| *** ericw has joined #openstack-infra | 00:00 | |
| *** boris-42 has quit IRC | 00:02 | |
| *** ericw has quit IRC | 00:02 | |
| *** UtahDave has joined #openstack-infra | 00:11 | |
| *** torgomatic has quit IRC | 00:22 | |
| *** CaptTofu has joined #openstack-infra | 00:22 | |
| *** torgomatic has joined #openstack-infra | 00:23 | |
| *** ericw has joined #openstack-infra | 00:25 | |
| *** dims has joined #openstack-infra | 00:28 | |
| *** CaptTofu has quit IRC | 00:34 | |
| *** ericw has quit IRC | 00:37 | |
| *** CaptTofu has joined #openstack-infra | 00:51 | |
| *** UtahDave has quit IRC | 00:54 | |
| pabelanger | mordred, odd, looks like saltstack removed your patch https://github.com/emonty/salt/commit/3867ffb3efff2df97b4bb82b9fb2afa703786a47 | 00:55 |
|---|---|---|
| pabelanger | here is the revert: https://github.com/saltstack/salt/pull/3933 | 00:55 |
| pabelanger | not sure why, but trying to get it merged back in | 00:55 |
| pabelanger | since I am hitting the same m2crypto issue | 00:56 |
| *** xchu has joined #openstack-infra | 00:59 | |
| morganfainberg | pabelanger, iirc there were a couple things that pycrypto at the time couldn't do | 01:13 |
| morganfainberg | pabelanger, i _think_ they plan on getting rid of m2crypto soon(ish) again. | 01:14 |
| morganfainberg | pabelanger, i asked about this recently because i was trying to deploy saltstack on a venv on ubuntu | 01:15 |
| pabelanger | morganfainberg, Ya, the same. I am trying to add floating ip support to salt-cloud. Ended up hacking round the issue with tox and a forked repo on github | 01:16 |
| pabelanger | Added -e git+https://github.com/ametaireau/M2Crypto#egg=M2Crypto-0.21.1 to my requirements.txt file | 01:17 |
| morganfainberg | pabelanger, yeah | 01:17 |
| fungi | ick | 01:20 |
| *** dims has quit IRC | 01:31 | |
| *** dcramer_ has quit IRC | 01:32 | |
| *** cody-somerville has joined #openstack-infra | 01:51 | |
| mordred | pabelanger: http://tarballs.openstack.org/nova/M2Crypto-0.21.2.tar.gz | 02:01 |
| pabelanger | mordred, Nice, thanks | 02:02 |
| mordred | pabelanger: there's the patched version of M2Crypto we used in OpenStack before we were able to get off of it | 02:02 |
| *** CaptTofu has quit IRC | 02:02 | |
| *** boris-42 has joined #openstack-infra | 02:03 | |
| *** thomasm has joined #openstack-infra | 02:03 | |
| mordred | fungi, clarkb: around? | 02:07 |
| clarkb | ish. just arrived in portland on my way to bend | 02:08 |
| clarkb | the weather hates me | 02:08 |
| mordred | clarkb: small change: https://review.openstack.org/#/c/48805/ - any objection to me ninja-ing it? | 02:09 |
| mordred | clarkb: enjoy bend btw! | 02:10 |
| *** guohliu has quit IRC | 02:11 | |
| clarkb | mordred no oppisition from me. do tou need a nodepool scripts in config update too? | 02:11 |
| mordred | clarkb: do ? | 02:11 |
| mordred | do I? | 02:11 |
| * mordred goes to look | 02:12 | |
| *** CaptTofu has joined #openstack-infra | 02:12 | |
| clarkb | mordred: for caching | 02:12 |
| mordred | clarkb: eventually, just getting it in to d-g though will at least get it on the node before the job starts | 02:12 |
| mordred | thanks for reminding me though | 02:12 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Add pypi-mirror to the devstack-gate cache list https://review.openstack.org/48812 | 02:14 |
| *** CaptTofu has quit IRC | 02:17 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add pypi-mirror to the devstack-gate cache list https://review.openstack.org/48812 | 02:17 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/devstack-gate: Use select-mirror in devstack-gate https://review.openstack.org/48064 | 02:21 |
| *** wchrisj_ has joined #openstack-infra | 02:27 | |
| *** thomasm has quit IRC | 02:32 | |
| *** thomasm has joined #openstack-infra | 02:35 | |
| *** thomasm has quit IRC | 02:52 | |
| *** mrda has joined #openstack-infra | 03:07 | |
| *** guohliu has joined #openstack-infra | 03:18 | |
| *** alexpilotti has quit IRC | 03:21 | |
| *** mordredtest has joined #openstack-infra | 03:22 | |
| mordredtest | ok. this is really weird | 03:22 |
| mordredtest | thunderbird has grown IRC support | 03:22 |
| mordredtest | and twitter | 03:22 |
| mordredtest | and XMPP | 03:22 |
| morganfainberg | mordredtest, it … wha? | 03:22 |
| mordredtest | yeah | 03:22 |
| mordredtest | I mean | 03:22 |
| mordredtest | I'm talking to you right | 03:23 |
| mordredtest | from | 03:23 |
| mordredtest | THUNDERBIRD | 03:23 |
| morganfainberg | hehe, that is just bizzare | 03:23 |
| mordredtest | yah | 03:23 |
| mordred | mordredtest: I wonder if thunderbird does notifications | 03:23 |
| mordred | nope | 03:23 |
| morganfainberg | that is unfortunate | 03:23 |
| morganfainberg | kind of a serious gap for any reasonable irc client | 03:24 |
| mordredtest | it does underline names | 03:24 |
| mordredtest | when someone spoke to me | 03:24 |
| mordredtest | but | 03:24 |
| mordredtest | I mean | 03:24 |
| mordredtest | why? | 03:24 |
| *** vipul is now known as vipul-away | 03:24 | |
| morganfainberg | mordredtest, not sure, but i think i need to try this now. | 03:24 |
| morganfainberg | just for the novelty | 03:24 |
| mordred | it's worth looking at at least | 03:25 |
| *** AlexF has joined #openstack-infra | 03:25 | |
| mordred | I do enjoy that their twitter support looks like IRC | 03:25 |
| *** rockyg has quit IRC | 03:25 | |
| lifeless | mordred: there are console clients that do that too | 03:25 |
| lifeless | mordred: that said, wtf; wasn't the value proposition of thunderbird to be 'do one thing well' ? | 03:26 |
| mordred | lifeless: yup to both things | 03:26 |
| morganfainberg | lifeless, ++ | 03:26 |
| morganfainberg | isn't that a value proposition for most good apps these days? | 03:26 |
| mordred | until they decide to do all the other things | 03:27 |
| mordred | dhellmann: could we cut a new release of oslo.sphinx please? | 03:27 |
| lifeless | mordred: pony request | 03:27 |
| lifeless | mordred: I'd like something to nag when branches have unreleased code on them | 03:27 |
| mordred | lifeless: I'm trying follow up on getting rid of the last things in openstack that depend on d2to1 transitively | 03:27 |
| *** rickerc_ has joined #openstack-infra | 03:28 | |
| mordred | lifeless: speaking of - could I have a new diskimage-builder release? | 03:28 |
| *** guitarza1 has joined #openstack-infra | 03:28 | |
| mordred | lifeless: and I updated this: https://review.openstack.org/#/c/42514/ and would like it merged and relased too | 03:28 |
| *** cherriges has joined #openstack-infra | 03:29 | |
| *** rickerc has quit IRC | 03:29 | |
| *** Protux has quit IRC | 03:29 | |
| *** odyi has quit IRC | 03:29 | |
| *** SpamapS has quit IRC | 03:29 | |
| *** guitarzan has quit IRC | 03:29 | |
| *** Protux has joined #openstack-infra | 03:29 | |
| *** rickerc_ is now known as rickerc | 03:29 | |
| *** cherriges is now known as odyi | 03:29 | |
| *** odyi has quit IRC | 03:29 | |
| *** odyi has joined #openstack-infra | 03:29 | |
| *** SpamapS has joined #openstack-infra | 03:29 | |
| *** SpamapS has quit IRC | 03:29 | |
| *** SpamapS has joined #openstack-infra | 03:29 | |
| lifeless | mordred: point me at the how-to-release docs again ? | 03:29 |
| *** AlexF has quit IRC | 03:29 | |
| mordred | lifeless: git tag -s $tagname ; git push gerrit $tagname | 03:30 |
| mordred | (the -s is important, because it signs the tag, and non-signed tags are not recognized as being real) | 03:30 |
| *** mordredtest has quit IRC | 03:32 | |
| mordred | WOW. | 03:33 |
| mordred | morganfainberg: the thunderbird twitter thing does not have tab completion of nicks | 03:34 |
| mordred | oh - yes it does. its just slow | 03:34 |
| morganfainberg | mordred, that would be a serious failing if didn't have that. | 03:34 |
| mordred | morganfainberg: yeah. it seems to be working onw | 03:35 |
| mordred | now | 03:35 |
| * mordred has removed the IRC connection, but is going to keep the twitter one for a bit to see how that works out | 03:35 | |
| morganfainberg | maybe had to cache a list followers/follwoed | 03:35 |
| *** ericw has joined #openstack-infra | 03:37 | |
| mordred | prolly | 03:37 |
| *** ericw has quit IRC | 03:41 | |
| lifeless | mordred: is that a addon or core thunderbird ? | 03:47 |
| lifeless | mordred: anyhow, what do you think of my nag thing pony request? | 03:47 |
| mordred | lifeless: might be premature for the openstack projects that have 6 months cycles | 03:47 |
| mordred | lifeless: I liek the theory-ish though | 03:48 |
| lifeless | mordred: we could set a release policy per branch | 03:48 |
| mordred | lifeless: nod | 03:48 |
| lifeless | mordred: like 'continuous', 'more than X weeks inventory', 'every X weeks' | 03:48 |
| mordred | lifeless: would be helpful for things that get sporadic and potentially low amounts of dev | 03:48 |
| mordred | like pbr and git-review | 03:49 |
| mordred | there's a patch in pbr right now that fixes a bug for someone, for instance | 03:49 |
| mordred | but in general, I hope we do not do much movement on pbr | 03:49 |
| mordred | so a 'nag me about getting a release done' makes sense | 03:49 |
| *** HenryG has quit IRC | 03:52 | |
| lifeless | I also want a semver automatic release script | 03:54 |
| lifeless | where should first drafts of these be put ? | 03:54 |
| mordred | what's semver? | 03:57 |
| mordred | and I'd make patches to config - because you'd need zuul config to trigger them | 03:57 |
| clarkb | semantic versioning. why not put the nag script in as a commit hook locally? | 03:57 |
| mordred | so, I'd think it would want to be a script in slave_scripts, with a job(s) in jjb config and a zuul entry | 03:57 |
| mordred | clarkb: sure, but I mean, what does "semver automatic release script" mean? | 03:58 |
| clarkb | I'm not sure that a script that bugs you is a general thing everyone eants | 03:58 |
| mordred | clarkb: I agree. but I could see it being an opt-in script that a project could request | 03:59 |
| mordred | I'd opt in to it for pbr, for instance | 04:00 |
| mordred | I forget that I have local patches installed on my laptop all the time | 04:00 |
| *** anteaya has joined #openstack-infra | 04:01 | |
| mordred | lifeless: can you explain "semver automatic release script" | 04:01 |
| lifeless | clarkb: because releasing is not coupled to the person doing the commit | 04:01 |
| mordred | lifeless: the individual words all make sense, but the combination of them do not | 04:01 |
| anteaya | I'm fine with a script that bugs mordred | 04:01 |
| lifeless | mordred: oh so | 04:01 |
| mordred | (btw, I would love it if they would remove the use of - from the semver spec) | 04:02 |
| mordred | since that had well documented uses that conflict with his WELL before he wrote that | 04:02 |
| lifeless | mordred: when I went to release dib | 04:03 |
| lifeless | mordred: I had to: | 04:03 |
| lifeless | pull | 04:03 |
| lifeless | check existing tags | 04:03 |
| lifeless | assess whether there was forward or backwards incompatible changes | 04:03 |
| lifeless | pick a new number based on the last two items + semver rules. | 04:03 |
| lifeless | do the release. | 04:03 |
