Opened 9 years ago
Closed 5 years ago
#3509 closed defect (wontfix)
OMC cannot compile libraries on OSX
Reported by: | grant@… | Owned by: | sjoelund.se |
---|---|---|---|
Priority: | high | Milestone: | Future |
Component: | Build Environment | Version: | |
Keywords: | OSX, libraries, | Cc: |
Description
So I've been trying to compile OM on OS X for a while and finally got past the usual error 2 point. Now make gets stuck while trying to compile the libraries, or that is what I think is happening as 4 omc processes start running and ran for more than 12 hours without any changes in the command line. The commands I used to get to this point are as follows:
First install all brew dependencies as per readme.
brew install gnu-sed xz PATH=$PATH:/usr/local/Cellar/gettext/0.19.6/bin/ autoconf ./configure LDFLAGS="-L/usr/local/opt/gettext/lib -L/usr/local/Cellar/lp_solve/5.5.2.0/lib" CFLAGS="-I/usr/local/opt/gettext/include -I/usr/local/Cellar/lp_solve/5.5.2.0/include" CC=clang CXX=clang++ make -j6 omc
The last block of text in the command line looks like this:
OPENMODELICA_BACKEND_STUBS=1 /Applications/Xcode.app/Contents/Developer/usr/bin/make -f Makefile generate-files-in-steps OMC=/Users/grantstephens/Repos/OpenModelica/build/bin/omc /Applications/Xcode.app/Contents/Developer/usr/bin/make -f Makefile templates scripting /Applications/Xcode.app/Contents/Developer/usr/bin/make -f Makefile --no-print-directory -C /Users/grantstephens/Repos/OpenModelica/OMCompiler/Compiler/Template /Applications/Xcode.app/Contents/Developer/usr/bin/make -f Makefile --no-print-directory -C /Users/grantstephens/Repos/OpenModelica/OMCompiler/Compiler/Script /Users/grantstephens/Repos/OpenModelica/build/bin/omc OpenModelicaScriptingAPI.mos ** AbsynDumpTpl template compilation ** ** CodegenUtil template compilation ** /Users/grantstephens/Repos/OpenModelica/build/bin/omc CodegenUtil.tpl > CodegenUtil.mo.log || (cat CodegenUtil.mo.log && false) /Users/grantstephens/Repos/OpenModelica/build/bin/omc AbsynDumpTpl.tpl > AbsynDumpTpl.mo.log || (cat AbsynDumpTpl.mo.log && false) ** CodegenC template compilation ** ** CodegenC template compilation ** ** CodegenFMUCommon template compilation ** /Users/grantstephens/Repos/OpenModelica/build/bin/omc CodegenCFunctions.tpl > CodegenCFunctions.mo.log || (cat CodegenCFunctions.mo.log && false) /Users/grantstephens/Repos/OpenModelica/build/bin/omc CodegenFMUCommon.tpl > CodegenFMUCommon.mo.log || (cat CodegenFMUCommon.mo.log && false) /Users/grantstephens/Repos/OpenModelica/build/bin/omc CodegenC.tpl > CodegenC.mo.log || (cat CodegenC.mo.log && false) /bin/sh: line 1: 99712 Illegal instruction: 4 /Users/grantstephens/Repos/OpenModelica/build/bin/omc CodegenUtil.tpl > CodegenUtil.mo.log make[6]: *** [CodegenUtil.mo] Error 1 make[6]: *** Waiting for unfinished jobs.... /bin/sh: line 1: 99719 Illegal instruction: 4 /Users/grantstephens/Repos/OpenModelica/build/bin/omc CodegenC.tpl > CodegenC.mo.log make[6]: *** [CodegenC.mo] Error 1
The CodegenC.mo.log is empty...
If any more info is needed please let me know.
Attachments (1)
Change History (8)
comment:1 follow-up: ↓ 2 Changed 9 years ago by adrpo
comment:2 in reply to: ↑ 1 Changed 9 years ago by grant@…
Thank you for the quick reply. I was able to compile that version, however when running the test-suite, 1382 of the 2689 failed- I'm assuming this is not normal- any suggestions?
Replying to adrpo:
Yes. We know about this issue but we haven't found a solution yet.
I suggest for now you take this commit as the next one breaks the compilation:
https://github.com/OpenModelica/OpenModelica/commit/2804a382451e7f2afbc514a59925823cef6668da
> git checkout 2804a382451e7f2afbc514a59925823cef6668dashould work.
The commit that doesn't work is:
https://github.com/OpenModelica/OpenModelica/commit/762e099f22b0a28a516daf422263da6ef02f2631
comment:3 follow-up: ↓ 4 Changed 9 years ago by adrpo
Yeah, that's not normal. What kind of errors are you getting?
comment:4 in reply to: ↑ 3 Changed 9 years ago by grant@…
Nothing out of the ordinary, but no expert in the matter. Makelog now attached.
Replying to adrpo:
Yeah, that's not normal. What kind of errors are you getting?
comment:5 Changed 9 years ago by dersh
I had noticed a failure of Modelica->Fluid->Examples->BranchingDynamicPipes that occurs in 133-gf27d20b_0 but does not occur in 106-ga101226_0 I'm not sure which commit in that range broke it, but I would not be at all surprised if there were a bunch of examples and tests that fail starting in that range.
comment:6 Changed 8 years ago by RexFuzzle.
So I think I have found the issue. I am able to compile and run all the tests if I uninstall open-mpi and hwloc. Don't know if its been fixed in the mean time, but thought I'd add it here for completeness.
comment:7 Changed 5 years ago by casella
- Resolution set to wontfix
- Status changed from new to closed
I guess this ticket is obsolete by now.
Yes. We know about this issue but we haven't found a solution yet.
I suggest for now you take this commit as the next one breaks the compilation:
https://github.com/OpenModelica/OpenModelica/commit/2804a382451e7f2afbc514a59925823cef6668da
should work.
The commit that doesn't work is:
https://github.com/OpenModelica/OpenModelica/commit/762e099f22b0a28a516daf422263da6ef02f2631