On 13/04/2011 20:50, Karel Gardas wrote:
On any builder you can see all the logs from the build root page. In
that particular case it is:
http://darcs.haskell.org/ghcBuilder/builders/kgardas-opensolaris-x86-head/190.html
for the output of all tests, just click on "10: testing".
I see that gcc is complaining about the lack of a final newline in the
stub C file we generate. I'll see if I can fix that.
I also notice that you have the standard set of failures associated with
not having a UTF-8 locale. The testsuite driver tries to set the locale
to UTF-8, but this probably isn't working on your system for some
reason. You might want to look at the code in
testsuite/driver/runtests.py, search for "UTF8" and see if you can get
that to work properly.
Cheers,
Simon
Anyway, since I think all the builders' builds are running from clean
source tree I'm also afraid that this is not only solaris issue, but
also the issue on freebsd. See for example x86 results here:
http://darcs.haskell.org/ghcBuilder/builders/pgj/330.html
Thanks,
Karel
On 04/13/11 06:37 PM, Edward Z. Yang wrote:
In that case, these may be legit regressions for Solaris. In particular,
it looks like the previously failing driver programs are no longer
failing.
It would be useful if we could get output from the failing test programs;
is that logged anywhere?
Edward
Excerpts from Karel Gardas's message of Wed Apr 13 11:14:20 -0400 2011:
Unfortunately cleaning the build tree does not help. At least I've
completely rm -rf tempbuild subdirectory on builder and still I get 39
unexpected failures, see:
http://darcs.haskell.org/ghcBuilder/builders/kgardas-opensolaris-x86-head/190/11.html
Thanks,
Karel
On 04/13/11 01:52 PM, Edward Z. Yang wrote:
I suggest cleaning the build trees and letting them try again.
I think we fixed the first set in the test suite, and the second set
is a problem
with build dependencies and the RTS.
Edward
Excerpts from Karel Gardas's message of Wed Apr 13 07:11:25 -0400 2011:
Edward,
if I recall correctly you have hinted on cvs-ghc@ mailing list that
you
still need to solve Txxxx and driverYYYY failures caused by some of
your
patches. I'm curious is it still the case? If not, then I would
like to
bring your attention to freebsd[3] and opensolaris[4] based builders
where such failures are still presented and were not presented at the
beginning of April. For example speaking about opensolaris builder
which
is run on my workstation and I care about it. The builder runs with 12
unexpected failures for a long time till April 5[1]. Then on April 6
there was a build failure due to some issue in makefiles and then on
April 7 driverXXXX start to fail:
driver016(normal)
driver017(normal)
driver028(normal)
driver035(normal)
driver045(normal)
contributing to total 17 unexpected failures and finally on April 9 we
got to 38 unexpected failures[2]. Very similar pattern you can also
observe on FreeBSD based HEAD builder [3] except that the number of
failures in total is a little bit lower.
I'm asking you because of your hint on mailing list and since I'm not
able to judge what patch exactly causes the failures yet. Do you think
that's some of yours or shall I report those new failures to ghc trac?
Thanks!
Karel
[1]:
http://darcs.haskell.org/ghcBuilder/builders/kgardas-opensolaris-x86-head/183/11.html
[2]:
http://darcs.haskell.org/ghcBuilder/builders/kgardas-opensolaris-x86-head/185/11.html
[3]: http://darcs.haskell.org/ghcBuilder/builders/pgj/
[4]:
http://darcs.haskell.org/ghcBuilder/builders/kgardas-opensolaris-x86-head/
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
