Ah -- so bindists are doing more than relocating the files.
I have transplanted gcc to my environment and it is going much better. I now
get (for WAY=optc) the following:
OVERALL SUMMARY for test run started at Sun Sep 11 18:45:07 BST 2011
2612 total tests, which gave rise to
14419 test cases, of which
0 caused framework failures
13106 were skipped
1280 expected passes
22 expected failures
0 unexpected passes
11 unexpected failures
Unexpected failures:
4038(optc)
Chan001(optc)
MVar001(optc)
QSem001(optc)
QSemN001(optc)
SampleVar001(optc)
T3087(optc)
cgrun025(optc)
maessen_hashtab(optc)
regex003(optc)
space_leak_001(optc)
I think I am only approximating the build environment with 2011.2.0.1:
* 5 of the tests (Chan-001-->SampleVar001) fail to find
Debug.QuickCheck
* a further 3 static failures (T3087, maessen_hashtab & regex003)
* cgrun025 bombs out with an irrefutable pattern failure
* 4038 fails to provide any output
* I couldn't see what was wrong with space_leak_001
Is there any bindist documentation -- I have been trying to locate it
without success. I have made a bindist but what I get doesn't seem to
overlap with the contents of the distributed bindist tarballs.
Also my experiments indicate that when you tell ./configure where to get gcc
it still runs /usr/bin/gcc, as does the build. In both instances, if
/usr/bin/gcc is missing the configuration/build fails. (This is with
ghc-7.0.3.) Is this right or am I mistaken?
Thanks!
Chris
-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Ian Lynagh
Sent: 11 September 2011 13:39
To: [email protected]
Subject: Re: 7.0.3 test suite giving -fno-toplevel-reorder on cc1
On Sat, Sep 10, 2011 at 10:58:16PM +0100, Chris Dornan wrote:
>
> (One of the things I am concerned about is mismatches between the
> gcc-version used to build and the version installed on the target
> system - they are quite different in this case.)
Do you mean you've moved a GHC tree from one machine to another?
If you instead make a bindist on the machine you built it on (make
binary-dist) and then install the bindist on the other machine, then the
right gcc flags should be used.
Thanks
Ian
_______________________________________________
Cvs-ghc mailing list
<mailto:[email protected]> [email protected]
<http://www.haskell.org/mailman/listinfo/cvs-ghc>
http://www.haskell.org/mailman/listinfo/cvs-ghc
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
