15:00:15 <vishalmanchanda> #startmeeting horizon
15:00:15 <opendevmeet> Meeting started Wed Feb 22 15:00:15 2023 UTC and is due to finish in 60 minutes.  The chair is vishalmanchanda. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:15 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:15 <opendevmeet> The meeting name has been set to 'horizon'
15:01:29 <vishalmanchanda> hi, anyone around for horizon weekly meeting?
15:06:35 <vishalmanchanda> I guess, we have  no topic for discussion for today's meeting.
15:06:54 <vishalmanchanda> Just few updates from my sides, if anyone look at meeting logs
15:07:41 <vishalmanchanda> I have purpose  a patch to cut a release of horizon for antelope cycle.
15:08:27 <vishalmanchanda> I am hoping if we can merge Floating-Ip-PortForwarding feature.
15:08:30 <vishalmanchanda> https://review.opendev.org/c/openstack/horizon/+/868368
15:08:52 <vishalmanchanda> let me know if any other patch you are planning to land in antelope cycle.
15:09:46 <vishalmanchanda> rdopiera: tmazur amotoki :^^
15:10:18 <amotoki> hi
15:10:22 <amotoki> anything to discuss?
15:10:25 <vishalmanchanda> amotoki: hello
15:11:00 <vishalmanchanda> I am just asking the core team if they can review Floating-Ip-PortForwarding feature.
15:11:11 <vishalmanchanda> It's look good to me now.
15:11:21 <amotoki> re: FIP port fw patch, at least I am against the version check in the code as we cover such kind of thing in requirements.txt so far.
15:11:58 <amotoki> do we ahve any reason we need to have a version check in the code rather than requirements.txt?
15:12:25 <vishalmanchanda> amotoki: you mean this https://review.opendev.org/c/openstack/horizon/+/868368/12/openstack_dashboard/api/neutron.py#938 ?
15:12:59 <amotoki> vishalmanchanda: yes
15:13:47 <amotoki> having version requirements in multiple places would be confusing
15:13:54 <amotoki> this is the main reason
15:14:30 <vishalmanchanda> amotoki: I am not sure of that, I just mentioned in my review comment that we use microversion for cinder and nove but not sure if we can follow the same in neutron.
15:14:53 <amotoki> neutron does not have microversion mechanism
15:15:08 <vishalmanchanda> amotoki: ok.
15:15:35 <amotoki> when we support a specfic feature, we depend on a specific version of neutronclient or newer
15:15:52 <amotoki> so it is reasonable to require neutroncleint >= x.y.z
15:16:43 <amotoki> neutron API consumers need to check which neutron APi extensions aree available to check which features are avaialable
15:17:32 <vishalmanchanda> amotoki: api extension is also added by the author.
15:18:16 <amotoki> it is not surprising that a newer release of horizon depends on a newer version of neutronclient
15:18:16 <vishalmanchanda> amotoki: Could you please drop these comments on the patch about not using version check in code?
15:18:25 <vishalmanchanda> amotoki: or I can do that.
15:18:31 <amotoki> vishalmanchanda: sure
15:18:36 <vishalmanchanda> amotoki: thanks
15:19:02 <amotoki> in addition, the neturon team plans to switch python bindings for neutron from neutronclient to openstacksdk
15:19:42 <vishalmanchanda> amotoki: yeah I remember you mentioned that already in previous meeting.
15:19:57 <amotoki> i think FIP port-fw is supported by openstacksdk, but it would be nice if the author can confirm it is implemented in sdk already
15:20:14 <vishalmanchanda> amotoki: make sense.
15:20:18 <amotoki> I will also add a comment on this
15:20:27 <vishalmanchanda> amotoki: cool
15:21:17 <vishalmanchanda> amotoki: Do you have any topic to discuss?
15:21:23 <amotoki> one thing
15:21:43 <amotoki> does anyone look at https://bugs.launchpad.net/horizon/+bug/2007574 in detail?
15:22:42 <amotoki> Following a discussion in setuptools, it seems that the issue exist around all xstatic packages
15:23:24 <amotoki> it looks like that the way used in xstatic packages are discouraged now
15:23:50 <vishalmanchanda> amotoki: hmm, I didn't receive a notification about this bug. looking at the bug now.
15:25:28 <vishalmanchanda> amotoki: will go through the github discussion, thanks for bringing it here.
15:25:34 <amotoki> it was posted to openstack-discuss too
15:25:42 <amotoki> https://lists.openstack.org/pipermail/openstack-discuss/2023-February/032244.html
15:27:24 <vishalmanchanda> amotoki: sorry, I missed that mail.
15:27:44 <amotoki> there are two aspects on this
15:28:14 <amotoki> the one is taht we need to raise it as a broader discussion around xstatic
15:28:58 <amotoki> the other one is whether we need to depends on xstatic (as JS lib distributor which mean we continue to ship newer release of xstatic version of JS libs)
15:29:52 <amotoki> personally I am not so motivated on the first point, although it needs to be addressed in a short-term
15:30:25 <amotoki> at the moment kolla team pins setuptools version and it does not impact them immediately
15:30:56 <amotoki> but it should be resolved soon
15:34:08 <vishalmanchanda> amotoki: just want to confirm they are facing this issue for XStatic-Moment-Timezone pypi only?
15:34:41 <amotoki> vishalmanchanda: no. it is not limited to a specif xstatic packages
15:34:53 <amotoki> the second package to be installed would hit the issue
15:35:43 <amotoki> anyway horizon installation would fail
15:36:00 <vishalmanchanda> yeah.
15:36:41 <vishalmanchanda> Don't we have any job in our CI that check such issues?
15:38:29 <vishalmanchanda> I am investigating this bug now and let's see if found some solution.
15:38:32 <amotoki> i haven't check how these dependenies are checked
15:39:42 <amotoki> but it is not limited to openstack world so it would need more disucciosn in the xstatic world
15:40:24 <vishalmanchanda> yeah because many other people using these pypi facing the same issues.
15:41:06 <amotoki> I have some referneces
15:41:09 <amotoki> https://github.com/pypa/setuptools/issues/3824
15:41:26 <amotoki> https://github.com/pypa/setuptools/pull/2149
15:42:13 <amotoki> the first one was raised by an author who filed the horizon bug
15:42:32 <amotoki> the latter one is mentioned during the discussion in the first one
15:42:55 <vishalmanchanda> thanks for the references.
15:43:48 <amotoki> that's all from me
15:44:02 <vishalmanchanda> As we have no other topic to discuss, ending this meeting
15:44:34 <amotoki> I believe rdopiera has his opinion on the xstatic issue too
15:44:47 <amotoki> I will continue to check it too
15:45:24 <vishalmanchanda> thanks
15:45:36 <vishalmanchanda> amotoki: Also, Please drop your comment about version-checking and opensdk and any other asap https://review.opendev.org/c/openstack/horizon/+/868368
15:45:49 <vishalmanchanda> #endmeeting