Julia has been on my radar for quite some time and I’ve had a couple of goes at installing it, but never quite got there. Decided to have another bash using the Arch Linux AUR package. Found one stumbling block in that the julia-git package has a dependency on librmath, itself a AUR package, that unfortunately conflicts with R. That’s a deal breaker for me, since I want R installed. Had a look into it and it seems Julia uses a patched version of RMath anyway and so it makes no sense to use the AUR package. Altered the PKGBUILD script and set USE_SYSTEM_RMATH=0 and “Hey presto!” it built and installed.

Now I just need to get around to using it.

Ideally I’d also like to get this built and installed on NetBSD, but as with Rust, I think my first stumbling block is probably LLVM: I can’t get the pkgsrc version to build.

[EDIT: 2014-02-11] I’ve given up on Julia again for now. The AUR package is great because it finally enabled me to get Julia installed and running, but I fell at the second hurdle as far as using it for anything useful. I thought I’d try to porting a nested bootstrap from R to Julia, but the DataFrame package gives me a whole bunch of errors. First these when using DataFrames:

Warning: New definition 
	copy!(AbstractDataArray{T,N},Any) at /home/me/.julia/v0.3/DataArrays/src/abstractdataarray.jl:48
is ambiguous with: 
	copy!(AbstractArray{T,3},AbstractArray{T,3}) at multidimensional.jl:142.
To fix, define 
before the new definition.

Followed by loads like these:

WARNING: deprecated syntax "x[i:]" at /home/me/.julia/v0.3/DataFrames/src/formula.jl:58.
Use "x[i:end]" instead.

And then, when trying to actually import data with readtable:

ERROR: no method readtable!(ParsedCSV, IOStream, Int32, ParseOptions{ASCIIString,UTF8String})
 in readtable at /home/me/.julia/v0.3/DataFrames/src/io.jl:648

I’m going to have a play with Rust instead since I noticed a binary is now available for Arch.