| lifeless | mordred: so I'd like something that: | 04:04 |
| mordred | yes. I agree | 04:04 |
| mordred | that would be a good script | 04:04 |
| lifeless | - lets me say 'diskimage-builder minor' | 04:04 |
| lifeless | and it goes and does it all | 04:04 |
| mordred | yuip | 04:04 |
| lifeless | in fact even that, do-release openstack/diskimage-builder major/minor/point. | 04:04 |
| mordred | I'd like it to be able to suggest something | 04:05 |
| mordred | by looking at the api between the two, and telling me something is different | 04:05 |
| lifeless | so putting scripts that one might run locally too in config is a little weird | 04:05 |
| mordred | indeed. if the script itself has general use outside of config | 04:06 |
| mordred | then putting it elsewhere is a good idea | 04:06 |
| mordred | do you think the nag script has general use? | 04:06 |
| lifeless | a 'check X and report whether it should do a release' script seems general. | 04:06 |
| mordred | (I think the release code wants to live in pbr, fwiw, because I already have some of the other bits in there) | 04:06 |
| lifeless | glueing that into projects.yaml is openstack specific. | 04:06 |
| mordred | yup | 04:07 |
| *** ericw has joined #openstack-infra | 04:08 | |
| lifeless | so | 04:09 |
| lifeless | is there another infra tree it might go in - a generic release/code gardening kinda place? | 04:09 |
| lifeless | mordred: what - in the semver spec? | 04:11 |
| mordred | lifeless: for pre-releases | 04:11 |
| lifeless | oh | 04:12 |
| lifeless | point 11 | 04:12 |
| mordred | and 9 | 04:12 |
| mordred | I've considered filing a bug | 04:12 |
| mordred | but everytime I've ever given feedback to anyone who works at github, I've been met with MASSIVE arrogance and disdain | 04:13 |
| *** ericw has quit IRC | 04:13 | |
| mordred | so I preemptively give up | 04:13 |
| mordred | I'd love it if semver was usable by linux hackers | 04:13 |
| mordred | and I'd love it if someone else would fix hat | 04:13 |
| mordred | that | 04:13 |
| mordred | as soon as they do, I will promote it with gusto | 04:13 |
| mordred | in the meantime, I'm happy to say "versioing similar to semver rules, except with different syntax choices" | 04:14 |
| lifeless | is there a common prerelease ruleset for 'linux' ? | 04:14 |
| lifeless | AFAIK only deb has a prerelease syntax at all. | 04:14 |
| mordred | nope | 04:15 |
| mordred | and you're right | 04:15 |
| mordred | I like pep440's approach, since it's not in conflict with existing schemes | 04:15 |
| mordred | N[.N]+[{a|b|c|rc}N][.postN][.devN] | 04:15 |
| mordred | if you choose to use it, you don't produce syntactically incorrect versions for other things | 04:16 |
| mordred | hrm. it's on github | 04:17 |
| mordred | maybe I'll just fork it | 04:17 |
| lifeless | hmm, I'm not sure pep440 will be interepreted correctly on debian | 04:18 |
| lifeless | but I haven't finished reading the full thing yet | 04:18 |
| lifeless | (by dpkg --compare-versions specifically) | 04:19 |
| lifeless | yup, it's incompatible | 04:19 |
| lifeless | mordred: the prerelease format in pep440 is interpreted as postrelease by debian. | 04:20 |
| mordred | lifeless: I'm certain that it's incompatible | 04:20 |
| mordred | lifeless: in terms of version math | 04:20 |
| mordred | lifeless: but semver is actually a syntax error on rpm | 04:21 |
| lifeless | mordred: thats fine | 04:21 |
| lifeless | mordred: IMNSHO | 04:21 |
| *** vipul-away is now known as vipul | 04:22 | |
| lifeless | mordred: since AFAIK rpm doesn't have pre-release support, it's appropriate to error when mapping that to rpm. | 04:22 |
| lifeless | mordred: and for debian it should be mapped to ~ | 04:22 |
| mordred | what would be GREAT is if we could actually start to align these in to one system | 04:23 |
| mordred | let - be reserved to indicate (re)distributor additions and patches | 04:23 |
| lifeless | mmm | 04:23 |
| mordred | pep440's use of a restricted set of alpha chars to denote prerelease is nice | 04:24 |
| mordred | and is something that could be _added_ to debian version math | 04:24 |
| lifeless | as long as there is a way to compile any given versioning scheme to (debian, python, rpm) | 04:24 |
| lifeless | then I'm happy | 04:24 |
| mordred | yah | 04:24 |
| lifeless | I really don't give a rats whether there is one or a dozen different schemes | 04:25 |
| mordred | I do, because is just assinine | 04:25 |
| mordred | s/is/it is/ | 04:25 |
| lifeless | pep440 is a meta-scheme | 04:25 |
| lifeless | it's the union of everything | 04:26 |
| mordred | yup. but also a paring down to something comprehensile | 04:26 |
| mordred | as a format, I think it's great | 04:26 |
| mordred | as a set of rules for _when_ you change things, I think semver is great | 04:26 |
| *** zaro0508 has joined #openstack-infra | 04:26 | |
| mordred | what I think would be fantastic is the union of the two of them | 04:26 |
| lifeless | right, but it doesn't describe backwards compat / api change/bugfix interactions | 04:27 |
| lifeless | so | 04:27 |
| lifeless | to me semver (and the libtool version rules which are similar) is in one problem space | 04:27 |
| lifeless | and pep440 is another | 04:27 |
| mordred | I agree with you (and I believe that's what I just said) | 04:27 |
| lifeless | and deb/rpm are closer to pep440 than semver | 04:28 |
| lifeless | sure | 04:28 |
| lifeless | not debating | 04:28 |
| mordred | k | 04:28 |
| mordred | :) | 04:28 |
| lifeless | just thinking out loud | 04:28 |
| lifeless | the reason I say I don't care if tere are a dozen schemes in the semver space | 04:28 |
| lifeless | is that if there are a dozen such schemes, there are a dozen distinct versioning-driven-by-code-change-categorisation policies | 04:29 |
| lifeless | and so that makes sense | 04:29 |
| *** mrda has quit IRC | 04:29 | |
| lifeless | the primary thing is that it be possible to compile the output of such a scheme to the target platforms; if you can't, live is very very unpleasant | 04:30 |
| lifeless | mordred: https://fedoraproject.org/wiki/Packaging:NamingGuidelines#Pre-Release_packages | 04:30 |
| lifeless | mordred: so there does appear to be -a- way for rpm | 04:33 |
| mordred | https://github.com/emonty/semver | 04:37 |
| mordred | there's my first step towards a modification | 04:37 |
| mordred | I have not yet updated the BNF grammar | 04:39 |
| *** ericw has joined #openstack-infra | 04:39 | |
| mordred | but that's possibly because I do not care about the BNF grammer :) | 04:39 |
| *** ericw has quit IRC | 04:44 | |
| lifeless | mordred: so I think it works, but I still don't see the point of changing it :) | 04:45 |
| mordred | lifeless: because I don't see the original point of using a - when there were other options | 04:54 |
| mordred | lifeless: it's like, if you're going to make something new and wonderful, you could also have the humility to at least pretend to look at the existing world a LITTLE bit | 04:55 |
| mordred | and maybe not be compatible with _everyone_ | 04:55 |
| mordred | but with a mild amount of work, you could be compatible with _someone)_ | 04:55 |
| mordred | but it depends on whether or not you want to collaborate | 04:56 |
| mordred | or make a new shiny thing that causes everybody to follow your choices | 04:56 |
| lifeless | mordred: hmm | 05:03 |
| lifeless | mordred: so semver to me seems pretty natural | 05:03 |
| lifeless | mordred: But thats my Debian background probably. | 05:03 |
| mordred | lifeless: the - in semver seems massively arrogant to me | 05:03 |
| lifeless | mordred: hyphens in Debian separate different parts of the version | 05:04 |
| lifeless | mordred: debian and upstream | 05:04 |
| mordred | right. but debian and redhat both use the - for that purpose | 05:04 |
| mordred | so there is already an ecosystem meaning for the - separatoin | 05:04 |
| mordred | a very clear one | 05:04 |
| lifeless | right, but since you need a compile step from semver to most other things /anyway/ | 05:05 |
| lifeless | the - isn't a problem, and the conceptual separation is something everyone is used to. | 05:05 |
| mordred | it's an interesting point | 05:05 |
| mordred | I think for me I'd hope that semver would be step one | 05:05 |
| mordred | and the distros starting to grok it would be step two | 05:06 |
| mordred | with an eventual goal being that there is no compile step | 05:06 |
| mordred | so anything that adds another need for a compile step | 05:06 |
| mordred | to me is a bug | 05:06 |
| mordred | because I want to get rid of the compile steps | 05:06 |
| lifeless | mordred: Thats a lovely dream, I don't think it's feasible - look at pep440 which introduces *yet another* such compiler | 05:06 |
| mordred | but pep440 + semver + using - separation for distro additions | 05:07 |
| mordred | could actually be an easy to understand, easy to parse set of rules | 05:08 |
| lifeless | mordred: doesn't help at all, because you have to map prerelease across, which requires represenation specific interpreters | 05:08 |
| mordred | I disagree | 05:08 |
| lifeless | e.g. in deb ~, in rpm the semver prerelease goes into the 'distro additions' segment | 05:08 |
| mordred | right. so, if we get a sane upstream version thing that people can use | 05:09 |
| mordred | then next step is to get the distros to start groking | 05:09 |
| mordred | there is no compelling reason for them to re-align to anything right now | 05:09 |
| mordred | because there is nothing to realign to | 05:09 |
| mordred | it would be pointless | 05:09 |
| lifeless | the problem is that distros are the union of all requirements | 05:09 |
| lifeless | so there si there is little reason for them to realign ever | 05:10 |
| mordred | I believe you and I have different worldviews and different goals, perhaps | 05:10 |
| lifeless | they can / will just map new things into their system, and adjust their system just-enough to cope | 05:10 |
| mordred | I view things like the divergence between the version systems used by debian and redhat as a fundamental bug in free software | 05:10 |
| mordred | not a feature | 05:10 |
| mordred | or an interesting quirk | 05:10 |
| *** ericw has joined #openstack-infra | 05:10 | |
| lifeless | I don't view it as a feature, thats for sure | 05:11 |
| mordred | these things should communicate | 05:11 |
| mordred | which is why I almost can be a fan of semver | 05:11 |
| lifeless | but there's a bunch of pragmatic issues even if the syntax and interpreters were identical that make direct reuse of versions problematic | 05:11 |
| lifeless | sonames, for instance. | 05:11 |
| mordred | get rid of the -, and I'm a fanboi and would talk about it at converences | 05:11 |
| mordred | conferences | 05:11 |
| lifeless | can I paraphrase? 'The - in semver makes version numbers ambiguous when combined with rpm or dpkg versions and thats bad' | 05:12 |
| mordred | yes | 05:13 |
| mordred | that is an excellent paraphrase | 05:13 |
| mordred | if I see a - in a version, I assume there is a distro part | 05:13 |
| mordred | which may be _my_ debian background coming through :) | 05:13 |
| *** ericw has quit IRC | 05:15 | |
| mordred | lifeless: thanks. I've used that sentence in an update I just pushed | 05:16 |
| mordred | I think that very clearly describes the impetus | 05:16 |
| lifeless | ok, so - I get that, but given the situation *today* it's not a functional problem. The question is 'would it be a functional problem if distros wanted to support it directly'? | 05:16 |
| lifeless | And | 05:16 |
| lifeless | 'will it confuse users'? | 05:16 |
| lifeless | The confusion of users is a significant aspect. | 05:16 |
| lifeless | And one I hadn't really cottoned onto | 05:17 |
| mordred | functional problem I think is solvable in a few different ways | 05:17 |
| mordred | the confusion I believe I now understand is the thing that was bugging me | 05:17 |
| mordred | I like it when conversations lead to revalations! | 05:18 |
| lifeless | :) | 05:18 |
| mordred | I think I'm going to more explicitly steal from pep440 in mine by indicating that pre-releases only really can use a|b|c|rc | 05:19 |
| mordred | because, again, since the issue is about potential confusion | 05:19 |
| mordred | I think it's worth having actual rules there, rather than "put whatever you want after the - it doesn't really carry meaning" | 05:20 |
| mordred | or, it carries meaning due to sorting rules, but none that's immediately and clearly apparent | 05:20 |
| mordred | to a human | 05:20 |
| lifeless | so you can say a1 right ? | 05:20 |
| mordred | yup | 05:20 |
| mordred | pre-release version section is indicated by (a|b|c|rc)[0-9]* | 05:21 |
| mordred | and is considered a modifier on the thing to its left | 05:21 |
| mordred | so 1.2.3.a3.a4 would be the fourth alpha release leading towards the third alpha release of 1.2.3 | 05:22 |
| mordred | and is silly | 05:22 |
| mordred | but possible if you need it | 05:22 |
| mordred | nah. if this is intended for clarity, then I think it needs to be clear like the first three | 05:24 |
| mordred | and really specific | 05:24 |
| mordred | I think that having one and only one pre-release section is clearer, even if parsers could parse a more complex thing | 05:24 |
| * anteaya has had enough fun for one day and will be going to sleep now, thanks | 05:25 | |
| *** anteaya has quit IRC | 05:25 | |
| mordred | bah! | 05:25 |
| mordred | you can never have enough fun when talking about versioning schemes! | 05:25 |
| lifeless | I agree, but interpretation #2 | 05:25 |
| mordred | lifeless: can you restate what you think? | 05:27 |
| lifeless | no matter what you do, talking about versioning schemes will never be enough fun | 05:28 |
| lifeless | mordred: last two lines were just humour | 05:29 |
| mordred | :) | 05:32 |
| *** AlexF has joined #openstack-infra | 05:40 | |
| lifeless | mordred: so what is oslo.version ? | 05:40 |
| mordred | lifeless: the version consuming portion of pbr broken out into its own library | 05:41 |
| mordred | so that pbr can just be the build portion | 05:41 |
| mordred | and oslo.version the runtime portion | 05:41 |
| mordred | either should be able to be used without the other | 05:41 |
| lifeless | ok | 05:41 |
| lifeless | so it reads pkg-info etc | 05:41 |
| lifeless | ? | 05:41 |
| mordred | yup | 05:41 |
| mordred | as well as /etc/$project/version.ini (for the distro override thing) | 05:42 |
| *** ericw has joined #openstack-infra | 05:42 | |
| mordred | althoguh I think that needs clearer semantics, tbh | 05:42 |
| lifeless | yeah, though I still don't get the distro thing :P | 05:43 |
| lifeless | bbiab | 05:43 |
| *** ericw has quit IRC | 05:46 | |
| *** davidhadas_ has joined #openstack-infra | 05:50 | |
| mordred | lifeless, clarkb, ttx: ok. I made a bunch of updates to https://github.com/emonty/semver/blob/master/semver.rst | 05:56 |
| mordred | plus side - openstack's versioning is mostly compliant | 05:57 |
| mordred | maybe I should move that file into the pbr tree | 05:57 |
| mordred | and then make a tool like lifeless was requesting earlier which will make it easy to make a local tag that matches types of the above rules | 05:57 |
| *** SergeyLukjanov has joined #openstack-infra | 06:04 | |
| *** AlexF has quit IRC | 06:06 | |
| *** ericw has joined #openstack-infra | 06:13 | |
| lifeless | mordred: you should | 06:14 |
| lifeless | mordred: 'tox' in pbr just hated on me | 06:15 |
| lifeless | Downloading/unpacking argparse (from -r /home/robertc/work/nova/requirements.txt (line 5)) | 06:15 |
| lifeless | You are installing an externally hosted file. Future versions of pip will default to disallowing externally hosted files. | 06:15 |
| lifeless | is one thing | 06:15 |
| lifeless | oh, wait. | 06:16 |
| lifeless | EWTF am I doing in the nova tree | 06:16 |
| lifeless | so it failed on boto etc, but thats irrelevant as I was just nuts | 06:16 |
| *** ericw has quit IRC | 06:17 | |
| lifeless | mordred: http://paste.ubuntu.com/6169970/ | 06:19 |
| lifeless | mordred: any thoughts ? | 06:19 |
| mordred | lifeless: looking | 06:27 |
| *** zaro0508 has quit IRC | 06:28 | |
| mordred | lifeless: are you running that outside of a git repo? and/or is git there? | 06:28 |
| lifeless | there is git there and there is a repo | 06:28 |
| lifeless | my global config deliberately has no user details | 06:28 |
| lifeless | so I don't commit as <personal> in a work repo or vice verca | 06:28 |
| mordred | gotcha | 06:29 |
| lifeless | sounds like pbr isn't isolating itself from the environment enough | 06:29 |
| mordred | I wonder if that is causing a problem | 06:29 |
| mordred | yeah | 06:29 |
