Hi Gilles, Il 30/10/2016 20:57, Gilles Filippini ha scritto: > Antonio Valentino a écrit le 30/10/2016 à 18:16 : >> Hi Gilles, >> >> Il 29/10/2016 17:17, Gilles Filippini ha scritto: >>> On 2016-10-25 09:42, Jerome Kieffer wrote: >>>> On Tue, 25 Oct 2016 08:13:43 +0100 >>>> Ghislain Vaillant <ghisvail@gmail.com> wrote: >>>> >>>>> >>>>> What about h5py ? Have you tried it with hdf5 v1.10 ? >>>>> >>>> >>>> There are many things going around pytables, h5py and hdf5: >>>> >>>> * h5py (2.6) supports the new features of hdf5 1.10 (SWMR mainly) >>>> >>>> * pytable is useful but no more heavily developed. It has been agreed >>>> with the dev of h5py to make pytable depend on h5py to have only 1 >>>> wrapper for hdf5 in python and pytables will just offer a different >>>> API. >>>> >>>> This said, I don't know how advanced this migation. >>> >>> This option was first discussed at SciPy 2015 [1], then officially >>> announced after SciPy 2016 [2]. >>> >>> [1] https://groups.google.com/forum/#!topic/pytables-users/GBVkvGexSag >>> [2] https://groups.google.com/forum/#!topic/h5py/IAaiyHOiJKE >>> >>> Unfortunaly it is far from achieved and the dedicated 'pt4' branch of >>> the pytables repo is stalled since late august. Thus the 'pytables overs >>> h5py' option seems dead for Stretch release. >> >> Yes I can confirm that pt4 is not an option for Stretch. >> The PyTable project is currently moving very slowly due to lack of time >> of the maintainers (including me). >> >>> This left us with the 'embed hdf5 1.8.16 into the pytable package' option. >>> >>> Antonio, any thought about this? >> >> In theory pytables should be able to work with HDF5 1.10 using the old >> 1.8 API. >> Unfortunately it seems that there is some compatibility issue (see also >> [1]) that none of the PyTables maintainers had the time to fix. >> IMHO a patch for PyTables v3.3 to fix compatibility with HDF5 1.10 would >> be a better option if someone has some time to dedicate to it. > > The HDF5 1.10 package in experimental was built with > "--with-default-api-version=v18". But there is one major difference with > 1.8.x releases: hid_t is now 64 bit integer instead of standard int. > > I gave a try patching pytables v3.3 in some obvious places, with no > success at all. Any help on this side would be greatly appreciated. OK, I think I have fixed all main issues with HDF5 1.10. I'm currently testing my patches and I will open a PR upstream ASAP (in an hour I hope). >> I would take the embedded hdf5 1.8.16 option as a fallback solution also >> because it could cause non trivial issues if someone uses pytables (with >> hdf51.8.16) and h5py (with hdf5 1.10) in the same program. >> >> Do we have a timeline for this? > > As I understand it, the deadline is the Stretch transition freeze: 5th > of november 2016. oh, it is very close Do you have some work already done for pytables v3.3 packaging? >> [1] https://github.com/PyTables/PyTables/issues/545 > > Thanks, > > _g. > > -- Antonio Valentino
Attachment:
signature.asc
Description: OpenPGP digital signature