| *** sdague has quit IRC | 00:15 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/releases: Final Oslo releases for Mitaka (except oslo.messaging) https://review.openstack.org/284430 | 00:16 |
|---|---|---|
| dims | dhellmann : just in case you are around ^^ | 00:16 |
| dhellmann | dims : checking | 00:18 |
| dhellmann | oof, lots of them :-) | 00:18 |
| dhellmann | dims : I'll wait for the list-changes job to finish and look over the logs first thing tomorrow | 00:18 |
| dims | the delta changes should be small | 00:18 |
| dhellmann | yeah | 00:18 |
| dims | dhellmann : the last change was a commit message change only so you can check list-changes output from the run before | 00:19 |
| dhellmann | dims : ack | 00:19 |
| dims | tomorrow morning is ok too | 00:19 |
| *** cody-somerville has quit IRC | 00:21 | |
| dhellmann | dims : +1 | 00:25 |
| dims | dhellmann : ok let me cut it in a couple of hours and deal with fallout tonight itself. thanks | 00:25 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: run yamllint on deliverable files during validation https://review.openstack.org/284120 | 00:54 |
| *** dims has quit IRC | 01:52 | |
| *** amotoki has joined #openstack-release | 02:29 | |
| *** amotoki has quit IRC | 03:19 | |
| *** amotoki has joined #openstack-release | 03:30 | |
| *** amotoki has quit IRC | 03:35 | |
| *** amotoki has joined #openstack-release | 03:40 | |
| *** amotoki has quit IRC | 04:01 | |
| *** amotoki has joined #openstack-release | 04:13 | |
| *** dims has joined #openstack-release | 05:52 | |
| *** cody-somerville has joined #openstack-release | 05:54 | |
| *** doug-fish has joined #openstack-release | 06:00 | |
| openstackgerrit | Merged openstack/releases: Final Oslo releases for Mitaka (except oslo.messaging) https://review.openstack.org/284430 | 06:03 |
| *** cody-somerville has quit IRC | 06:34 | |
| *** dims has quit IRC | 06:43 | |
| *** doug-fish has quit IRC | 06:54 | |
| *** amotoki has quit IRC | 08:53 | |
| *** amotoki has joined #openstack-release | 09:10 | |
| *** pcaruana has joined #openstack-release | 10:42 | |
| *** sdague has joined #openstack-release | 11:06 | |
| *** dtantsur|afk is now known as dtantsur | 11:09 | |
| dhellmann | ttx: if you're around, we can chat about release testing | 12:18 |
| ttx | dhellmann: around now | 12:47 |
| dhellmann | ttx: I spent some time yesterday thinking about release testing. We should probably test once with each type of release (normal, alpha, beta, rc) even though we've done the other types previously in other repos. Did you have anything more specific in mind? | 12:50 |
| ttx | nope. I just wanted to make sure we did a 0b3 -> 0rc1 -> 0rc2 -> final | 12:51 |
| ttx | and we cut the stable branch at 0rc1 | 12:51 |
| ttx | to test all the tooling | 12:51 |
| ttx | "normal" releases we can re-test but we exercise that part often enough | 12:52 |
| ttx | I think we need ~1hour together to go through the steps | 12:52 |
| ttx | I'm available today before 16 utc | 12:53 |
| dhellmann | ttx: ack. I created a stable/mitaka branch in release-test earlier this week to test a patch to the script for automating that step. I just used "HEAD" as the version though | 12:53 |
| dhellmann | ttx: ok, I'm going to make some breakfast, and then I'll line up release requests for all of those cases | 12:53 |
| ttx | we could remove the branch and/or create stable/newton instead if we want a clean slate | 12:54 |
| ttx | or stable/meiji | 12:54 |
| *** gordc has joined #openstack-release | 12:54 | |
| *** pcaruana has quit IRC | 12:55 | |
| dhellmann | if you want to look at https://review.openstack.org/284427 and the patches it created, we could throw all of them away and delete the branch | 12:55 |
| dhellmann | brb | 12:55 |
| ttx | ok, will have a look | 12:55 |
| *** pcaruana has joined #openstack-release | 13:09 | |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test an alpha release https://review.openstack.org/284711 | 13:16 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test beta three release https://review.openstack.org/284712 | 13:16 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test first release candidate https://review.openstack.org/284713 | 13:16 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test second release candidate https://review.openstack.org/284714 | 13:16 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test final release https://review.openstack.org/284715 | 13:16 |
