dev-lang/icc fails to merge at the install phase with the following error: >>> Install icc-15.0.2.164 into /var/tmp/portage/dev-lang/icc-15.0.2.164/image/ category dev-lang * Cleaning out documentation ... [ ok ] * Cleaning out examples ... [ ok ] * Cleaning out eclipse plugin ... [ ok ] !!! doman: opt/intel/composerxe-2015_update2.2.164/man/man1/* does not exist doman failed * Tagging icc ... [ ok ] >>> Completed installing icc-15.0.2.164 into /var/tmp/portage/dev-lang/icc-15.0.2.164/image/ * QA Notice: file does not exist: * * doman: opt/intel/composerxe-2015_update2.2.164/man/man1/* does not exist strip: x86_64-pc-linux-gnu-strip --strip-unneeded -R .comment -R .GCC.command.line -R .note.gnu.gold-version opt/intel/composerxe-2015_update2.2.164/compiler/lib/mic/libchkp.so opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/libchkp.so opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/cilk_db.so opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/libgfxoffload.so opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/libchkp.so opt/intel/composerxe-2015_update2.2.164/compiler/lib/mic/libchkpwrap_w.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/cilk_db.so opt/intel/composerxe-2015_update2.2.164/compiler/lib/mic/libchkpwrap_h_w.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/mic/libchkpwrap_h.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/mic/libchkpwrap.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/libchkpwrap_w.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64_gfx/libsvml.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/libchkpwrap_h_w.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/libchkpwrap_h.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/libchkpwrap.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/gfxbegin.o opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/crt/crtend.o opt/intel/composerxe-2015_update2.2.164/compiler/lib/intel64/crt/crtbegin.o opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/libchkpwrap_w.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/libchkpwrap_h_w.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/libchkpwrap.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/libchkpwrap_h.a opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/crt/crtend.o opt/intel/composerxe-2015_update2.2.164/compiler/lib/ia32/crt/crtbegin.o ecompressdir: bzip2 -9 /usr/share/man ecompressdir: bzip2 -9 /usr/share/doc Traceback (most recent call last): File "/usr/lib/python-exec/python3.4/emerge", line 50, in <module> retval = emerge_main() File "/usr/lib64/python3.4/site-packages/_emerge/main.py", line 1154, in emerge_main return run_action(emerge_config) File "/usr/lib64/python3.4/site-packages/_emerge/actions.py", line 3188, in run_action emerge_config.args, spinner) File "/usr/lib64/python3.4/site-packages/_emerge/actions.py", line 495, in action_build retval = mergetask.merge() File "/usr/lib64/python3.4/site-packages/_emerge/Scheduler.py", line 1023, in merge rval = self._merge() File "/usr/lib64/python3.4/site-packages/_emerge/Scheduler.py", line 1412, in _merge self._main_loop() File "/usr/lib64/python3.4/site-packages/_emerge/Scheduler.py", line 1389, in _main_loop self._event_loop.iteration() File "/usr/lib64/python3.4/site-packages/portage/util/_eventloop/EventLoop.py", line 270, in iteration if not x.callback(f, event, *x.args): File "/usr/lib64/python3.4/site-packages/portage/util/_async/PipeLogger.py", line 92, in _output_handler self.wait() File "/usr/lib64/python3.4/site-packages/_emerge/AsynchronousTask.py", line 57, in wait self._wait_hook() File "/usr/lib64/python3.4/site-packages/_emerge/AsynchronousTask.py", line 175, in _wait_hook self._exit_listener_stack.pop()(self) File "/usr/lib64/python3.4/site-packages/_emerge/SpawnProcess.py", line 168, in _pipe_logger_exit self.wait() File "/usr/lib64/python3.4/site-packages/_emerge/AsynchronousTask.py", line 57, in wait self._wait_hook() File "/usr/lib64/python3.4/site-packages/_emerge/AsynchronousTask.py", line 175, in _wait_hook self._exit_listener_stack.pop()(self) File "/usr/lib64/python3.4/site-packages/_emerge/EbuildPhase.py", line 290, in _post_phase_exit _post_src_install_soname_symlinks(self.settings, out) File "/usr/lib64/python3.4/site-packages/portage/package/ebuild/doebuild.py", line 2329, in _post_src_install_soname_symlinks soname_deps.add(entry) File "/usr/lib64/python3.4/site-packages/portage/util/_dyn_libs/soname_deps.py", line 63, in add "Missing multilib category data: %s" % entry.filename) AssertionError: Missing multilib category data: /opt/intel/composerxe-2015_update2.2.164/compiler/lib/mic/libchkp.so Reproducible: Always
Please give us all information.
Created attachment 399600 [details] emerge --info
Attached above is the emerge --info. The build log does not contain any new information than what I've already posted. The problem is that there seems to be a check for multilib data on libraries that are meant for use with the Xeon Phi coprocessor, which obviously won't have that, since they don't run on the host x86/x86_64 processor at all. I couldn't see where that check is being made, but it needs to be disabled for all libraries whose architecture is k1om (Intel's MIC architecture). Please let me know anything else you need to know, and how I may help with this.
By the way, this is not a problem with the ebuild itself, but with new tests recently introduced by portage. I tried downgrading portage to 2.2.14 and emerging dev-lang/icc with that and it worked just fine.
(In reply to Guilherme Amadio from comment #2) > Created attachment 399600 [details] > emerge --info We need "emerge --info dev-lang/icc" to know your USE flag.
Created attachment 399602 [details] emerge --info dev-lang/icc Here it is.
Well, except for portage version, which was the latest when it failed. With version 2.2.14 it works.
Could you guys explain to me what is happening?
(In reply to Justin Lecher from comment #8) > Could you guys explain to me what is happening? By "you guys" do you mean myself? I am not sure what you would like to know. In comment #3 I explain what is happening. Portage is checking for x64 multilib data on libraries that are not meant for x64, but for k1om (MIC architecture). We should probably copy portage developers and ask what can be done to fix this. The problem is not in the ebuild itself. With portage 2.2.14 it merges fine, because the tests aren't there yet, but on the latest version of portage, the new tests make emerging icc fail with the error shown in the bug description.
(In reply to Guilherme Amadio from comment #9) > (In reply to Justin Lecher from comment #8) > > Could you guys explain to me what is happening? > > By "you guys" do you mean myself? I am not sure what you would like to know. > I meant the portage devs. > In comment #3 I explain what is happening. Portage is checking for x64 > multilib data on libraries that are not meant for x64, but for k1om (MIC > architecture). > We should probably copy portage developers and ask what can be done to fix > this. The problem is not in the ebuild itself. With portage 2.2.14 it merges > fine, because the tests aren't there yet, but on the latest version of > portage, the new tests make emerging icc fail with the error shown in the > bug description. Yeah, I can reproduce it, but I don't understand what is wrong. Perhaps we need some whitelisting here, as this is a binary package and installed as upstream intend it to be. So nothing we can do something against.
*** This bug has been marked as a duplicate of bug 543818 ***
(In reply to Justin Lecher from comment #10) > Yeah, I can reproduce it, but I don't understand what is wrong. The patch for bug 543818 handles it by generating a "QA Notice: Unrecognized ELF file(s)" eqawarn message if the files are not matched by QA_PREBUILT. > Perhaps we need some whitelisting here, as this is a binary package and > installed as upstream intend it to be. So nothing we can do something > against. The intel-sdp eclass sets QA_PREBUILT="${INTEL_SDP_DIR}/*", which should suppress the QA Notice for this package.
Thank you @zmedico for your comment. The unknown libraries/binaries are for the Xeon Phi coprocessor, so they are not meant to run on the host (as mentioned in the other bug).