| lifeless | fixtures.TempHomeDir + explicit git global config | 06:29 |
| lifeless | should do the ticket | 06:29 |
| lifeless | I will spin that up shortly | 06:29 |
| lifeless | but first, I've had a fully 'on' day, time for a little time out. | 06:30 |
| mordred | mmm. yes. time out | 06:31 |
| mordred | I'm snacking and beering - if you don't have it sorted when I next am on, I'll also take a look | 06:32 |
| *** ericw has joined #openstack-infra | 06:43 | |
| *** AlexF has joined #openstack-infra | 06:45 | |
| *** guohliu has quit IRC | 06:49 | |
| *** ericw has quit IRC | 06:52 | |
| *** salv-orlando has quit IRC | 06:52 | |
| *** tkammer has joined #openstack-infra | 06:55 | |
| *** afazekas_zz is now known as afazekas | 06:59 | |
| *** tvb has quit IRC | 07:03 | |
| *** tvb has joined #openstack-infra | 07:03 | |
| *** tvb has quit IRC | 07:03 | |
| *** tvb has joined #openstack-infra | 07:03 | |
| *** AlexF has quit IRC | 07:05 | |
| *** ericw has joined #openstack-infra | 07:18 | |
| *** SergeyLukjanov has quit IRC | 07:22 | |
| *** ericw has quit IRC | 07:23 | |
| *** guohliu has joined #openstack-infra | 07:24 | |
| *** alexpilotti has joined #openstack-infra | 07:46 | |
| *** tvb has quit IRC | 07:48 | |
| lifeless | mordred: isn't it uhm 0230 for you ? | 07:48 |
| lifeless | ish | 07:48 |
| *** thingee is now known as thingee_zzz | 07:52 | |
| openstackgerrit | lifeless proposed a change to openstack-dev/pbr: Fix test_changelog when git isn't globally setup. https://review.openstack.org/48828 | 07:55 |
| lifeless | mordred: ^ | 07:55 |
| *** SergeyLukjanov has joined #openstack-infra | 08:07 | |
| *** tvb has joined #openstack-infra | 08:14 | |
| *** tvb has quit IRC | 08:14 | |
| *** tvb has joined #openstack-infra | 08:14 | |
| *** kpavel has joined #openstack-infra | 08:18 | |
| *** kpavel_ has joined #openstack-infra | 08:19 | |
| *** kong has joined #openstack-infra | 08:22 | |
| *** kpavel has quit IRC | 08:23 | |
| *** kpavel_ is now known as kpavel | 08:23 | |
| openstackgerrit | lifeless proposed a change to openstack-dev/pbr: Switch away from tearDown for BaseTests. https://review.openstack.org/48831 | 08:45 |
| openstackgerrit | lifeless proposed a change to openstack-dev/pbr: Add a test for command registration. https://review.openstack.org/48833 | 08:59 |
| *** yonglihe_ has quit IRC | 09:02 | |
| *** cody-somerville has quit IRC | 09:15 | |
| *** mrda has joined #openstack-infra | 09:25 | |
| *** SergeyLukjanov has quit IRC | 09:46 | |
| *** mrda has quit IRC | 10:14 | |
| lifeless | mordred: so pbr seems clearly the wrong place for next version stuff I was ponying around | 10:20 |
| *** SergeyLukjanov has joined #openstack-infra | 10:21 | |
| openstackgerrit | lifeless proposed a change to openstack-dev/pbr: Add a test for command registration. https://review.openstack.org/48833 | 10:21 |
| *** xchu has quit IRC | 10:22 | |
| *** ericw has joined #openstack-infra | 10:28 | |
| *** ericw has quit IRC | 10:38 | |
| *** guohliu has quit IRC | 10:39 | |
| *** alexpilotti has quit IRC | 10:39 | |
| *** tvb has quit IRC | 10:51 | |
| *** stevebaker has quit IRC | 10:57 | |
| *** stevebaker has joined #openstack-infra | 10:57 | |
| *** tvb has joined #openstack-infra | 10:57 | |
| *** tkammer has quit IRC | 11:11 | |
| *** CaptTofu has joined #openstack-infra | 11:21 | |
| *** ericw has joined #openstack-infra | 11:35 | |
| *** ericw has quit IRC | 11:40 | |
| *** SergeyLukjanov has quit IRC | 11:44 | |
| *** hashar has joined #openstack-infra | 11:45 | |
| lifeless | mordred: where do you want bugs on cookiecutter? tox fails in a fresh cut. | 11:55 |
| lifeless | Complete output from command python setup.py egg_info: | 11:55 |
| lifeless | error in setup command: Error parsing /home/robertc/work/git-next-version/setup.cfg: Exception: Versioning for this project requires either an sdist tarball, or access to an upstream git repository. | 11:55 |
| *** alexpilotti has joined #openstack-infra | 12:00 | |
| *** hashar has quit IRC | 12:01 | |
| *** tkammer has joined #openstack-infra | 12:05 | |
| *** ericw has joined #openstack-infra | 12:06 | |
| *** ericw has quit IRC | 12:11 | |
| *** qba73 has joined #openstack-infra | 12:40 | |
| *** tvb has quit IRC | 12:41 | |
| *** che-arne has joined #openstack-infra | 12:44 | |
| *** SergeyLukjanov has joined #openstack-infra | 12:59 | |
| *** SergeyLukjanov has quit IRC | 13:06 | |
| *** nicedice has quit IRC | 13:08 | |
| *** basha has joined #openstack-infra | 13:08 | |
| *** nicedice has joined #openstack-infra | 13:08 | |
| *** basha has quit IRC | 13:17 | |
| *** basha has joined #openstack-infra | 13:18 | |
| *** michchap_ has quit IRC | 13:19 | |
| *** michchap has joined #openstack-infra | 13:22 | |
| *** thomasbiege has joined #openstack-infra | 13:22 | |
| *** michchap has joined #openstack-infra | 13:24 | |
| *** SergeyLukjanov has joined #openstack-infra | 13:30 | |
| *** basha has quit IRC | 13:31 | |
| *** thomasbiege has quit IRC | 13:33 | |
| *** che-arne has quit IRC | 13:35 | |
| *** SergeyLukjanov has quit IRC | 13:39 | |
| *** SergeyLukjanov has joined #openstack-infra | 13:44 | |
| *** SergeyLukjanov has quit IRC | 13:50 | |
| *** Guest98622 is now known as Vivek | 13:56 | |
| *** Vivek has joined #openstack-infra | 13:56 | |
| *** qba73 has quit IRC | 14:11 | |
| *** tjones has joined #openstack-infra | 14:15 | |
| *** tjones has quit IRC | 14:19 | |
| *** guohliu has joined #openstack-infra | 14:19 | |
| *** dims has joined #openstack-infra | 14:28 | |
| *** SergeyLukjanov has joined #openstack-infra | 14:44 | |