| dhellmann | ttx: I think we're ready to go with these patches ^^ -- we have to land them one at a time for the validation to work on the follow-ups | 13:37 |
| ttx | looking | 13:38 |
| dhellmann | I used the same sha for most of them (we don't have a lot of commits in that repo) | 13:39 |
| ttx | Wondering if making them in the _independent directory is going to affect anything (vs. mitaka) | 13:39 |
| dhellmann | oh, hmm, I hadn't thought of that | 13:39 |
| dhellmann | I can move them over, just to be safe. we can delete the file when we're done to clean up the release reporting site | 13:40 |
| ttx | I don't think we have any usage for that, beyond announcement and publication data | 13:40 |
| ttx | right, probably better | 13:40 |
| ttx | looking at the SHAs vs. branches | 13:41 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test an alpha release https://review.openstack.org/284711 | 13:41 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test final release https://review.openstack.org/284715 | 13:41 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test second release candidate https://review.openstack.org/284714 | 13:41 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test first release candidate https://review.openstack.org/284713 | 13:41 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test beta three release https://review.openstack.org/284712 | 13:41 |
| dhellmann | yeah, so like I said I used the same sha for most of them because we don't have a lot of commits but we can add some quickly to make the scenario more realistic if you think it matters | 13:42 |
| ttx | I'd like to test that if a commit is only on stable/mitaka it still finds it for rc2 | 13:42 |
| dhellmann | ok, sure | 13:42 |
| dhellmann | we've done stable releases already, but it's good to verify | 13:42 |
| dhellmann | so when we get to that point, and have the branch, we can create a new commit and tag that instead of what's in the file now | 13:43 |
| ttx | right, one step at a time | 13:43 |
| ttx | we can do the a1 -> rc1 | 13:43 |
| ttx | though arguably we could just skip a1 since we won't do that | 13:44 |
| ttx | but that shouldn't hurt | 13:44 |
| dhellmann | do you want to run them, to make sure there's no bias about my copy of the tools or something? | 13:44 |
| ttx | yep | 13:44 |
| ttx | I'm on it | 13:44 |
| ttx | ok, so a1 first | 13:45 |
| dhellmann | ok, I'd like to use my copy of the branch creation script to make the branch when we get to that point | 13:45 |
| dhellmann | ttx: perhaps we should delete the stable/mitaka branch first? | 13:47 |
| *** dims has joined #openstack-release | 13:47 | |
| ttx | perhaps we should, but we can't | 13:48 |
| ttx | we'll need help | 13:48 |
| dhellmann | ttx: oh, right. well, let's just use a different name | 13:49 |
| ttx | change is stuck in test queue | 13:49 |
| dhellmann | I don't think anything checks the name of the branch, just the DAG | 13:49 |
| dhellmann | aw | 13:50 |
| ttx | yeah, the only potential glitch is between the releases file directory name and branch name | 13:50 |
| ttx | but I don't think there is anything linking the two | 13:51 |
| dhellmann | that might be worth adding, but we don't do it now | 13:51 |
| ttx | ideally we would submit "newton/release-test.yaml" and create stable/newton | 13:51 |
| dhellmann | I could make up a fake name and resubmit the patches using that, just to be safe | 13:51 |
| ttx | sure, starting with 0b3 | 13:52 |
| ttx | no need to redo 0a1 now that it's in the pipe ? | 13:52 |
| dhellmann | oh, sure, we could have folks request branches through this | 13:52 |
| dhellmann | yeah, that won't matter, I'll update the others | 13:52 |
| ttx | that will actually speed up things, since the 0b3 will be able to pass tests on its own | 13:52 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test final release https://review.openstack.org/284715 | 13:53 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test second release candidate https://review.openstack.org/284714 | 13:53 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test first release candidate https://review.openstack.org/284713 | 13:53 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test beta three release https://review.openstack.org/284712 | 13:53 |
| dhellmann | ttx: it won't, because the previous tag won't exist yet | 13:54 |
| ttx | oh, you moved it over, ok | 13:54 |
| * dims grabs a cup of coffee and reports for work :) | 13:54 | |
| dhellmann | good morning, dims. we're exercising the release tools this morning | 13:55 |
| dims | cool | 13:55 |
| ttx | (especially the rc stuff which is largely untested) | 13:55 |
| dhellmann | ttx: isn't that what you meant? to move the file? | 13:55 |
| ttx | dhellmann: it works too, but I thought you would just create a meiji file with the 0b3 in it. | 13:56 |
| dhellmann | ttx: oh, I see, sure. that would have been faster, I guess. | 13:56 |
| ttx | it's fine, we are stuck with the check queue anyway | 13:56 |
| dhellmann | yeah | 13:56 |
| dims | zuul is a sea of red - looks like sdague has a review for fixing | 13:58 |
| dims | ansible | 13:58 |
| dhellmann | oh, yeah, I saw that email. I have zuul filtered to "release" to watch these tests, and they're all working ok | 13:59 |
| dhellmann | I wonder if we got in after his fix, or if we're just not affected | 13:59 |
| ttx | ok, looks like we are a go | 14:01 |
| openstackgerrit | Merged openstack/releases: test an alpha release https://review.openstack.org/284711 | 14:01 |
| * dhellmann dons his Thunderbirds outfit | 14:01 | |
| ttx | tag applied | 14:03 |
| dhellmann | pre-release jobs are queued | 14:03 |
| dhellmann | ah, right, no announce email for pre-releases | 14:03 |
| dhellmann | I wonder if we want to change that for rc? | 14:03 |
| ttx | I would say no | 14:04 |
| dhellmann | don't you usually send an announcement of some sort? | 14:04 |
| ttx | what would you list on the "fixes" there ? All mitaka fixes ? post-FF fixes ? | 14:04 |
| dhellmann | my first reaction is to say post-ff but I see your point | 14:04 |
| ttx | it's slightly tricky | 14:05 |
| dhellmann | yeah | 14:05 |
| ttx | I'm fine sending my canned email on RCs | 14:05 |
| dhellmann | if we want to do it, we could add a flag to leave out the list of changes | 14:05 |
| dhellmann | but -- right -- a personal email might be better for that case | 14:05 |
| ttx | so, for RC2s it would not be a bad idea to have the list of post-RC1 fixes | 14:06 |
| ttx | since in theory people are testing RCs and want to know what changed since the last one | 14:06 |
| dhellmann | we could use the announce script manually to compile those | 14:06 |
| ttx | ok, so 0a1 seems to have worked as designed | 14:07 |
| ttx | oh, our meiji stuff will breal releases.o.o | 14:07 |
| ttx | break | 14:07 |
| ttx | shall we make it newton so that the publication to releases.o.o works ? | 14:08 |
| dhellmann | the doc build is working for me locally, what do you think will be broken? | 14:08 |
| ttx | No error due to missing meiji index pages ? | 14:09 |
| dhellmann | no | 14:09 |
| dhellmann | it's driven by the existing files, so the new dir is ignored until a file is added | 14:10 |
| dhellmann | I could do that in the update to the b3 tag if you want | 14:10 |
| dhellmann | we have to recheck it anyway | 14:10 |
| ttx | yeah, that way we can test the releases.o.o site output as well | 14:10 |
| ttx | just a sec though | 14:10 |
| ttx | 0a1 tarball link is broken on http://releases.openstack.org/mitaka/index.html | 14:11 |
| ttx | https://tarballs.openstack.org/release-test/release-test-0.0.1.0a1.tar.gz 404 | 14:11 |
| ttx | it's stored at https://tarballs.openstack.org/release-test/openstack-release-test-0.0.1.0a1.tar.gz | 14:11 |
| ttx | looks like a setup.cfg glitch | 14:12 |
| ttx | which we should probably just ignore for this test | 14:13 |
| dhellmann | no, I put openstack-release-test in the setup.cfg to give it a unique name upstream | 14:14 |
| dhellmann | let me see where I get that tarball name | 14:14 |
| ttx | tarball name is probably derived from repository name, but the built tarball depends on setup.cfg | 14:15 |
| ttx | Do we have real world case where the two differ ? Like glance_-store or oslo.-_test ? | 14:16 |
| dhellmann | python-cliff and cliff | 14:16 |
| dhellmann | maybe? maybe that's just on launchpad | 14:16 |
| dhellmann | yeah, that's not one | 14:17 |
| ttx | cliff and glance-store seem to survive it well | 14:17 |
| dhellmann | yeah, ok, maybe it's safe to ignore that and keep an eye out for a real case | 14:17 |
| ttx | ack | 14:18 |
| ttx | let's move on to 0b3... are you moving everything to newton ? | 14:18 |
| dhellmann | I'm testing my doc changes locally | 14:18 |
| ttx | or creating index pages for meiji | 14:18 |
| dhellmann | no, I added a meiji page | 14:18 |
| ttx | ok | 14:18 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test final release https://review.openstack.org/284715 | 14:19 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test second release candidate https://review.openstack.org/284714 | 14:19 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test first release candidate https://review.openstack.org/284713 | 14:19 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test beta three release https://review.openstack.org/284712 | 14:19 |
| dims | ttx : dhellmann : how do deal with requirements requests? especially new packages at this stage of the release? | 14:20 |
| ttx | dims: still ok until next week | 14:21 |
| dhellmann | dims: I'm reluctant to add new requirements now. What is it? | 14:21 |
| dims | https://review.openstack.org/#/c/284183/ | 14:21 |
| ttx | but yeah, only things that are actually useful :) | 14:21 |
| dims | and https://review.openstack.org/#/c/277893/ | 14:21 |
| dims | haha | 14:21 |
| ttx | test requirement bah | 14:22 |
| dhellmann | ugh, the scapy thing is going to come back to bite us as soon as someone adds it as a real dependency to a project | 14:22 |
| dhellmann | ttx: we have no way to prevent a project from adding scapy as a runtime dependency and it's GPL | 14:22 |
| dhellmann | dims : +2 on pytun | 14:23 |
| ttx | you mean because it's listed in g-r as a test requirement ? | 14:23 |
| dims | right dhellmann | 14:23 |
| dhellmann | ttx: we don't differentiate between regular and test requirements any more in the global list | 14:23 |
| dims | yep | 14:23 |
| ttx | we differebtiate, but we don't enfore it | 14:24 |
| ttx | oops | 14:24 |
| dhellmann | where do we differentiate? | 14:24 |
| ttx | comment in the file ? "# Testing tools below, which are typically in test-requires.txt" | 14:24 |
| dhellmann | ok, I suppose technically we do differentiate without enforcing it | 14:25 |
| ttx | so in theory everything below line 279 should not be put in requirements.txt, only in test-req | 14:25 |
| dhellmann | the question is, how big of a deal is that for unofficial projects? or even official projects that aren't in the tc-approved-release? | 14:25 |
| ttx | but we don't prevent crossusage yes | 14:26 |
| ttx | unofficial projects can do what they want | 14:26 |
| dhellmann | dragonflow is now official | 14:26 |
| ttx | official projects, the clarification we recently pushed says "not ok" | 14:26 |
| dhellmann | should we do some sort of review as part of the release process? | 14:27 |
| ttx | I think it's more due diligence that the Foundation needs to conduct | 14:27 |
| ttx | i.e. periodically check | 14:28 |
| dhellmann | ok | 14:28 |
| ttx | My recent clarification of licensing work was the first step | 14:28 |
| dhellmann | it ought to be easy enough to add to the check-requirements job since we have license comments for all of them now | 14:28 |
| dhellmann | if the project is official, enforce no gpl in requirements.txt | 14:28 |
| dims | good point dhellmann that's a good spot | 14:29 |
| ttx | it's doable indeed | 14:29 |
| dhellmann | anyway, on this one it's failing some tests | 14:29 |
| ttx | My second step was to add a few outlier licenses to the list of acceptable ones, so that the current global-requirements.txt actualkly conforms to the rule | 14:29 |
| ttx | Then 3rd step was to do due diligence on current project requirements | 14:30 |
| ttx | Then 4th step was to actually dig into the code using analyzers to check for random copyrigth violations | 14:30 |
| ttx | copy-pasted code and such | 14:30 |
| * dhellmann nods | 14:30 | |
| ttx | so yes at step 3 we could contemplate enforcing a bit more | 14:31 |
| dhellmann | all of that sounds like a good plan | 14:32 |
| * dims is all in | 14:32 | |
| ttx | but compared to last year where we just had no idea if we were complying with our rule (and what exactly the rule was) it's already an improvement | 14:32 |
| ttx | now we have a pretty good idea, we just need to tie a few lose ends, then check across the board | 14:33 |
| ttx | still waiting on the check queue | 14:34 |
| *** amotoki has quit IRC | 14:37 | |
| *** mriedem has joined #openstack-release | 14:40 | |
| * ttx takes a quick break | 14:52 | |
| dhellmann | ttx: the b3 patch is in the gate now | 15:07 |
| openstackgerrit | Merged openstack/releases: test beta three release https://review.openstack.org/284712 | 15:08 |
| ttx | alright | 15:08 |
| ttx | tagged | 15:09 |
| dhellmann | ok, is this where we wanted to create the branch? | 15:11 |
| ttx | dhellmann: so for the RC1 I think we would process the tagging request first and then create the branch on the same SHA | 15:11 |
| ttx | or the other way around? | 15:11 |
| dhellmann | in the past we've created the branch using the tag as the argument, not the sha | 15:11 |
| ttx | ok that works | 15:12 |
| ttx | so we should first process the tag | 15:12 |
| dhellmann | at least that's what we did for libs, because that enforces that there is a tag to start the branch | 15:12 |
| dhellmann | ok, so we probably need to recheck the rc1 patch then | 15:12 |
| ttx | ok willdo | 15:12 |
| ttx | release publication looks good on 0b3 | 15:13 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 15:15 | |
| * dhellmann resumes waiting for the check queue | 15:16 | |
| dhellmann | ttx, dims: do you want to review https://review.openstack.org/284427 while we wait? | 15:16 |
| ttx | on it | 15:17 |
| ttx | TIL today... .title() | 15:18 |
| dims | magic? :) "$(python -c "print('$shortbranch'.title())")" | 15:20 |
| dhellmann | dims : I found a way to do it with sed, too, but this seemed easier to understand | 15:20 |
| ttx | how about ${shortbranch[@]^} ? | 15:22 |
| dims | dhellmann : ack. i ran it by hand to understand what it does :) had not seen title() in my python adventures :) | 15:22 |
| ttx | so much more redable :) | 15:22 |
| ttx | readable* | 15:22 |
| dhellmann | ttx: does that do the same thing? | 15:22 |
| ttx | capitalizes first letter | 15:22 |
| dhellmann | sure, we can change it to that if it makes more sense | 15:23 |
| dims | ttx : that does not change release-tools to Release-Tools | 15:23 |
| ttx | ah hm | 15:24 |
| ttx | oh well | 15:24 |
| dims | :) +2 as is | 15:24 |
| dhellmann | dims : release-tools isn't the name of the branch though | 15:24 |
| dhellmann | our branches, for now, are all 1 word | 15:24 |
| dims | ah. so we are good either way | 15:25 |
| ttx | +2ed | 15:25 |
| openstackgerrit | Merged openstack-infra/release-tools: use unique tmp space dir for make_stable_branch.sh https://review.openstack.org/284426 | 15:30 |
| openstackgerrit | Merged openstack-infra/release-tools: update make_stable_branch.sh to add reno pages https://review.openstack.org/284427 | 15:30 |
| *** sdague has quit IRC | 15:32 | |
| openstackgerrit | Merged openstack/release-test: Update reno for stable/mitaka https://review.openstack.org/284425 | 15:40 |
| ttx | dhellmann: note that I replied to a few questions on #openstack-dev this morning which you should probably repeat the answers for in your weekly post | 15:52 |
| ttx | (if you haven't sent it already) | 15:52 |
| ttx | about FFEs | 15:53 |
| ttx | and M3 date | 15:53 |
| dhellmann | ttx: I'll look for those | 16:06 |
| dhellmann | ttx: I have a patch ready to fix the tarball link issue, but I'll wait to submit it until the rc1 patch is in to avoid triggering an accidental rebase | 16:06 |
| *** dims has quit IRC | 16:06 | |
| *** sdague has joined #openstack-release | 16:14 | |
| dhellmann | ttx: I reviewed the channel logs. I thought we were going to create the stable branches at rc1 but not change the ACLs (so normal stable policy would apply) | 16:16 |
| dhellmann | ttx: I've made some notes in the reminders section of R-6 in our etherpad, if have a couple of minutes to take a look | 16:20 |
| dhellmann | ttx: I think it's going to be a bit before that rc1 patch is ready to merge. I'm going to step away and take care of a few things. | 16:24 |
| ttx | yeah, looks like we'll have to continue this tomorrow | 16:26 |
| dhellmann | ttx: yeah, I'll make a point of getting up early tomorrow to see if we can try to finish before the queue starts backing up again | 16:33 |
| ttx | added a few comments on the etherpad | 16:33 |
| dhellmann | I saw, I'm moving the ones you suggested pushing to next week | 16:35 |
| sdague | fyi, it looks like setuptools is ruining requirements unit tests | 16:45 |
