-
Notifications
You must be signed in to change notification settings - Fork 35
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
version.jl fatal: No names found, cannot describe anything. #28
Comments
These cholmod errors are likely because of an old version of suite-sparse. This formula needs to be updated to version 4.2.1. Can you change Don't worry about the |
I changed suite-sparse-julia.rb and suite-sparse64-julia.rb to SuiteSparse-4.2.1, these the errors I get this time clang++ -mmacosx-version-min=10.6 -shared -Xlinker -all_load libsuitesparseconfig.a libspqr.a -o /private/tmp/julia-vWKO/usr/lib/libspqr.dylib -L/usr/local/opt/suite-sparse64-julia/lib -L/usr/local/opt/arpack64-julia/lib -L/usr/local/opt/openblas64-julia/lib -L/usr/local/opt/readline/lib -L/usr/local/lib -L/opt/X11/lib -headerpad_max_install_names -L/private/tmp/julia-vWKO/usr/lib -lcholmod -lcolamd -lamd -lopenblas && I check the content of /private/tmp/julia-vWKO/usr/lib/ and there are the following libs There is no libspqr.dylib. Would you give me any further directions to dig in? |
I have a suspicion that we need to add We can attempt to do this by adding in more
|
I added the those lines to suite-sparse-julia.rb but the errors remain. We have to somehow tell SuiteSparse to compile dynamic lib. |
I have updated my SuiteSparse formula with the changes, tested on an OSX 10.8.2 machine, and everything seems to be working. Try |
Thanks for spending time working on the formula! I was not sure how to update tap formulas with brew update. I tried this way
I also wanted to try 64bit compilation, so I copied suite-sparse-julia.rb to suite-sparse64-julia.rb and have changed openblas-julia openblas64-julia at line 9. The updated formula indeed works, I was able to compile and run julia. While the installation process I got these suspicious lines
Running brew test -v julia I got this
Will it somehow affect julia? |
I tried to install 32bit and it passes the tests.
By the way what is red colored lines mean? Shall we forget about 64 bit version? |
Yes, that is likely why. I've updated the 64-bit build of Suite-Sparse-Julia to 4.2.1, that should solve the errors you mention above. Can you The 64-bit build of Julia is a funny distinction. It's not that it is any faster, or able to hold more memory overall than the 32-bit version. Rather, it means that the dimensions of arrays can be 64-bit (which most BLAS/LAPACK packages do not support) so that you can have arrays larger than 2^32 by 2^32. If you don't deal with such enormous matrices, the 64-bit version holds no benefit for you. |
I've tried again, now it is not building at all. I am getting the same errors as before
Regarding the necessity of 64bit build, I agree that one rarely needs it. Although I have arrays in Matlab that reach 1Gb (1,000,000 of 128 vectors where each component is a double type) that is quite close to the limit of 4Gb in 32bits systems. |
Can you verify that the lines adding in the |
I have the same problem (see issue 24) now. Both -DNCAMD -DNPARTITION are present in suite-sparse64-julia.rb. |
Yep, having the same problem. |
I figured this one out; my options were getting overridden by the same line that differentiates the 64 and 32-bit builds. I've uploaded a fix, should fix the 64-bit build for you all now! |
Gotta love how github auto-closes issues like that. :P Just post here again if you keep on having trouble, I'm going to leave this closed for now |
Perfectly builds 64bit version, no problem at all. |
It is also working for me. Thanks! |
Getting the following messages
version.jl
fatal: No names found, cannot describe anything.
linalg.jl
Warning: replacing module Base
version.jl
fatal: No names found, cannot describe anything.
Similar issue was stated here JuliaLang/julia#2747
Building Julia outside of the homebrew environment does not produce such errors.
Despite the above fatal errors Julia builds but running the test
$brew test -v julia
produces the following output
/usr/local/Cellar/julia/HEAD/bin/julia runtests.jl all
From worker 4: * numbers
From worker 2: * core
From worker 9: * remote
From worker 5: * strings
From worker 7: * corelib
From worker 8: * hashing
From worker 6: * unicode
From worker 3: * keywordargs
From worker 9: * iostring
From worker 3: * arrayops
From worker 9: * linalg
From worker 6: * blas
From worker 8: * fft
From worker 2: * dsp
From worker 5: * sparse
From worker 6: * bitarray
From worker 7: * random
From worker 7: * math
From worker 8: * functional
From worker 7: * bigint
From worker 5: * sorting
From worker 8: * statistics
From worker 2: * spawn
From worker 7: * parallel
From worker 2: [stdio passthrough ok]
From worker 2: * priorityqueue
From worker 3: * arpack
From worker 2: * bigfloat
From worker 8: * file
From worker 7: * perf
From worker 8: * suitesparse
From worker 4: * version
Warning: Possible conflict in library symbol cholmod_start
Warning: Possible conflict in library symbol cholmod_sort
Warning: Possible conflict in library symbol cholmod_norm_sparse
Warning: Possible conflict in library symbol cholmod_check_sparse
Warning: Possible conflict in library symbol cholmod_sdmult
Warning: Possible conflict in library symbol cholmod_analyze
Warning: Possible conflict in library symbol cholmod_factorize
From worker 8: exception on 8: ERROR: type c_CholmodFactor has no field Perm
From worker 8: in CholmodFactor at linalg/cholmod.jl:243
From worker 8: in cholfact at linalg/cholmod.jl:646
From worker 8: in cholfact at linalg/cholmod.jl:880
From worker 8: in runtests at /usr/local/Cellar/julia/HEAD/share/julia/test/testdefs.jl:5
From worker 8: in anonymous at multi.jl:416
From worker 8: at suitesparse.jl:121
ERROR: type c_CholmodFactor has no field Perm
at suitesparse.jl:121
at /usr/local/Cellar/julia/HEAD/share/julia/test/runtests.jl:18
Error: julia: failed
The text was updated successfully, but these errors were encountered: