On Mon, Dec 16, 2019 at 6:41 PM Alfredo Deza <adeza@xxxxxxxxxx> wrote: > > On Mon, Dec 16, 2019 at 6:37 PM Sage Weil <sweil@xxxxxxxxxx> wrote: > > > > The py3 battle goes well. A bunch of tests have been pinned to > > ubuntu nodes (because they have both py2 and py3). Including s3-tests. > > > > However... there are a bunch of rgw tests combine s3-tests (needs ubuntu > > for py2) and vlagrind (needs centos for valgrind notcmalloc build). We > > could > > > > 1- disable these tests > > 2- add a ubuntu notcmalloc builder and see if valgrind is better in bionic > > than it was in the xenial and earlier > > 3- convert s3-tests to py3 already > > > > I'm hoping/begging/pleading that someone will take a serious look at > > s3-tests and tell me how hard it is. > > > > I would love to merge the py3 change tomorrow (Tuesday). > > If we want to merge soon, then ignoring these until they are fixed > sounds like the best past forward. > > Ali mentioned he had these almost done for py3 - I wonder how much is > left, I can help out there if needed. Ali, I was looking at the python3 work in https://github.com/ceph/s3-tests/commits/wip-python-3-port How much more do you think we need there? Let me know if I can help out. > > > > > Thanks! > > sage > > > > > > > > On Mon, 16 Dec 2019, Sage Weil wrote: > > > > > I'm working on getting the test suite to run on the py3 branch, starting > > > with rados. One of the failures I've run into is that s3-tests runs with > > > python 2 instead of 3. This needs to be converted asap in order for us to > > > make the jump to python 3. > > > > > > I started with a few trivial changes here > > > > > > https://github.com/ceph/s3-tests/pull/333 > > > > > > but I don't really know what I'm doing. Can someone more familiar with > > > s3-tests and/or python take over? > > > > > > One question I had was around the s3tests vs s3tests_boto3 directories... > > > which is the one that matters, and why is the old one still around? Or > > > are they both still used and maintained? > > > > > > My other question is how the branches work. I'm guessing we should > > > actually make the change to teh master branch and then cherry-pick it to > > > ceph-master? I'm surprised how far apart master and ceph-master are. > > > > > > Thanks! > > > sage > > > _______________________________________________ > > > Dev mailing list -- dev@xxxxxxx > > > To unsubscribe send an email to dev-leave@xxxxxxx > > > > > > > > _______________________________________________ > > Dev mailing list -- dev@xxxxxxx > > To unsubscribe send an email to dev-leave@xxxxxxx > > _______________________________________________ Dev mailing list -- dev@xxxxxxx To unsubscribe send an email to dev-leave@xxxxxxx