| sdague | so requirements jobs can't pass | 16:45 |
| dhellmann | sdague : ? | 16:46 |
| sdague | https://jenkins04.openstack.org/job/gate-requirements-python27/992/console | 16:48 |
| dhellmann | is gerrit down? | 16:49 |
| sdague | no | 16:49 |
| sdague | it's just slow | 16:50 |
| sdague | or, maybe... | 16:50 |
| dhellmann | was there a new setuptools release or something? I'm not able to reproduce that error locally. | 16:51 |
| *** dtantsur is now known as dtantsur|afk | 16:51 | |
| sdague | yes | 16:51 |
| sdague | http://lists.openstack.org/pipermail/openstack-dev/2016-February/087486.html some of the debug info is there. But there was a setuptools release which tried to make requirements parsing stricter | 16:53 |
| sdague | that mostly made it wrong | 16:53 |
| dhellmann | sdague : I see 20.2.1 in that job but not on pypi | 16:53 |
| sdague | perhaps it was removed? | 16:53 |
| sdague | it will still be in the infra mirror | 16:53 |
| dhellmann | perhaps -- trying to force install that is failing | 16:53 |
| dhellmann | yeah | 16:53 |
| dhellmann | do we need someone from infra to remove that, or will it be flushed when the cache is updated? | 16:54 |
| *** dims has joined #openstack-release | 16:54 | |
| *** amotoki has joined #openstack-release | 16:54 | |
| sdague | I don't know | 16:54 |
| sdague | I got as far as "there is a fail for a class of jobs that looks like it's 100%, lets go tell the interested subgroup" | 16:55 |
| dhellmann | k | 16:55 |
| dhellmann | sdague : just found a tweet announcing the package being removed | 17:04 |
| * dhellmann sighs | 17:04 | |
| openstackgerrit | Doug Hellmann proposed openstack/releases: run yamllint on deliverable files during validation https://review.openstack.org/284120 | 17:25 |
| *** david-lyle has joined #openstack-release | 17:31 | |
| *** gordc has quit IRC | 17:44 | |
| dims | ugh something went stale to break requirements jobs - http://logs.openstack.org/53/284553/1/check/gate-requirements-python27/bad7d74/testr_results.html.gz | 17:45 |
| *** amotoki has quit IRC | 17:56 | |
| *** gordc has joined #openstack-release | 18:47 | |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test final release https://review.openstack.org/284715 | 18:49 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test second release candidate https://review.openstack.org/284714 | 18:49 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: test first release candidate https://review.openstack.org/284713 | 18:49 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: check for retagging in validation https://review.openstack.org/284869 | 18:49 |
| openstackgerrit | Doug Hellmann proposed openstack/releases: support tarball names that do not match the repository name https://review.openstack.org/284870 | 18:49 |
| dhellmann | dims : that's a bad version of setuptools. it has been deleted from pypi, but we need infra to remove it from the mirror (I've asked, but not checked if they had time to do it yet) | 18:50 |
| dims | yep, mordred cleaned it up | 18:50 |
| dhellmann | dims : cool, thanks | 18:51 |
| mordred | dhellmann: it would be eversogreat if we could find a productive way to communicating that deleting releases is not a solution | 19:05 |
| mordred | dhellmann: I'm not going to attempt to becuase I'm sure I'll be too abrasive | 19:06 |
| mordred | but if someone who is nicer than me could communicate that too many people depend too heavily on this stuff for reverts that don't come with positive version bumps to work | 19:06 |
| mordred | that would be fantastic | 19:06 |
| mordred | I mean - people running CD on things out there will now have a broken setuptools installed and no real remediation other than manually sending out a downgrade to whatever they're running | 19:08 |
| dhellmann | mordred : perhaps we can work together on a blog post about the right way to deal with it? | 19:09 |
| dhellmann | mordred : my guess is posting something to developerworks and then re-publicizing it heavily would get lots of attention | 19:09 |
| mordred | dhellmann: ++ | 19:10 |
| mordred | dhellmann: that sounds like a great idea | 19:11 |
| mordred | because I really do want to communicate to people why it's a bad idea in a way that doesn't come across like I'm a crusty angry greybeard | 19:11 |