| *** schwicht has joined #openstack-infra | 14:57 | |
| *** schwicht has quit IRC | 15:02 | |
| *** ericw has joined #openstack-infra | 15:02 | |
| *** schwicht has joined #openstack-infra | 15:03 | |
| *** thomasm has joined #openstack-infra | 15:11 | |
| *** SergeyLukjanov has quit IRC | 15:18 | |
| *** thomasm has quit IRC | 15:22 | |
| mordred | lifeless: on oslo please | 15:25 |
| *** ericw has quit IRC | 15:32 | |
| *** dcramer_ has joined #openstack-infra | 15:34 | |
| *** kpavel has quit IRC | 15:37 | |
| *** basha has joined #openstack-infra | 15:39 | |
| dhellmann | mordred: oslo.sphinx 1.1? | 15:48 |
| *** dcramer_ has quit IRC | 15:49 | |
| *** basha has quit IRC | 15:50 | |
| mordred | dhellmann: that would be nice - oslo.sphinx 1.0 is still using d2to1 version of pbr | 15:51 |
| mordred | and d2to1 | 15:51 |
| dhellmann | sure, I can do that now -- I was verifying the version would be ok, since I'm not as familiar with the semver rules | 15:52 |
| dhellmann | mordred: looks like the changes are all packaging related, except the intersphinx file thing; nothing should break backwards compatibility | 15:53 |
| * dhellmann crosses fingers | 15:53 | |
| *** salv-orlando has joined #openstack-infra | 15:55 | |
| dhellmann | mordred: 1.1 tag pushed | 15:56 |
| mordred | dhellmann: woot! | 15:56 |
| mordred | dhellmann: I made an openstack friendly semver fork: https://github.com/emonty/semver/blob/master/semver.rst if you want to read :) | 15:57 |
| dhellmann | reading | 15:57 |
| dhellmann | mordred: the bit that always trips me up is where dependency changes fall in terms of "backwards-compatible" | 15:57 |
| mordred | dhellmann: that's an interesting question | 15:58 |
| mordred | dhellmann: I think that if the contract from the thing in question to the thing consuming it changes, it would not be backwards-compatible | 15:58 |
| dhellmann | the cliff/pyparsing thing is a good example | 15:59 |
| dhellmann | I thought bumping pyparsing to 2.0.1 fixed the bug that it wasn't possible to install the same versions of things under py2 and py3 | 15:59 |
| mordred | but what broke in openstack was people's version pins of pyparsing, right? | 16:00 |
| dhellmann | well, yeah, quantumclient should never have had anything specified for that | 16:00 |
| dhellmann | because they were not using it directly | 16:00 |
| dhellmann | and cliff had the right rules in its setup.py | 16:00 |
| mordred | right. I don't think that was a contract violation on your part | 16:00 |
| mordred | I believe it was that quantumclient was reaching in past your contract | 16:01 |
| mordred | and you violated that thing | 16:01 |
| dhellmann | that's what I thought; sdague disagreed and thought I should have used 1.5 instead of 1.4.5 | 16:01 |
| mordred | but- dependency tree as api contract is an interesting topic | 16:01 |
| dhellmann | it doesn't really matter, but I'm trying to understand how people interpret the guidelines to avoid conflicts in the future :-) | 16:02 |
| mordred | sdague: I'd love a dissenting opinion | 16:02 |
| mordred | ++ | 16:02 |
| mordred | _this_ is actually an area that I don't think semver really covers | 16:02 |
| dhellmann | btw, do you know the state of that fix for the stable/grizzly gate? I have a cliff release that does not use pyparsing ready, but as dreamhost has broken gates twice this week with releases to our libs, I'd like to hold off a bit before making another ;-) | 16:02 |
| *** dcramer_ has joined #openstack-infra | 16:03 | |
| dhellmann | I was under the impression someone was going to "magic" the fix to remove the pin for pyparsing in quantumclient in stable | 16:03 |
| mordred | oh, well, we did a couple of quantumclient releases | 16:04 |
| mordred | and it now just depends on neutronclient | 16:04 |
| mordred | but I believe there are some issues with the symbols being exposed that are causing different stable gate failures | 16:05 |
| mordred | and, well, landing things through the gate hasn't been fun for a couple of days | 16:05 |
| dhellmann | mordred: ok, the symbol thing isn't related to cliff or pyparsing, so that's good at least | 16:06 |
| mordred | yah! | 16:06 |
| dhellmann | mordred: does our git setup support grokmirror? | 16:17 |
| mordred | dhellmann: I don't know what grokmirror is (no, but I'd like to learn) | 16:17 |
| dhellmann | mordred: https://www.kernel.org/mirroring-kernelorg-repositories.html | 16:18 |
| dhellmann | I'm looking for a way to get a copy of all of our repositories without scraping the cgit UI | 16:19 |
| dhellmann | https://github.com/mricon/grokmirror | 16:21 |
| jeblair | dhellmann: not to distract from looking at grokmirror which might be a good idea, but you may be able to get what you need immediately with 'ssh review.openstakck.org -p 29418 gerrit ls-projects' | 16:21 |
| mordred | jeblair: so, grokmirror actually looks interesting and I could see us even making use of it in d-g | 16:21 |
| mordred | (on very short examination) | 16:22 |
| mordred | from what it looks like, the server maintains a json file listing the repos and last modified time | 16:22 |
| mordred | (there's a tool for that) | 16:22 |
| mordred | so things wanting to mirror grab the file first, and then just run git remote update on the repos that have actually changed | 16:23 |
| jeblair | mordred: yeah, also putting it on jenkins slaves for faster local cloning. | 16:23 |
| mordred | yup | 16:23 |
| jeblair | anyway... sunday. | 16:23 |
| mordred | totally. have a sunday | 16:23 |
| mordred | dhellmann: tl;dr - no, we don't - but it seems like a worthwhile thing to add to the infrastructure | 16:23 |
