| opendevreview | Cyril Roelandt proposed openstack/oslo.config master: Fix tests for Python3. https://review.opendev.org/c/openstack/oslo.config/+/805090 | 00:42 |
|---|---|---|
| *** rpittau|afk is now known as rpittau | 07:08 | |
| *** redrobot1 is now known as redrobot | 07:59 | |
| stephenfin | zzzeek: Any chance you'll have time to plough through a few more of these SQLA 2.0 changes for oslo.db before EOW? https://review.opendev.org/q/project:openstack/oslo.db+is:open+topic:sqlalchemy-20 (CC: hberaud) | 09:38 |
| stephenfin | zzzeek: Also, if you've docs on alternatives to what I'm doing here, I'd appreciate it. Don't need a solution but just enough pointers to get me going :) https://review.opendev.org/c/openstack/oslo.db/+/804775 | 09:40 |
| opendevreview | Merged openstack/oslo.config master: Fix tests for Python3. https://review.opendev.org/c/openstack/oslo.config/+/805090 | 10:29 |
| opendevreview | Merged openstack/oslo.db master: requirements: Bump sqlalchemy lower constraint https://review.opendev.org/c/openstack/oslo.db/+/804705 | 11:02 |
| ykarel | zaneb, hi | 14:21 |
| ykarel | zaneb, was looking at packaging autopage for RDO as next cliff release will need it. And found that you have copr repo for it https://copr.fedorainfracloud.org/coprs/zaneb/autopage/package/python-autopage/ | 14:23 |
| ykarel | we rebuild in RDO from fedora builds, are you also planning to add that package to Fedora rawhide? | 14:24 |
| zaneb | ykarel: yeah, I hadn't submitted a Fedora package request for it yet because I wasn't sure if that would make your life easier or harder :) | 14:24 |
| ykarel | or already in process, i didn't find bz for it though | 14:24 |
| zaneb | happy to do it if it would make it easier | 14:24 |
| ykarel | yes please we would need it, we rebuild for centos8-stream from fedora srpms | 14:25 |
| ykarel | i tried scratch build against c8-stream with your spec https://github.com/zaneb/autopage/blob/rpm/python-autopage.spec and worked fine | 14:26 |
| zaneb | ok, that also answers my other question about whether to submit the pure pyproject-rpm-macros one from https://github.com/zaneb/autopage/blob/fedora/python-autopage.spec or the combined fedora/centos one from https://github.com/zaneb/autopage/blob/rpm/python-autopage.spec :) | 14:27 |
| ykarel | yes would need combined one | 14:34 |
| zaneb | ykarel: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=1995676 | 15:24 |
| zzzeek | @stephenfin: will take a look. we got a bunch merged right? will get some more | 15:24 |
| stephenfin | we sure did :) thanks! | 15:24 |
| ykarel | zaneb, Thanks | 15:26 |
| *** rpittau is now known as rpittau|afk | 15:51 | |
| zzzeek | @stephenfin: the neutron issue im not sure how to fix right now because i can't get a simple reproducer | 16:58 |
| zzzeek | @stephenfin: basically somethign uses the session in that case, and they have to call session.commit() | 16:58 |
| zzzeek | @stephenfin: the session will "Autobegin" the first time it's used. if they are doing db operations that they need completed before that method is called then they should commit | 16:59 |
| stephenfin | Ok, great, I can work on a reproducer. Out of curiosity, what happens if I call session.commit() when auto-commit is enabled? | 16:59 |
| zzzeek | @stephenfin: but someone knowlegeable about neutron would need to show us what the methods being called are, because the unit tests at least run that method without issue | 16:59 |
| zzzeek | you probably get an error | 17:00 |
| stephenfin | Drat, I was hoping it would be a no-op /o\ | 17:00 |
| zzzeek | so the fix in ntruon would be, they set autocommit=False on their session up front, then make sure they commit() when they expected it to | 17:00 |
| stephenfin | It sounds like I won't be able to approach things this way. I'll have to deprecate autocommit first and force everyone to set it to False | 17:00 |
| stephenfin | yeah | 17:00 |
| zzzeek | stephenfin: you can check "if session.transacvtion: session.commit()" | 17:00 |
| zzzeek | but we need to find in neutron hwere theyu are currently expecting autocommit to occur and that has to be made explicit | 17:01 |
| zzzeek | neutron is the most difficult openstack project to approach | 17:01 |
| stephenfin | Uup. I'll have to lean on the neutron folks for that because I've no idea how their API and DB layers work | 17:01 |
| stephenfin | *Yup | 17:01 |
| stephenfin | I've got guys on the inside though ;) | 17:01 |
| zzzeek | also the one w/ the most problems | 17:02 |
| zzzeek | ok great | 17:02 |
| zzzeek | there is a TODO for that wrapper from kevinbenton, "TODO remove all use of this flag" | 17:02 |
| zzzeek | the wrapper there is acvtually OK, they should leave it | 17:02 |
| zzzeek | they just need to make sure commits are explicit where expected | 17:02 |
| stephenfin | ack, can do that | 17:03 |
| stephenfin | Thanks for taking a look :) | 17:05 |
| *** mtreinish_ is now known as mtreinish | 17:22 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!