| dhellmann | mordred : "So you've released a broken package to PyPI. What now?" | 19:11 |
| mordred | yah | 19:11 |
| mordred | step one: relax. it happens. | 19:12 |
| dhellmann | mordred : https://etherpad.openstack.org/p/so-youve-released-a-broken-package-to-pypi | 19:12 |
| mordred | step two: realize that just like with email, there is no way to recall a message once it's been sent (sorry MSFT, that's not really a thing) | 19:12 |
| *** cody-somerville has joined #openstack-release | 19:22 | |
| *** cody-somerville has joined #openstack-release | 19:22 | |
| mordred | dhellmann: it seems short - but I don't have a ton more to say that isn't just repeating with more words | 19:25 |
| dhellmann | mordred : a specific example might be good? I'll noodle on it a bit, too | 19:26 |
| mordred | yah | 19:28 |
| *** krotscheck is now known as krotscheck_dcm | 19:50 | |
| dims | dhellmann : last 2 examples - eventlet 0.18.1 and today's setuptools 20.2.1 | 20:29 |
| dhellmann | mordred , dims : I did a little work on it, added an intro and some more words. I'm not sure there's a lot else to say, though, as you point out. | 20:53 |
| mordred | dhellmann: I think your edits are great | 20:55 |
| dhellmann | mordred : where should we publish it? do you know the developerworks folks? | 20:55 |
| mordred | I do not know them | 20:55 |
| dhellmann | we could put it on one of our blogs, too, of course | 20:55 |
| dhellmann | maybe that would come out better, actually | 20:55 |
| mordred | maybe yours? you've got more general exposure in the python world | 20:56 |
| mordred | and I can tweet/retweet it | 20:56 |
| dhellmann | yeah, I can throw a co-written-by line in there | 20:56 |
| mordred | ++ | 20:56 |
| dhellmann | mordred : can you see https://doughellmann.com/blog/?p=4025&preview=true | 20:58 |
| mordred | dhellmann: nope | 20:59 |
| dhellmann | hmm, ok. I'm linking your name to inaugust.com, is that the best site? | 21:00 |
| dhellmann | mordred : I updated the etherpad with a formatted version | 21:04 |
| dhellmann | to which I also made a few further tweaks in wording | 21:04 |
| lifeless | more folk delete-to-fixing ? | 21:05 |
| mordred | dhellmann: yah | 21:05 |
| dhellmann | lifeless : setuptools 20.2.1 | 21:05 |
| mordred | dhellmann: it all just disappeared from the etherpad | 21:05 |
| lifeless | yeah, there's going to be repeated disruption there while we *finally* actually get the unified parser in place | 21:05 |
| dhellmann | gah | 21:05 |
| mordred | oh - there it is | 21:05 |
| lifeless | but deleting aint the answer | 21:06 |
| mordred | lifeless: yes. this is essentially what we've written :) | 21:06 |
| mordred | lifeless: doug helped me turn my virtiol and mouth-foaming into something possibly productive | 21:07 |
| dhellmann | mordred : ok, try again | 21:07 |
| dhellmann | lifeless : https://etherpad.openstack.org/p/so-youve-released-a-broken-package-to-pypi | 21:07 |
| dhellmann | mordred : I'll wait to give you a chance to read the final draft before I hit publish. | 21:09 |
| mordred | dhellmann: I still can't see the thing on doughellman.com | 21:09 |
| dhellmann | mordred : I updated the contents of the etherpad correctly | 21:09 |
| lifeless | +1 | 21:10 |
| dhellmann | so those match what I have in my wordpress editor window | 21:10 |
| mordred | cool | 21:10 |
| mordred | I think it looks great | 21:10 |
| dhellmann | ok, I am clicking publish now | 21:10 |
| mordred | woot | 21:10 |
| dhellmann | https://doughellmann.com/blog/2016/02/25/so-youve-released-a-broken-package-to-pypi-what-do-you-do-now/ | 21:10 |
| mordred | and tweeted/retweeted | 21:12 |
| dhellmann | several retweets already | 21:21 |
| dims | ouch. late to the party. sorry had to go get some food :) | 21:37 |
| dims | dhellmann : one library left to release oslo.messaging, hoping to do that late evening today | 22:07 |
| openstackgerrit | Merged openstack/release-test: Fix the bug URL in README.rst https://review.openstack.org/279719 | 22:20 |
| jroll | dhellmann: good stuff :) | 22:25 |
| openstackgerrit | Lingxian Kong proposed openstack/releases: Release mistralclient 1.3.0 https://review.openstack.org/284975 | 22:54 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 23:23 | |
| *** mriedem has quit IRC | 23:23 | |
| *** gordc has quit IRC | 23:29 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!