I can't reproduce the problem. julia> versioninfo() Julia Version 0.5.0 Commit 3c9d753 (2016-09-19 18:14 UTC) Platform Info: System: Linux (x86_64-pc-linux-gnu) CPU: Intel(R) Core(TM) i5-3320M CPU @ 2.60GHz WORD_SIZE: 64 BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Sandybridge) LAPACK: libopenblas64_ LIBM: libopenlibm LLVM: libLLVM-3.7.1 (ORCJIT, ivybridge)
-- mb On Mon, Oct 3, 2016 at 12:46 PM, Fabrice Collard <[email protected]> wrote: > Hello everyone, > > I recently upgraded to Julia 5.0, and experienced something really weird. > Assume I generate a matrix X of random numbers, such that > > X=randn(10,100,4); > > and let me build the median of, say, the first element of the table along > its first dimension > > tmp=median(X[1,:,:],1) > > This works w/o any problem, and i can similarly build > > tmp=median(X[2,:,:],1) or tmp=median(X[5,:,:],1) > > But now let's use a loop to do it for each element > > for i in 1:10 > tmp=median(X[i,:,:],1) > println(tmp) > end > > Then I get the following message > > LoadError: AssertionError: frame.inferred > while loading In[187], in expression starting on line 2 > > in typeinf_ext(::LambdaInfo) at ./inference.jl:1645 > > > > where line 2 actually refers to the line starting by for. I must do something > wrong, but I cannot see it. Any help would be greatly appreciated. > > > Thanks a lot > > > Fabrice > >
