Hi,
I have uploaded 7.10.3rc1 to experimental.
The good news:
GHCi on armhf works better. This affects doctests: Previously we had
Running 1 test suites...
Test suite doctests: RUNNING...
doctests: user error (Language.Haskell.GhciWrapper.close:
Interpreter exited with an error (ExitFailure (-11)))
Test suite doctests: FAIL
Test suite logged to: dist-ghc/test/bytes-0.15.0.1-doctests.log
and now we have
Running 1 test suites...
Test suite doctests: RUNNING...
Examples: 19 Tried: 19 Errors: 0 Failures: 0
Test suite doctests: PASS
Test suite logged to: dist-ghc/test/bytes-0.15.0.1-doctests.log
1 of 1 test suites (1 of 1 test cases) passed.
Thanks to Erik de Castro Lopo for fixing this upstream.
Cabal runs the test suites without relying on a threaded runtime. I
cooked up a patch (https://github.com/haskell/cabal/pull/2913) to Cabal
and included it in our GHC package. So instead of
Running 1 test suites...
Test suite active-tests: RUNNING...
E: Caught signal ‘Terminated’: terminating immediately
we now have
Running 1 test suites...
Test suite active-tests: RUNNING...
(0 tests) (1
test) (2 tests) (3
tests) (4
tests) (5
tests) (6
tests) (7
ests) (197
tests) (198
tests) (199
tests) atTime/fun
+++ OK, passed 200 tests.
[..lots of output..]
Test suite active-tests: PASS
Test suite logged to: dist-ghc/test/active-0.2.0.4-active-tests.log
(this is on s390x). The bug has (strangely) also affected architectures
with threading, such as powerpc, but at least it is not affecting us
greatly.
The bad news:
7.10.3 should have fixed
ghc: panic! (the 'impossible' happened)
(GHC version 7.10.2.20151030 for powerpc-unknown-linux):
PPC.Ppr: Shift by 32 bits is not allowed.
on powerpc (https://ghc.haskell.org/trac/ghc/ticket/10870), but it did
not (see for example haskell-vector-alorigthms).
The above fix related to GHCi on arm broke GHC completely on armel. :-(
Both of them are release critical, and would prevent a quick upload-to-
unstable-and-then-directly-migrate-to-testing.
Recently, Erik has been the only one able and willing to fix
architecture related bugs in GHC (thanks a lot!), but we cannot expect
him to fix all of them. So this is a call for help!
Greetings,
Joachim
--
Joachim "nomeata" Breitner
Debian Developer
nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: F0FBF51F
JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata
Attachment:
signature.asc
Description: This is a digitally signed message part