Package: bitshuffle
Followup-For: Bug #955456
Incidentally, Thorsten, another way of handling h5py and bitshuffle's
build against hfd5-mpi is to explicitly access h5py-mpi.
That is, patch out bitshuffle to replace
import h5py
with
import h5py._debian_h5py_mpi as h5py
(in bitshuffle/h5.pyx
On 2020-04-06 20:23, Gilles Filippini wrote:
Drew Parsons a écrit le 06/04/2020 à 05:08 :
Do we want to use this as a solution? Or would you prefer an
environment
variable that h5py can check to allow mpi invocation on a serial
process?
I let this decision up to you.
I think it's
Control: tags -1 + patch
Gilles Filippini a écrit le 06/04/2020 à 14:23 :
> Drew Parsons a écrit le 06/04/2020 à 05:08 :
>> On 2020-04-06 09:56, Drew Parsons wrote:
>>> On 2020-04-06 01:48, Gilles Filippini wrote:
Drew Parsons a écrit le 05/04/2020 à 18:57 :
>
> Another option is to
Drew Parsons a écrit le 06/04/2020 à 05:08 :
> On 2020-04-06 09:56, Drew Parsons wrote:
>> On 2020-04-06 01:48, Gilles Filippini wrote:
>>> Drew Parsons a écrit le 05/04/2020 à 18:57 :
Another option is to create an environment variable to force h5py to
load the mpi version even
On 2020-04-06 09:56, Drew Parsons wrote:
On 2020-04-06 01:48, Gilles Filippini wrote:
Drew Parsons a écrit le 05/04/2020 à 18:57 :
Another option is to create an environment variable to force h5py to
load the mpi version even when run in a serial environment without
mpirun. Easy enough to set
On 2020-04-06 01:48, Gilles Filippini wrote:
Drew Parsons a écrit le 05/04/2020 à 18:57 :
Another option is to create an environment variable to force h5py to
load the mpi version even when run in a serial environment without
mpirun. Easy enough to set up, though I'm interested to see if
Drew Parsons a écrit le 05/04/2020 à 18:57 :
> On 2020-04-05 22:34, Gilles Filippini wrote:
>> I suspect a mismatch between mpi and serial build-depdencies:
>>
>> python3-h5py-serial + libhdf5-dev => OK
>> python3-h5py-serial + libhdf5-openmpi-dev => KO
>> python3-h5py-mpi + libhdf5-dev
On 2020-04-05 22:34, Gilles Filippini wrote:
I suspect a mismatch between mpi and serial build-depdencies:
python3-h5py-serial + libhdf5-dev => OK
python3-h5py-serial + libhdf5-openmpi-dev => KO
python3-h5py-mpi + libhdf5-dev => KO
python3-h5py-mpi +
Drew Parsons a écrit le 05/04/2020 à 15:28 :
> On 2020-04-05 20:38, Gilles Filippini wrote:
>>>
>>> But it must be something else from these new h5py upstream patches
>>> that's leading to any other bitshuffle errors (the ones apart from the
>>> file-not-found error).
>>
>> Nope. Seems related to
On 2020-04-05 20:38, Gilles Filippini wrote:
But it must be something else from these new h5py upstream patches
that's leading to any other bitshuffle errors (the ones apart from the
file-not-found error).
Nope. Seems related to the h5py 'serial' flavour only. It appears for
the first time
Drew Parsons a écrit le 03/04/2020 à 15:08 :
> On 2020-04-03 20:13, Gilles Filippini wrote:
>>
>> This is not related to the ongoing hdf5 transition, but to the recent
>> uploads of h5py >= 2.10.0-3. Before triggering the hdf5 transition I've
>> checked that bitshuffle builds successfully. It was
On 2020-04-03 20:13, Gilles Filippini wrote:
This is not related to the ongoing hdf5 transition, but to the recent
uploads of h5py >= 2.10.0-3. Before triggering the hdf5 transition I've
checked that bitshuffle builds successfully. It was against h5py
2.10.0-2 by then.
The change in
Hi,
On Wed, 01 Apr 2020 11:47:33 +0800 Drew Parsons wrote:
> Package: bitshuffle
> Version: 0.3.5-3
> Severity: serious
> Justification: FTBFS
> Control: block 954654 by -1
>
> bitshuffle fails to build against hdf5: 1.10.6+repack-1, because
> tmp_test_filters.h5 does not exist
>
> OSError:
Package: bitshuffle
Version: 0.3.5-3
Severity: serious
Justification: FTBFS
Control: block 954654 by -1
bitshuffle fails to build against hdf5: 1.10.6+repack-1, because
tmp_test_filters.h5 does not exist
OSError: Unable to open file (unable to open file: name =
'tmp_test_filters.h5', errno = 2,
14 matches
Mail list logo