Re: ghc_7.4.1-1_hurd-i386.changes ACCEPTED into unstable
Yep, at last!
We've been missing ghc support for a long time, due to ghc 6.8 having
odd races when forking & piping, and for some reason I don't remember
6.10 wouldn't compile either.
Quite a few people attempted fixing it, others tried to cross-build from
Linux: lately Svante and I spent quite some time on it, only to realize
that there are serious issues in the bootstrap process (linux-i386
bootstrap itself bugs out).
And then I tried again to use the old ghc 6.6 and run ./configure &&
make in ghc 6.10. Which worked partially, but enough to get a compiler,
which 6.12 happily compiled completely with. From then, as Svante was
advised by upstream, 7.0.4, and now 7.4, which got ACCEPTED.
My guess is that it's thanks to piling fix after fix that 6.10
eventually managed to build fine (posix signals, bugs in locks, etc.)
That'll result in the buildds being quite busy in the coming days: there
are a few hundred haskell packages to build, see the 5% drop on the
buildd graph due to the recent haskell binNMUs in the other archs:
https://buildd.debian.org/stats/graph-week-big.png
Really a great piece of news, thanks everybody who have tried to fix the
issue.
Samuel
Reply to: