[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Is There Anybody Out There?



Hi Christoph.

I made some test builds with OpenJDK-7 on kfreebsd-amd64.

The old error was reproducible with sbuild's configuration you gave me (on fasch server) and I tried to find out which sbuild parameter is problematic for openjdk-7, but the last week something happened to kfreebsd-amd64 (unstable) and the build process of openjdk-7 gave another error even with dpkg-buildpackage which worked before. The log file is here - http://os.oles.biz/openjdk/openjdk-7_7~u3-2.1-3_kfreebsd-amd64.build.gz

Best regards
Georgi

On 03/20/2012 12:04 PM, Christoph Egger wrote:
Hi all!

   I am seeing rather urgent kfreebsd problems unsolved for some time now
and I definitely won't manage to solve these by myself. So hopefully
some people reading here also want to make kfreebsd+wheezy happening and
can free some time for it.

  - #654783 is now open for more than 2 months. Blocking python builds
    for approximately that time and now blocking transitions (v-sim build
    for netcdf transition)

  - ruby testsuite is disabled -- we're running blind there

  - Both of these test suites fail due to problems also visible outside
    the test suite. llvm's testsuite (written in python) is special-cased
    on kfreebsd and forced to run single-threaded there, mkvtoolnix'
    build-system is written in ruby and isn't completing unless forced to
    -j1

  - Despite quite some effort from the maintainer (thanks!) openjdk-7
    still doesn't build on the autobuilders and an increasing portion of
    java software doesn't work with gcj. Admittedly, this will probably
    need attention from my side as it's somewhat related to buildds, but
    I'm open for any input and help!

   At least for the python testsuite not much knowledge is needed at
all. Just building python2.7 source package untill it fails and looking
at the source of the failing test should reveal -- in most cases -- some
short series of thread+mostly IO patterns one needs to convert to C, see
if it still fails and report as bug against eglibc.

Regards

     Christoph



Reply to: