| *** dims_ has joined #openstack-relmgr-office | 00:53 | |
| *** dimsum__ has quit IRC | 00:56 | |
| *** dims_ has quit IRC | 01:03 | |
| *** dimsum__ has joined #openstack-relmgr-office | 01:05 | |
| *** dimsum__ has quit IRC | 01:08 | |
| *** dimsum__ has joined #openstack-relmgr-office | 01:08 | |
| *** doug-fish has joined #openstack-relmgr-office | 02:04 | |
| *** doug-fi__ has joined #openstack-relmgr-office | 02:06 | |
| *** doug-fis_ has quit IRC | 02:06 | |
| openstackgerrit | Tony Breeds proposed openstack/releases: Juno: release python-glanceclient 0.14.3 https://review.openstack.org/235967 | 02:09 |
|---|---|---|
| *** doug-fish has quit IRC | 02:09 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/releases: Oslo Releases for Week of Oct 19th 2015 https://review.openstack.org/236770 | 02:21 |
| *** dimsum__ has quit IRC | 02:24 | |
| *** dimsum__ has joined #openstack-relmgr-office | 02:24 | |
| *** dimsum__ has quit IRC | 02:51 | |
| *** dimsum__ has joined #openstack-relmgr-office | 02:51 | |
| *** dimsum__ has quit IRC | 03:30 | |
| *** dimsum__ has joined #openstack-relmgr-office | 03:31 | |
| *** dimsum__ has quit IRC | 03:31 | |
| *** dimsum__ has joined #openstack-relmgr-office | 03:31 | |
| *** dimsum__ has quit IRC | 03:37 | |
| *** dimsum__ has joined #openstack-relmgr-office | 03:38 | |
| *** dimsum__ has quit IRC | 03:57 | |
| openstackgerrit | Steve Martinelli proposed openstack/releases: Release openstackclient 1.8.0 https://review.openstack.org/236792 | 04:37 |
| openstackgerrit | Steve Martinelli proposed openstack/releases: Release openstackclient 1.8.0 https://review.openstack.org/236792 | 04:41 |
| *** stevemar_ has quit IRC | 04:51 | |
| openstackgerrit | Steve Martinelli proposed openstack/releases: release keystoneauth 1.2.0 https://review.openstack.org/236797 | 04:52 |
| *** stevemar_ has joined #openstack-relmgr-office | 04:53 | |
| openstackgerrit | Steve Martinelli proposed openstack/releases: release keystoneauth 1.2.0 https://review.openstack.org/236797 | 04:55 |
| *** dimsum__ has joined #openstack-relmgr-office | 05:54 | |
| *** dimsum__ has quit IRC | 05:59 | |
| *** stevemar_ has quit IRC | 06:00 | |
| *** stevemar_ has joined #openstack-relmgr-office | 06:01 | |
| *** stevemar_ has quit IRC | 06:04 | |
| *** dimsum__ has joined #openstack-relmgr-office | 06:56 | |
| *** dimsum__ has quit IRC | 07:00 | |
| *** jhesketh has quit IRC | 07:02 | |
| *** jhesketh has joined #openstack-relmgr-office | 07:05 | |
| *** fesp has joined #openstack-relmgr-office | 07:17 | |
| *** fesp has quit IRC | 07:36 | |
| *** stevemar_ has joined #openstack-relmgr-office | 08:02 | |
| *** stevemar_ has quit IRC | 08:06 | |
| *** dtantsur|afk is now known as dtantsur | 08:30 | |
| tonyb | What would be involved in adding a new library to the realease repo such that the workflow for tagging and pushing things just works? | 08:42 |
| tonyb | I'm thinking of yaml2ical. | 08:42 |
| tonyb | c | 08:57 |
| *** dimsum__ has joined #openstack-relmgr-office | 08:59 | |
| *** dimsum__ has quit IRC | 09:04 | |
| *** openstack has joined #openstack-relmgr-office | 09:19 | |
| *** stevemar_ has joined #openstack-relmgr-office | 10:03 | |
| *** stevemar_ has quit IRC | 10:06 | |
| *** dimsum__ has joined #openstack-relmgr-office | 10:07 | |
| *** dimsum__ is now known as dims | 10:52 | |
| *** amotoki has joined #openstack-relmgr-office | 10:55 | |
| amotoki | dhellmann: ttx: I noticed that reviews of stable/liberty branch of nova, neutron and more can be approved only by <project>-milestone team. | 10:58 |
| amotoki | Is it time to switch <project>-milestone to <project>-stable-maint team? | 10:58 |
| amotoki | or is there any policy change? | 10:58 |
| *** gordc has joined #openstack-relmgr-office | 11:21 | |
| *** doug-fish has quit IRC | 11:22 | |
| *** doug-fish has joined #openstack-relmgr-office | 11:25 | |
| *** krotscheck_ is now known as krotscheck | 11:48 | |
| dims | ttx: dhellmann: oslo releases for this week - https://review.openstack.org/#/c/236770/ | 11:52 |
| *** amotoki has quit IRC | 12:00 | |
| *** stevemar_ has joined #openstack-relmgr-office | 12:04 | |
| *** stevemar_ has quit IRC | 12:07 | |
| *** bswartz has quit IRC | 12:18 | |
| *** dims has quit IRC | 12:27 | |
| *** dims has joined #openstack-relmgr-office | 12:28 | |
| *** Kiall has quit IRC | 12:42 | |
| *** Kiall has joined #openstack-relmgr-office | 12:44 | |
| *** bswartz has joined #openstack-relmgr-office | 12:51 | |
| *** amotoki has joined #openstack-relmgr-office | 13:10 | |
| *** mestery has joined #openstack-relmgr-office | 13:24 | |
| *** superdan is now known as dansmith | 13:42 | |
| dhellmann | amotoki: we're in the process of moving over | 13:44 |
| amotoki | dhellmann: thanks for clarifying. no problem. | 13:45 |
| dhellmann | dims: ttx is out this week, but I'll take a look at that shortly | 13:47 |
| dims | dhellmann: thanks | 13:49 |
| openstackgerrit | Monty Taylor proposed openstack/releases: Release 1.9.0 of os-client-config https://review.openstack.org/236989 | 13:53 |
| *** mordred has joined #openstack-relmgr-office | 13:53 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:55 | |
| *** stevemar_ has joined #openstack-relmgr-office | 14:04 | |
| *** stevemar_ has quit IRC | 14:08 | |
| bswartz | have the ACLs been loosened for stable/liberty branches now that the release is out? | 14:18 |
| dims | bswartz: i still see "refs/heads/stable/liberty" in project-config, so i don't think so | 14:27 |
| dims | bswartz: let me put up a review | 14:28 |
| bswartz | dims: will the release team automatically update the acls to loosen them or do we need to request it? | 14:28 |
| bswartz | at this point stable/liberty should be under control of the stable-maint team and the core team, correct? | 14:29 |
| dims | bswartz: if i remember right it was one of the tasks to be done | 14:29 |
| dims | bswartz: let me get the ball rolling with a review | 14:29 |
| dims | bswartz: https://review.openstack.org/#/c/237012/ | 14:33 |
| dims | bswartz: ttx is out this week, dhellmann is around, so we'll get feedback on that soon | 14:33 |
| *** stevemar_ has joined #openstack-relmgr-office | 14:34 | |
| bswartz | dims: k thx | 14:34 |
| bswartz | dims: you change merged but the acl change doesn't seem to be in effect yet... | 15:44 |
| bswartz | your* | 15:44 |
| dims | bswartz: takes some time to propagate | 15:44 |
| dims | around 30 mins after merge is what i heard last time, we should ask on infra if doesn't | 15:45 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/releases: Release of python-manilaclient 1.4.1 https://review.openstack.org/237057 | 15:52 |
| *** armax has joined #openstack-relmgr-office | 15:52 | |
| mordred | dims, dhellmann: I know y'all is busy - but if I could bug you for https://review.openstack.org/#/c/236989/ at your convenience. I need it released so that I can land a patch to shade which should go in before I cut a 1.0 for it which I'd like to do soon because ansible 2.0 is coming out soon | 15:55 |
| mordred | (it's a lovely snake of depends) | 15:55 |
| dims | mordred: ack. i'll pick it up after the oslo meeting which starts shortly | 15:56 |
| mordred | dims: thanks! really appreciate it | 15:59 |
| *** sambetts has joined #openstack-relmgr-office | 16:26 | |
| sambetts | dhellmann: Hey | 16:26 |
| dhellmann | sambetts: hi | 16:26 |
| *** armax_ has joined #openstack-relmgr-office | 16:27 | |
| dhellmann | sambetts: so what's going on with your version numbers? | 16:27 |
| *** lifeless has quit IRC | 16:28 | |
| *** mikal has quit IRC | 16:29 | |
| sambetts | dhellmann: So originally when we release networking-cisco we release them using the 2015.0.0 style of version numbers, as was the fasion in OpenStack at the time, however now neutron has switch to SemVer we would like to switch also, but we expect that'll cause issues with the version ordering on PyPi | 16:29 |
| *** armax has quit IRC | 16:29 | |
| *** mestery has quit IRC | 16:29 | |
| *** armax_ is now known as armax | 16:29 | |
| *** lifeless has joined #openstack-relmgr-office | 16:30 | |
| *** mikal has joined #openstack-relmgr-office | 16:30 | |
| dhellmann | sambetts: yes, that's right. most of the distros will know to add an "epic" prefix to the number if they package it, but we didn't do that for pip packages | 16:31 |
| dhellmann | did you want to re-tag the liberty release or start with the new numbering for mitaka? | 16:31 |
| bdemers | I think the bigger question is _should_ we | 16:32 |
| dhellmann | that's really up to you | 16:32 |
| bdemers | if we keep the same package name “networking-cisco” i’m guessing that would break version range resolution right ? | 16:32 |
| mordred | dhellmann: I'm not sure pbr supports pip epochs | 16:33 |
| sambetts | dhellmann: we've not tagged our liberty release yet, so we want to start with the new numbering for liberty | 16:33 |
| dhellmann | mordred: we decided not to use them for other reasons, but that's good to know | 16:33 |
| mordred | I'm not positive though - I cannot remember - but I do know that infra tag releasing does not | 16:33 |
| mordred | dhellmann: also, for openstack servers, we don't release them to pypi in the first place, so the usecase that sambetts is bringing up didn't really come up | 16:34 |
| dhellmann | irrc, we decided that because we're not a distributor we wouldn't mess with epochs | 16:34 |
| mordred | (we didn't have any existing date-based entries in pypi) | 16:34 |
| dhellmann | mordred: it did for oslo.config and oslo.messaging, back in the day | 16:34 |
| mordred | oh right | 16:34 |
| mordred | how did we transition the pypi entries for them? is pip just doing the right thing in terms of current versions? | 16:34 |
| dhellmann | I think we just took the hit on the requirements specification, but maybe we didn't publish those versions to pypi | 16:35 |
| mordred | yeah - I don't see them on pypi | 16:35 |
| dhellmann | yeah | 16:35 |
| bdemers | did the package name change for oslo.* when it was updated ? | 16:35 |
| dhellmann | no | 16:35 |
| mordred | so this is a new problem we did not have with the other transitions | 16:36 |
| sambetts | Nice to know we can still come up with new spanners to throw | 16:38 |
| dhellmann | mordred: I think we can solve this with a requirements cap <2000.0 or something | 16:38 |
| bdemers | so it looks like 2013.1b5 was the previous format for oslo | 16:38 |
| mordred | yah - which is a "beta" version which pip does not consider in normal version ranges | 16:38 |
| bdemers | the <2000.0 isn’t a very user frieldnly way to fix it | 16:39 |
| mordred | looks like there is exactly one date-based release of networking-cisco on pypi: https://pypi.python.org/simple/networking-cisco/ | 16:39 |
| bdemers | what about changing all of the networking package nameing ? | 16:39 |
| bdemers | and then depecrate the others | 16:39 |
| dhellmann | not this week | 16:39 |
| bdemers | (that isn’t exactly use friendly either) | 16:39 |
| dhellmann | maybe next cycle | 16:39 |
| mordred | I tink it would be easier to just hide the 2015* releases on pypi | 16:40 |
| bdemers | yeah, but this still effects anyone with a pypi mirror | 16:40 |
| dhellmann | mordred: does pip ignore hidden releases? | 16:40 |
| sambetts | This describes the problem exactly but I don't know how it works with the openstack release system https://www.python.org/dev/peps/pep-0440/#version-epochs | 16:41 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/releases: Release of python-manilaclient 1.5.0 https://review.openstack.org/237057 | 16:41 |
| mordred | dhellmann: I believe so, yes | 16:41 |
| mordred | dhellmann: I'm testing right now | 16:41 |
| dhellmann | ok | 16:41 |
| dhellmann | I thought that was just a ui thing, but if hiding them works that seems like a good solution | 16:42 |
| sambetts | dhellmann: can you still explictly download a hidden version? | 16:42 |
| mordred | dhellmann: yup. just a ui thing. it does not work | 16:42 |
| bdemers | but that only works against pipy directly right? so any mirror would could still have this problem ? | 16:42 |
| bdemers | nm | 16:42 |
| bdemers | so if we go the epoch way, then PBR needs to be updated, and what else ? | 16:44 |
| dhellmann | bdemers, sambetts : think for liberty you should use date-based versioning so we have time to understand the implications of changing | 16:44 |
| dhellmann | bdemers: we don't do epochs of pypi packages for several reasons, pbr is the least of it | 16:44 |
| bdemers | ok | 16:45 |
| bdemers | so we could do. 2015.1.x for kilo and something sillly like 2015.10.x for libery | 16:45 |
| dhellmann | I think you'd do better to stick with the usual 2015.1 and 2015.2 | 16:46 |
| dhellmann | no sense inventing another system | 16:46 |
| bdemers | what if something in stable kilo needs a 2015.2 ? | 16:46 |
| sambetts | if it came to having to remove the old 2015 version from pypi, doing it while only having 1 of those types of release would be easier for damage control | 16:46 |
| mordred | bdemers: that's 2015.1.1 | 16:46 |
| bdemers | mordred: o | 16:47 |
| bdemers | mordred: i’m saying what if we need to minor bump on stable/kilo | 16:47 |
| sambetts | bdemers: the 2015 scheme was <year>.<release 1 or 2>.<fix> | 16:47 |
| bdemers | oh | 16:47 |
| bdemers | ok, nm | 16:47 |
| bdemers | then yeah, 2015.2.x | 16:48 |
| dhellmann | ok, before you tag, let's talk about removing the package | 16:48 |
| dhellmann | does anything currently depend on networking-cisco? | 16:49 |
| bdemers | are other project expected to hit this issue ? | 16:49 |
| dhellmann | which distributions are packaging it, if any? | 16:49 |
| bdemers | dhellmann: RDO and OSP | 16:49 |
| dhellmann | I don't know what OSP is | 16:49 |
| dhellmann | Oracle? | 16:49 |
| bdemers | Redhat’s supported release | 16:50 |
| bdemers | i can change the RPMs if the package name changes (that isn’t an issue) | 16:50 |
| dhellmann | so we have several options: | 16:50 |
| dhellmann | 1. release as date-based version | 16:50 |
| dhellmann | 2. rename and release as not-date-based version | 16:50 |
| dhellmann | 3. remove the existing release and release as not-date-based version | 16:51 |
| dhellmann | 4. release as not-date-based version without removing or renaming | 16:51 |
| dhellmann | others? | 16:51 |
| sambetts | I like 3, we could even re-release stable/kilo as 6.0.0 | 16:51 |
| dhellmann | has this package been around for 6 cycles? | 16:52 |
| mordred | I also like 3- it introduces a potential one-time hit for consumers - but that's not an ongoing hit - and is a smaller hit than name changes (and is only for existing deploys - a single force install) | 16:52 |
| dhellmann | we can figure out the version # when we decide which approach to take | 16:52 |
| sambetts | dhellmann: no, but we want to align with the neutron numbers | 16:52 |
| bdemers | personally, i’m of the opion that a release should be immutable | 16:53 |
| dhellmann | so what would break if we do 3? are any jobs for other projects depending on having a package with that version number installable from pypi? | 16:53 |
| mordred | bdemers: I'm also of that opinion ... | 16:53 |
| mordred | bdemers: I'm 'ok' with removing it from pypi because we would not be removing it from tarballs.o.o - so the 2015.1 tarball would still exist - and it would continue to exist in all the distro repos | 16:53 |
| mordred | so although it's not _ideal_ - it seems like the lesser of evils in this particular case | 16:54 |
| dhellmann | I wonder how many operators might end up broken if we delete it, though | 16:54 |
| bdemers | yeah, the idea of re-tagging kilo is intereseting too | 16:54 |
| mordred | dhellmann: I think that's tied to "how many operators are installing from pip not distro packages and also doing so with pip upgrades rather than throwaway virtualenvs or docker containers" | 16:55 |
| mordred | there is a _very_ specific set of humans who this will break | 16:55 |
| dims | mordred: done (https://review.openstack.org/#/c/236989/) | 16:56 |
| mordred | dims: thank you!!! | 16:56 |
| dhellmann | mordred: yes, it is specific, but how big is it? | 16:57 |
| dhellmann | it looks like we're all over the map for the network project versioning http://paste.openstack.org/show/476730/ | 16:57 |
| openstackgerrit | Merged openstack/releases: Release 1.9.0 of os-client-config https://review.openstack.org/236989 | 16:58 |
| bdemers | so if we _did_ rename, any ideas what that would be ? | 17:00 |
| bdemers | networking-cisco-sys ? | 17:00 |
| bdemers | networking-cisco2 (ugg) | 17:01 |
| *** openstackgerrit has quit IRC | 17:01 | |
| *** openstackgerrit has joined #openstack-relmgr-office | 17:02 | |
| sambetts | ugg ... it would be confusing to have to have another name :/ would we rename the git.openstack.org repo and everything too ? | 17:04 |
| bdemers | ugg | 17:04 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/releases: Release of python-manilaclient 1.5.0 https://review.openstack.org/237057 | 17:05 |
| bdemers | rock -> <- hard place | 17:05 |
| bdemers | so, it is really one single release right ? | 17:05 |
| bdemers | 2015.1.0 | 17:05 |
| bdemers | (though i just requested another release last week) | 17:05 |
| sambetts | Yup thats it, that all we've released properly so far, (except for like 2 0.0.X test releases) | 17:06 |
| dhellmann | how many consumers of that release, from pypi, do you think you have? | 17:10 |
| sambetts | dhellmann: according to PyPi 553 downloads in the last month, 116 downloads in the last week, and 2 downloads in the last day although I have no idea if thats useful :/ | 17:12 |
| dhellmann | that could all be our ci system | 17:12 |
| dhellmann | although I don't know if you have any jobs installing that package from pypi instead of source? | 17:12 |
| sambetts | bdemers: any ideas? ^ | 17:13 |
| bdemers | nope | 17:14 |
| dhellmann | who is likely to come with pitchforks and torches if you delete the thing that is keeping their environment running? | 17:15 |
| bdemers | i still think the bigger issue here is anybody how has mirrored/proxied internally | 17:16 |
| bdemers | which is a single external download from pypi, but many internally | 17:16 |
| bdemers | (i’m not sure if anybody is actually doing that…) | 17:17 |
| bdemers | ok, just for completeness, why not use an epoch ? | 17:19 |
| dhellmann | bdemers: it doesn't work with all of the packaging ecosystem, iiuc, and because we're not a distributor -- now that last one may apply less here since you're publishing to pypi | 17:23 |
| cp16net | dims: could you look at the 2 requirement patches we have to block oslo.utils 2.6.0? master: https://review.openstack.org/#/c/235591/ stable/liberty: https://review.openstack.org/#/c/235591/ | 17:24 |
| dhellmann | cp16net: you need to update the upper-constraints.txt file in that second patch | 17:25 |
| dhellmann | actually those links point to the same patch | 17:25 |
| cp16net | oh crap... | 17:25 |
| cp16net | stable/liberty: https://review.openstack.org/#/c/235588/ | 17:26 |
| cp16net | there it is. | 17:26 |
| bdemers | dhellmann: what does the openstack foundation consider a release ? | 17:26 |
| bdemers | (i.e. for apache it is more or less a source tarball) | 17:26 |
| dhellmann | bdemers: is that a philosophical question? what are you getting at? | 17:27 |
| bdemers | no, a pratical release question | 17:27 |
| bdemers | i.e. if you only care about a soruce tarball, the other packaging systems are not a concern | 17:28 |
| dhellmann | we make release artifacts in the form of tarballs,t oo | 17:28 |
| dhellmann | for things we publish to pypi, we care about those artifacts, too | 17:28 |
| bdemers | the integration falls onto who packages the tarball | 17:28 |
| bdemers | so if we _supported_ the epoch, it woudn’t cause any issues within the openstack foundation ? | 17:29 |
| bdemers | sorry to keep brining this up, if it is a dead horse, but it sounds like our best option from my point of view | 17:30 |
| cp16net | dhellmann: quick question those 2 patches were made separately. should we merge to master first then cherry-pick it to stable/liberty or the 2 patches ok separate? | 17:31 |
| openstackgerrit | Merged openstack/releases: Oslo Releases for Week of Oct 19th 2015 https://review.openstack.org/236770 | 17:31 |
| dhellmann | cp16net: as long as we get them both in, the order isn't important | 17:33 |
| cp16net | ok cool | 17:33 |
| dhellmann | cp16net: I approved the one for stable/liberty, and if you fix up the one for master I will approve that one, too | 17:33 |
| cp16net | just updated | 17:35 |
| dhellmann | cp16net: approved | 17:38 |
| cp16net | dhellmann: thanks! :) | 17:38 |
| * dhellmann disappears for another pre-summit errand | 17:39 | |
| lifeless | dhellmann: when you get back, can we talk reno and the blacklist? | 17:53 |
| *** armax has quit IRC | 18:10 | |
| stevemar_ | dhellmann: dims could i get eyes on: https://review.openstack.org/#/c/236797/ ? | 18:11 |
| stevemar_ | dhellmann: also, what can i do to not bother you with requests to review release patches all the time :) | 18:12 |
| *** dtantsur is now known as dtantsur|afk | 18:13 | |
| dims | stevemar_: what's what this channel is for...so ask away! | 18:35 |
| stevemar_ | dims: \o/ | 18:35 |
| stevemar_ | dims: i just didn't want to seem like i pester you guys | 18:35 |
| dims | stevemar_: i am still trying to figure out the best way to test a library before we release it. especially py27/py34 jobs of different projects. | 18:41 |
| *** armax has joined #openstack-relmgr-office | 18:41 | |
| dims | stevemar_: on its way | 18:45 |
| *** amotoki has quit IRC | 18:55 | |
| dhellmann | lifeless: back | 19:10 |
| dhellmann | stevemar_: that release looks good. dims were you taking that release or should I? | 19:12 |
| dims | dhellmann: took care of it | 19:14 |
| stevemar_ | dims: we more or less cover things in osc and ksa by using tempest dsvm jobs | 19:14 |
| lifeless | dhellmann: I don't understand why reno is in the requirements constraints blacklist | 19:14 |
| dhellmann | dims: that patch isn't merged, how did you do it? | 19:14 |
| dims | dhellmann: was waiting to check on pypi | 19:14 |
| lifeless | dhellmann: it doesn't seem to meet the (admittedly tribal knowledge - so I'll put up a patch to improve it) criteria | 19:14 |
| dims | it's there now | 19:14 |
| dhellmann | lifeless: because constraints are only applied to places that reno is not being used | 19:15 |
| dhellmann | lifeless: reno is only going to be in either the doc build or a special release notes build | 19:15 |
| lifeless | dhellmann: thats not a reason not to constrain it - its just we haven't finished the rollout of constraints | 19:15 |
| dhellmann | dims: you know there's a script to read a yaml file and run the release, right? I usually merge first so I can run that script. | 19:15 |
| dims | announce sent | 19:15 |
| lifeless | dhellmann: the docs build will be constrained very soon for instance - we have them enabled non-voting on neutron now | 19:16 |
| dims | dhellmann: ah. which one? | 19:16 |
| dhellmann | dims: release_from_yaml.py | 19:16 |
| dims | release_from_yaml.py gotcha | 19:16 |
| dhellmann | lifeless: ok, well, as a release-related tool I thought it would be better to not wedge ourselves on requirements specifications | 19:16 |
| dhellmann | dims: ok, I'm going to flush the rest of the release queue now | 19:17 |
| lifeless | dhellmann: I don't really follow that; anything in the gate can wedge us, and anything unconstrained can break multiple projects at once yet require multiple patches to fix | 19:17 |
| dims | dhellmann: ++ | 19:18 |
| lifeless | dhellmann: constraints makes that symmetrical - one patch to break, one patch to fix | 19:18 |
| openstackgerrit | Merged openstack/releases: release keystoneauth 1.2.0 https://review.openstack.org/236797 | 19:18 |
| dhellmann | lifeless: we keep building rules to enforce behaviors that we could just work around. This tool isn't going to break the gate. At worst it will break a post-merge job. | 19:18 |
| lifeless | dhellmann: if its used in docs jobs | 19:19 |
| lifeless | dhellmann: then its not post-merge | 19:19 |
| dhellmann | lifeless: it's not yet clear we want it in that job. | 19:19 |
| dhellmann | we would end up publishing release notes under docs.openstack.org/developer/ which seems not obvious | 19:19 |
| dhellmann | this is something I plan to discuss at the summit | 19:19 |
| lifeless | dhellmann: sec, elocal from daughter | 19:19 |
| openstackgerrit | Merged openstack/releases: Release of python-manilaclient 1.5.0 https://review.openstack.org/237057 | 19:22 |
| lifeless | dhellmann: food refusal, I'll be a while | 19:24 |
| dhellmann | lifeless: np | 19:25 |
| dhellmann | stevemar_: you have https://review.openstack.org/#/c/236792/ WIPed, did you want that release? | 19:26 |
| stevemar_ | dhellmann: not just yet, maybe in a few minutes | 19:27 |
| dhellmann | stevemar_: ack, just remove the WIP when you're ready | 19:27 |
| stevemar_ | just waiting on 2 more patches to merge (tests and release notes) | 19:27 |
| stevemar_ | yes sir | 19:27 |
| dhellmann | stevemar_: this release for keystonemiddleware includes a change to the way ids are processed for pycadf -- does that make it incompatible with something looking for those old ids? https://review.openstack.org/#/c/235999/2 | 19:28 |
| stevemar_ | dhellmann: yes, i suppose so... it used to add a prefix "openstack:", but now we are removing that since it just made things confusing | 19:30 |
| dhellmann | stevemar_: so the question then is whether you consider that data stream part of the API, and if ids for old things will change with this new version, and make this a major version bump | 19:31 |
| stevemar_ | gordc: thoughts? ^ i'm tempted to say that ids are an implementation detail and may change | 19:33 |
| stevemar_ | we're not removing the field and breaking a contract.. (just talking aloud) | 19:33 |
| dhellmann | stevemar_, gordc : I -1 the patch so we can discuss it there so we have a record of it in a place that's easy to find | 19:34 |
| gordc | stevemar_: how far up do i need to read?lol | 19:34 |
| dhellmann | I mean, we can discuss it here, too, but at least link to that in the review | 19:34 |
| stevemar_ | gordc: just a few lines :) | 19:34 |
| gordc | oh, so context | 19:34 |
| dhellmann | gordc: the cadf output of the new keystonemilddleware release appears to be backwards-incompatible, but I'm not sure what guarantees are made about that data | 19:34 |
| gordc | dhellmann: in ceilometer, we scope the results based on the value we capture in project_id and compare it to project_id the user is scoped too | 19:35 |
| openstackgerrit | Merged openstack/releases: mitaka: release python-novaclient 2.32.0 https://review.openstack.org/232683 | 19:35 |
| gordc | the problem is, the values from keystonemiddleware were all prefixed with openstack: | 19:36 |
| dhellmann | ok, but the question is really about whether other users of that data might break as a result of this change | 19:36 |
| stevemar_ | gordc: we just have to decide if it's a minor or major version bump | 19:36 |
| gordc | so we could never actually match to them as from ceilometer (and rest of openstack ecosystem) that namespace prefix is unknown | 19:36 |
| dhellmann | it sounds like ceilometer needs some way to deal with both versions of the data anyway, since you can't control what's happening in the middleware used by another service | 19:37 |
| gordc | oh. i'd probably make it major since i don't know how else it's being consumed | 19:37 |
| gordc | from ceilometer pov, it's minor because with the prefix, it's unusable. | 19:37 |
| stevemar_ | poor use of the term minor gordc :P | 19:38 |
| gordc | well minor as in i don't think we really need to worry about backward compat (since it's broken) | 19:39 |
| gordc | but yeah, until you lose a limb, it's all minor flesh wounds.lol | 19:39 |
| lifeless | "None shall pass!" | 19:40 |
| gordc | let me know if we choose major or minor, i need to apply same namespace drop to pycadf lib too. | 19:42 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: import liberty versions for unmanaged projects https://review.openstack.org/236156 | 19:42 |
| stevemar_ | gordc: i think dhellmann wants us to decide | 19:43 |
| stevemar_ | :) | 19:43 |
| gordc | stevemar_: major | 19:43 |
| stevemar_ | gordc: i'm okay with a major | 19:43 |
| stevemar_ | now to doc this all in the patch :P | 19:43 |
| gordc | stevemar_: cool cool | 19:43 |
| gordc | stevemar_: you see dvp from your condo? | 19:44 |
| stevemar_ | gordc: nope, i face north | 19:44 |
| gordc | dang. i wonder how busy the roads are. /me googles webcams. | 19:45 |
| bknudson | gordc: stevemar_: don't forget to vote | 19:45 |
| gordc | bknudson: i assume you watch "last week tonight"? | 19:46 |
| bknudson | gordc: I only watch 22 minutes or whatever it's called. | 19:46 |
| gordc | hahah kudos on the reference. | 19:46 |
| stevemar_ | gordc: what up? going to the game tonight? | 19:46 |
| gordc | stevemar_: trying to avoid the traffic from the game. | 19:47 |
| bknudson | gordc: stevemar_ I won't tell you whom to vote for. | 19:47 |
| * gordc already voted. | 19:48 | |
| gordc | they give us pencils to vote with... /me hopes no one discovers what an eraser is | 19:50 |
| stevemar_ | pencil and paper, so secure! | 19:52 |
| stevemar_ | bknudson: here ya go: https://www.youtube.com/watch?v=0V5ckcTSYu8 | 19:53 |
| openstackgerrit | Merged openstack/releases: Juno: release python-glanceclient 0.14.3 https://review.openstack.org/235967 | 19:57 |
| dims | stevemar_: no hanging chads? | 20:01 |
| stevemar_ | dims: definitely no hanging chads! | 20:02 |
| *** dims_ has joined #openstack-relmgr-office | 20:16 | |
| openstackgerrit | Brant Knudson proposed openstack/releases: keystonemiddleware 3.0.0 https://review.openstack.org/235999 | 20:19 |
| *** dims has quit IRC | 20:19 | |
| openstackgerrit | Steve Martinelli proposed openstack/releases: Release openstackclient 1.8.0 https://review.openstack.org/236792 | 20:23 |
| *** bswartz has quit IRC | 20:23 | |
| stevemar_ | bahhh of course the tempest neutron tests timed out :) | 20:27 |
| *** gordc has quit IRC | 20:38 | |
| lifeless | dhellmann: ok so | 20:47 |
| lifeless | dhellmann: I think you're saying you want to discuss reno-and-blacklist at the summit? | 20:47 |
| dhellmann | lifeless: I want to discuss where we publish reno's output at the summit, and that will have a bearing on the blacklist discussion | 20:48 |
| lifeless | dhellmann: My main fear is that we'll get things in the blacklist that can wedge the gate (vs impacting only single projects) | 20:48 |
| dhellmann | I view reno as a build tool, and I didn't think we wanted those constrained | 20:48 |
| lifeless | dhellmann: so I want us to be super conservative about whats in the blacklsit | 20:48 |
| lifeless | dhellmann: pbr is not blacklisted | 20:49 |
| lifeless | dhellmann: the only things I understood us to be blacklisting was things which *must* vary per-project such as linters. | 20:49 |
| dhellmann | ok, that makes some sense. I'd been treating that more broadly because I thought we wanted to minimize the number of places we had to change a version number if we ended up broken because of some build tool | 20:50 |
| lifeless | I'll put up a patch adding more reasoning and docs to master requirements | 20:51 |
| dhellmann | that would be helpful | 20:51 |
| lifeless | we can talk about reno more later, its not urgent | 20:51 |
| lifeless | [what got me onto this was the openstackdocstheme patch and talking to AJ aboutit | 20:51 |
| dhellmann | yeah, I thought that fit the same criteria: only the doc projects use it, it isn't a runtime dependency, it's part of the build, etc. | 20:53 |
| lifeless | we're mismatched on the criteria :) | 20:55 |
| * dhellmann nods | 20:55 | |
| *** stevemar_ has quit IRC | 21:22 | |
| *** stevemar_ has joined #openstack-relmgr-office | 21:22 | |
| *** stevemar_ has quit IRC | 21:22 | |
| *** stevemar_ has joined #openstack-relmgr-office | 21:23 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:01 | |
| *** dims_ has quit IRC | 22:15 | |
| *** bswartz has joined #openstack-relmgr-office | 22:33 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!