So let's switch to compiling Setup scripts once more, to give at least
a few more hs packages to be built on i386, until I figure out a fix
for rts/Linker.c.
Unfortunately, the current breakage also affects template haskell,
which still leads to a lot of unbuildable ports, for example
devel/hs-vector.
old-time and time. After comparing all the .hi files before and
after (using ghc --show-iface) I guess this changes had been triggered
by the suseconds_t change some weeks ago (my previous ghc build was
from october 23rd, so just a little bit before that change).
People following -current and using Haskell for serious work should
wait with updates until the new ghc and hs-packages are built and
copied out to the mirrors. Any Haskell additional libraries built
manually (i.e. using cabal something) should be rebuilt. If in
doubt, run ghc-pkg check to see if anything is broken for you.
the 64 bit time_t switch.
If anyone wants to beat me, grab ~kili/ghc-bootstrap.sh from cvs
(or http://openbsd.dead-parrot.de/ghc-bootstrap.sh), *read and try
to understand it*, downgrade the lang/ghc port to the last 7.4.
version (for example, cvs -q up -dP -D 2013-07-09) and then run
ghc-boostrap.sh (but be aware that you'll need a lot of time, disk
space, memory, and alcohol).
This may work or not (it worked for 5.3-current back in may). If
it doesn't work, either wait or go using some real programming
language.
is poisoned with code from some very early bootstrapper, which
causes failure when bootstrapping from 6.6 via 6.10, 6.12, 7.0
to 7.4. So just overwrite it with /bin/pwd.
While here, remove the commands that attempt to rename
*-x86_64-* to *-amd64-*, because it currently doesn't
work anyways (VERSION_DATE is ignored by the bootstrapper
build).
(e.g. large Fibonacci numbers): link GHC with its bundled gmp library
that has customized memory allocation functions.
While here, fix two other problems as well:
- undefined references to 'ffi_call_unix64', 'ffi_closure_unix64'
on amd64
- unknown symbol '__guard_local' (caused by gcc's switch from __guard
to __guard_local)
Joint work with kili@
OK kili@
may detect it but dpb(1) may remove it afterwards, which causes the
build to fail (spotted by naddy).
This *could* be done simpler (by adding TarCmd=/bin/tar to
CONFIGURE_ENV), but this would break the bootstrap target.
Some more dependency odditiies (dblatex, docboo*, libxslt may be
not necessary for the no_doc FLAVOR, also noticed by naddy@) will
be checked and fixed later.
Bulk builders suffer without this, and we already have a sligt
idea what's going wrong (new thread receives a signal before
__tfork returns and then it fails in sigreturn).
I hope to find some time to work on a regression test for this
problem (which isn't related at all to ghc).
Unfortunately, we can't completely remove or skip include/execvpe.h
and cbits/execvpe.c, because they also contain pPrPr_disableITimers(),
which is called from the Haskell code.
- Switch back to GhcThreaded=YES (default). It's stupid, it's not
required by GHC (IIRC it was required a long time ago back when
ghci used libreadline), but it may help finding rthread bugs.
Until now, I did run four complete builds of GHC and all Haskell
ports with dpb without any problems.
itself using the non-threaded rts. This doesn't fix the real problem,
but it makes it possible to build hs-* ports without filling up
dpb's build slots with stuck ghc processes.
ok naddy@ espie@
Note: I'll re-run the testsuite with this diff and hope to get some
reasonable results that may help us to identify the real problem.
matching the requirements of haskell-platform-2011.4.0.0.
Moving the xhtml library back to a separate port (www/hs-xhtml)
would be nice but it causes too much headache (like dependency
cycles with devel/haddock).
