________________________________ > IMP-users mailing list > IMP-users@salilab.org > https://salilab.org/mailman/listinfo/imp-users > >
To me would not make much difference. If I see a cool new feature added in develop branch I would do pull to give a try and go back to prev. version in case of problems. Just this time I really wanted to get the newest tag because IMP with RMF is much more fun :-)
Jan
On 06/26/2013 06:58 PM, Daniel Russel wrote: > 8am or 9am is often a good time to check out IMP as it is the last > night tests plus fixes for anything that was discovered to be broken :-) > > Would providing a tag/branch in git that is a more formal > representation of that be useful to people? Specifically, have a > "nightly" branch which is the version used in the last nightly test + > any fixes for specific problems discovered. > > > On Wed, Jun 26, 2013 at 9:53 AM, Jan Kosinski <kosinski@embl.de > mailto:kosinski@embl.de> wrote: > > OK, updated to 2.0.0-588-g04dc1f9 and I can enjoy functional IMP > again :-) > > I think the todays commits from drussel and maxbonomi made a > difference, thanks a lot! > > Jan > > > n 06/26/2013 05:14 PM, Ben Webb wrote: >> On 6/26/13 7:45 AM, Jan Kosinski wrote: >>> I have problems with compiling recent dev versions of IMP. The last >>> version I can build successfully is 2.0.0-544-gb4dfefb. >>> >>> The version 2.0.0-545-ga3aaf4a is the first that crashes >> ... >>> Is it a bug? The dev logs from >>> http://salilab.org/imp/nightly/logs/develop/20130625/x86_64-intel8/ >>> seem >>> ok. So maybe I build IMP wrong way? >> >> If you build from develop, you're going to hit compile problems >> from time to time, particularly when lots of things are going on, >> like right now. I'd suggest you check >> http://salilab.org/imp/nightly/results/ (as you seem to be doing >> already) and updating only when things look good there. The logs >> you point to are for 954614a11a, which is a few commits after >> yours, so most likely the decorator problems you ran into were >> fixed by that point. >> >> Ben > > > -- > Jan Kosinski, PhD > Structural and Computational Biology Unit > European Molecular Biology Laboratory (EMBL) > Meyerhofstrasse 1 > 69117 Heidelberg > Germany > > > _______________________________________________ > IMP-users mailing list > IMP-users@salilab.org mailto:IMP-users@salilab.org > https://salilab.org/mailman/listinfo/imp-users > > > > > _______________________________________________ > IMP-users mailing list > IMP-users@salilab.org > https://salilab.org/mailman/listinfo/imp-users
-
Ben Webb -
Daniel Russel -
Jan Kosinski