Please stabilize pytables-2.0.2. It has been fairly stable and is numarray free. src_test should do it. Thanks.
Ok, this package has incomplete dependencies. With sci-libs/hdf5-1.6.4(stable) the tests fail horribly. Also this message is given: *Warning*: HDF5 version is lower than recommended: 1.6.4 < 1.6.5 However, with sci-libs/hdf5-1.6.5 pytables' test suite works perfectly. So this patch: Index: pytables-2.0.2.ebuild =================================================================== RCS file: /var/cvsroot/gentoo-x86/dev-python/pytables/pytables-2.0.2.ebuild,v retrieving revision 1.2 diff -u -b -B -r1.2 pytables-2.0.2.ebuild --- pytables-2.0.2.ebuild 23 Feb 2008 20:22:58 -0000 1.2 +++ pytables-2.0.2.ebuild 23 May 2008 20:45:46 -0000 @@ -15,7 +15,7 @@ LICENSE="BSD" IUSE="doc examples" -DEPEND="sci-libs/hdf5 +DEPEND=">=sci-libs/hdf5-1.6.5 >=dev-python/numpy-1.0.3 dev-libs/lzo:2 app-arch/bzip2" should fix the problem.
ia64 is not stable
(In reply to comment #1) > Ok, this package has incomplete dependencies. Python team, can we raise the dependency and stabilise > 1.6.5 of HDF5?
(In reply to comment #1) > *Warning*: HDF5 version is lower than recommended: 1.6.4 < 1.6.5 > > However, with sci-libs/hdf5-1.6.5 pytables' test suite works perfectly. So this > patch: sci team is not able to mark something more recent stable. So we either wait until they sorted things out and come again or disable tests for the time being. Python team?
Even if hdf-1.6.5 is not ready to go stable yet, that doesn't stop pytables from depending on it correctly. Python? Could you fix this dependency problem?
I fixed the sci-libs/hdf5 dependencies and dropped ia64 from pytables 2.0.2 because it only has keywords for <=hdf5-1.6.2 Also bumped to pytables-2.0.3
Any news on this one?
Please reCC amd64 when there is something for us to do.
(In reply to comment #8) > Please reCC amd64 when there is something for us to do. > dito for ppc
please readd x86 when we can do something here.
Created attachment 195714 [details] latest beta called -2.2b1 - Jun 23, 2009 This is the new beta released yesterday. I renamed it 2.1.99 to conincide with gentoo naming conventions. I asked Francesc (pytables lead developer) to adapt a more conventional naming schema to avoid packaging difficulties.
pytables 2.1 is fine for me and also fixes failures with Python 2.5, we should revisit here.
Please stabilize dev-python/pytables-2.1.
This needs a newer hdf5, which sees no stabilisation progress in bug 245715.
Testing on x86: Builds fine and passes tests: $ python Python 2.6.4 (r264:75706, Dec 7 2009, 11:46:22) [GCC 4.3.4] on linux2 Type "help", "copyright", "credits" or "license" for more information. Welcome to rlcompleter2 0.95 for nice experiences hit <tab> multiple times >>> import tables >>> tables.test() -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= [ . . . ] ---------------------------------------------------------------------- Ran 6593 tests in 68.550s OK >>>
Created attachment 216896 [details] build.log (In reply to comment #15) > Testing on x86: Builds fine and passes tests: It fails for me still. For tests it needs a newer dev-python/numarray (1.5.2 is recommeneded), so only the ones marked in my build.log are still broken.
The tests failed because you had numarray and numeric installed. I now added a patch which skip these tests with the obsolete packages
Still it fails. There is only one error left (see below), if no numarray is installed, only numeric, so maybe a USE=test with blockers on numarray should be chosen. ====================================================================== ERROR: None (tables.netcdf3.tests.test_netcdf3.NetCDFFileTestCase2) ---------------------------------------------------------------------- Traceback (most recent call last): File "/var/tmp/portage/dev-python/pytables-2.1.2/work/tables-2.1.2/build/lib.linux-i686-2.6/tables/tests/common.py", line 248, in newmethod return oldmethod(self, *args, **kwargs) File "/var/tmp/portage/dev-python/pytables-2.1.2/work/tables-2.1.2/build/lib.linux-i686-2.6/tables/netcdf3/tests/test_netcdf3.py", line 168, in test_h5tonc nobjects, nbytes = f.nctoh5(self.filenc) File "/var/tmp/portage/dev-python/pytables-2.1.2/work/tables-2.1.2/build/lib.linux-i686-2.6/tables/netcdf3/__init__.py", line 476, in nctoh5 var = self.createVariable(varname,datatype,ncvar.dimensions,least_significant_digit=lsd,filters=filters) File "/var/tmp/portage/dev-python/pytables-2.1.2/work/tables-2.1.2/build/lib.linux-i686-2.6/tables/netcdf3/__init__.py", line 292, in createVariable var = NetCDFVariable(varname,self,datatype,dimensions,least_significant_digit=least_significant_digit,expectedsize=expectedsize,filters=filters) File "/var/tmp/portage/dev-python/pytables-2.1.2/work/tables-2.1.2/build/lib.linux-i686-2.6/tables/netcdf3/__init__.py", line 579, in __init__ raise ValueError, 'datatype must be one of %s'%_netcdftype_dict.keys() ValueError: datatype must be one of ['1', 's', 'd', 'f', 'i', 'c', 'F', 'l', 'D'] ---------------------------------------------------------------------- Ran 6560 tests in 53.848s
(In reply to comment #18) > Still it fails. There is only one error left (see below), if no numarray is > installed, only numeric, so maybe a USE=test with blockers on numarray should > be chosen. Actually no need of blockers, I removed all the numeric and numarray tests. It should only use numpy now. > ====================================================================== > ERROR: None (tables.netcdf3.tests.test_netcdf3.NetCDFFileTestCase2) > ---------------------------------------------------------------------- I can't reproduce this one on amd64. Anyone with x86 would like to try?
(In reply to comment #19) > (In reply to comment #18) > > Still it fails. There is only one error left (see below), if no numarray is > > installed, only numeric, so maybe a USE=test with blockers on numarray should > > be chosen. > > Actually no need of blockers, I removed all the numeric and numarray tests. It > should only use numpy now. > > > > ====================================================================== > > ERROR: None (tables.netcdf3.tests.test_netcdf3.NetCDFFileTestCase2) > > ---------------------------------------------------------------------- > > I can't reproduce this one on amd64. Anyone with x86 would like to try? (In reply to comment #19) > (In reply to comment #18) > > Still it fails. There is only one error left (see below), if no numarray is > > installed, only numeric, so maybe a USE=test with blockers on numarray should > > be chosen. > > Actually no need of blockers, I removed all the numeric and numarray tests. It > should only use numpy now. > > > > ====================================================================== > > ERROR: None (tables.netcdf3.tests.test_netcdf3.NetCDFFileTestCase2) > > ---------------------------------------------------------------------- > > I can't reproduce this one on amd64. Anyone with x86 would like to try? This seem to be an isolated thing on my machine. Marking stable for x86, thanks Thomas for the work.
