Bug#800466: FTBFS: 4 failed tests with ValueError: could not broadcast input array

[Emilio Pozuelo Monfort]

On Wed, 30 Sep 2015 18:56:35 +0200 Emilio Pozuelo Monfort <pochu at debian.org> wrote:
> On Tue, 29 Sep 2015 21:28:07 +0200 Gilles Filippini <pini at debian.org> wrote:
> > On Tue, 29 Sep 2015 20:46:42 +0200 Gilles Filippini <pini at debian.org> wrote:
> > > Source: pytables
> > > Version: 3.2.1-1
> > > Severity: serious
> > > Justification: FTBFS
> > > 
> > > -----BEGIN PGP SIGNED MESSAGE-----
> > > Hash: SHA256
> > > 
> > > Hi,
> > > 
> > > pytables FTBFS on a clean amd64 sid chroot. 4 tests fail with similar
> > > error messages:
> > > ======================================================================
> > > ERROR: None (tables.tests.test_tables.RecArrayRangeTestCase)
> > > - ----------------------------------------------------------------------
> > > Traceback (most recent call last):
> > >   File "/tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-2.7/tables/tests/test_tables.py", line 2169, in test01a_range
> > >     self.check_range()
> > >   File "/tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-2.7/tables/tests/test_tables.py", line 2042, in check_range
> > >     recarray = table.read(self.start, self.stop, self.step)
> > >   File "/tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-2.7/tables/table.py", line 1965, in read
> > >     arr = self._read(start, stop, step, field, out)
> > >   File "/tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-2.7/tables/table.py", line 1887, in _read
> > >     self.row._fill_col(result, start, stop, step, field)
> > >   File "tables/tableextension.pyx", line 1272, in tables.tableextension.Row._fill_col (tables/tableextension.c:15021)
> > > ValueError: could not broadcast input array from shape (2) into shape (0)
> > 
> > This is upstream issue #481 [1] which was fixed by commit 44dba04 [2].
> > 
> > [1] <https://github.com/PyTables/PyTables/issues/481>
> > [2] <https://github.com/PyTables/PyTables/commit/44dba04d7d72f150a91553f4eb455684dfef0913.patch>
> > 
> > I've successfully tested this patch, but then a python3.5 related error occurs:
> > 
> > Ran 5734 tests in 147.360s
> > 
> > OK (skipped=42)
> > + cd /tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-3.5
> > + env PYTHONPATH=. LOCPATH=/tmp/buildd/pytables-3.2.1/tmp-locales LC_ALL=en_US.UTF-8 python3.5 tables/tests/test_all.py -vvv
> > Traceback (most recent call last):
> >   File "tables/tests/test_all.py", line 10, in <module>
> >     import tables
> >   File "/tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-3.5/tables/__init__.py", line 123, in <module>
> >     from tables.file import File, open_file, copy_file, openFile, copyFile
> >   File "/tmp/buildd/pytables-3.2.1/build/lib.linux-x86_64-3.5/tables/file.py", line 31, in <module>
> >     import numexpr
> >   File "/usr/lib/python3/dist-packages/numexpr/__init__.py", line 40, in <module>
> >     from numexpr.expressions import E
> >   File "/usr/lib/python3/dist-packages/numexpr/expressions.py", line 45, in <module>
> >     from numexpr import interpreter
> > ImportError: cannot import name 'interpreter'
> > debian/rules:58: recipe for target 'override_dh_install' failed
> > make[1]: *** [override_dh_install] Error 1
> > make[1]: Leaving directory '/tmp/buildd/pytables-3.2.1'
> > debian/rules:26: recipe for target 'binary' failed
> > make: *** [binary] Error 2
> 
> That's just because numexpr hasn't been rebuilt for the python 3.5 transition

That's happened now, so this shouldn't fail at that stage.

Emilio