tech-toolchain archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: put fortran back in base?


I've brought it up and made it seem a bit dire, but it really isn't the case.
You can see an example of how severe the fallout of not having fortran is here:

it's just 235 packages broken from not having lang/g95.

I think pkgsrc fortran is sufficient, although I would love to get rid of the
default choice to g95.

It doesn't build on linux or on arm for some really vague "compiler is totally
fucked" reasons, which sound hard to fix (PR 51198, PR 51402). Also MIPS, but
that issue at least sounds workable :-)

When it does build, it may hit some internal compiler errors, like kamil had
when trying benchmarks/benchfft.

This is reasonable and kinda sily to fix - it's a 10 year old version of GCC.
It makes more sense to try to deal with versions of GCC that are supported..

It's a much bigger pkgsrc problem than a netbsd problem. all fortran variants
build fine on netbsd/x86, and most users don't run into any issues.

perhaps we can continue discussing why opting to use gfortran in pkgsrc is
problematic, joerg mentioned it's a problem to have much newer gcc, but I feel
like it can be dealt with.

Home | Main Index | Thread Index | Old Index