Hi all,

As anticipated last week, I spent a lot of this week working on the
build system. Mostly this was resurrecting UNIX binary distributions,
which now work (although there are doubtless some bits still missing).
Unfortunately, we don't normally know whereabouts in the tree we are,
e.g. when make is in utils/hsc2hs then it knows that to get to the top
you go to "../..", but it doesn't know how to get /from/ the top back
/to/ utils/hsc2hs. I've therefore passed a WHERE_AM_I variable around
when making bindists; not particularly pretty, but it works.

I've also been back and forth with Roman a bit about patches to simplify
the way we build dph; hopefully we'll have them applied soon. Plus I've
made various little build system fixes, e.g. utils/pwd now gets cleaned,
"make clean stage=2" works in compiler/ and ghc/, etc.

I spent another chunk of time trying to untangle the knot at the bottom
of the base package, and finished off the proposal to split off more
packages from base. I'll leave further untangling until after the
proposal has finished, as that'll make life a little easier.

I've reviewed and applied a large number of patches, mostly GHC API
documentation and minor refactoring patches from Max. I've also added a
new OS X buildbot slave c/o Malcolm; we have an offer of another from
Thorkil, but I haven't got to it yet.

Between the dph build system phonecall, the #ghc meeting, and other
discussions on IRC and the lists (e.g. problems building GHC, VCSs,
extensible exceptions in Cabal and nhc, CPP symbols for Cabal package
versions) I've also spent quite a bit of this week communicating with
various people.

Plus various little things, e.g. some bug triage, fixing extensible
exceptions problems in extralibs, fixing a problem with our darcs
posthooks (we had 3 posthooks, but only the first was getting; I've now
merged them into a single commandline so they all get run).

Not a good bug fixing week - I think the only bug I closed was #2509,
which Judah coincidentally filed while I was already validating a fix -
but we expected that with the plan to spend time on the base split-up
and bindist resurrection. I still have wiki build system docs to update,
and bootstrapping from HC files to resurrect, but I will try to get a
few more bugs closed this week coming as well.


Thanks
Ian

_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to