| mordred | and possibly make use of ourselves | 16:23 |
| dhellmann | mordred: ok, no hurry, I'll look at the gerrit tip jeblair tossed out (need to figure out why ssh is refusing my connection, probably an identity thing) | 16:24 |
| dhellmann | indeed, forgot the gerrit user name part | 16:24 |
| dhellmann | cool, I can script around this, thanks, jeblair | 16:25 |
| mordred | and in the meantime, you should be able to build a mirror fairly simply with for repo in `ssh review.openstakck.org -p 29418 gerrit ls-projects` ; do mkdir `dirname $repo` ; pushd `dirname $repo` ; git clone git://git.openstack.org/$repo; popd; done | 16:25 |
| mordred | hrm. you need a -p on the mkdir so that it doesn't fail when the dir is already there | 16:25 |
| mordred | but you get the idea | 16:25 |
| dhellmann | mordred: yep | 16:27 |
| mordred | https://bugs.launchpad.net/openstack-ci/+bug/1232799 | 16:30 |
| uvirtbot | Launchpad bug 1232799 in openstack-ci "add grokmirror support" [Undecided,New] | 16:30 |
| *** tvb has joined #openstack-infra | 16:37 | |
| morganfainberg | mordred, phsaw, stop offering "easy" and "in the mean time" solutions :P *plans to make similar script occur locally* | 16:39 |
| mordred | morganfainberg: :) | 16:39 |
| * mordred just realized that if we had grokmirror, I could just run it in my src dir on my laptop | 16:40 | |
| mordred | and always make sure that all of my local repos were up to date | 16:40 |
| * mordred is a bit excited | 16:40 | |
| morganfainberg | mordred, that makes far too much sense. | 16:41 |
| mordred | morganfainberg: yup | 16:41 |
| *** salv-orlando has quit IRC | 16:41 | |
| dstufft | mordred: have you hit any timeouts in pypi today? i know it was failing on thrift (dunno if you ran any jobs today tho) | 16:44 |
| *** tjones has joined #openstack-infra | 16:44 | |
| mordred | dstufft: lemme see if I've run anything that wanted to hit pypi | 16:46 |
| dhellmann | mordred: that's what I wanted it for :-) | 16:47 |
| mordred | http://paste.openstack.org/show/47675 | 16:47 |
| mordred | oops. bug. one sec | 16:48 |
| dhellmann | the mkdir has to come in the else block, no? | 16:48 |
| mordred | the -p should always succeed | 16:48 |
| dhellmann | oh, no, missed the dirname | 16:48 |
| mordred | god, if I really wanted to be awesome, I should have it gitreview -s if it clones | 16:49 |
| dhellmann | gitreview -s? | 16:49 |
| dhellmann | oh, for new repos | 16:50 |
| mordred | http://paste.openstack.org/show/47677 | 16:51 |
| mordred | there we go | 16:51 |
| mordred | that's working | 16:51 |
| mordred | (I have a remote setup for review.o.o port 29418) | 16:51 |
| mordred | Cloning into 'openstack/melange'... | 16:52 |
| mordred | fatal: Could not read from remote repository. | 16:52 |
| mordred | heh | 16:52 |
| mordred | we should really you know, figure out an opinion on that | 16:52 |
| ekarlso | mordred: melange ? | 16:52 |
| ekarlso | ain't that dead ? | 16:52 |
| mordred | yup | 16:52 |
| mordred | but we don't delete things from gerrit | 16:53 |
| ekarlso | ah | 16:53 |
| mordred | problem is, we're not mirroring it to git.o.o, because that's kinda silly | 16:53 |
| mordred | so it's in a weird limbo state | 16:53 |
| mordred | neat! that worked | 16:59 |
| mordred | takes 4 minutes to git remote update every repo in openstack :) | 17:09 |
| *** tkammer has quit IRC | 17:17 | |
| *** thingee_zzz is now known as thingee | 17:25 | |
| *** schwicht has quit IRC | 17:33 | |
| *** tjones has quit IRC | 17:39 | |
| *** davidhadas has quit IRC | 17:53 | |
| *** guitarza1 is now known as guitarzan | 17:58 | |
| sdague | mordred, dhellmann: so my take on why I think the pyparsing req change should have been a Y bump (X.Y.Z land)... | 17:58 |
| *** thingee is now known as thingee_zzz | 17:58 | |
| sdague | is that it was a stepwise X bump in dependency | 17:58 |
| sdague | 1.4.3 required == 1.5.7 and 1.4.4 required > 2.0.1 | 17:59 |
| sdague | there was no overlap in dependencies, so you couldn't satisfy the new software without taking a major upgrade of another package | 17:59 |
| sdague | triggering additional software installation at a major release level doesn't feel like a Z bump to me | 18:01 |
| mordred | ok. I can see that logic | 18:03 |
| sdague | dstufft: yeh, we hit one pypi timeout - http://logstash.openstack.org/#eyJzZWFyY2giOiJcIkVycm9yIFRoZSByZWFkIG9wZXJhdGlvbiB0aW1lZCBvdXQgd2hpbGUgZ2V0dGluZ1wiIiwiZmllbGRzIjpbXSwib2Zmc2V0IjowLCJ0aW1lZnJhbWUiOiI4NjQwMCIsImdyYXBobW9kZSI6ImNvdW50IiwidGltZSI6eyJ1c2VyX2ludGVydmFsIjowfSwic3RhbXAiOjEzODA0Nzc3NjIzMDh9 | 18:03 |
| mordred | perhaps we should add text around that in my new semver fork... | 18:03 |
| sdague | sure | 18:03 |
| sdague | it's all a judgement call, which is why I tried to word the email on list pretty softly about it | 18:04 |
| *** dims has quit IRC | 18:04 | |
| sdague | I just really want to go one week without having to deal with the consequences of dependency upgrades :) | 18:04 |
| mordred | yeah. I agree | 18:05 |
| *** fbo_away is now known as fbo | 18:06 | |
| mordred | I think that's the reason I've kind of OCD'd on this semver issue recently | 18:06 |
| mordred | I think it's really helpful to have clear guidelines | 18:06 |
| dstufft | lol | 18:06 |
| dstufft | versioning | 18:06 |
| dstufft | HAVE FUN | 18:06 |
| mordred | and I think we've found around here that making those less flexible and more explicit is more scalable | 18:06 |
| mordred | so a lot of "if you want to do this, then do this" bits are basically fail | 18:06 |
| mordred | dstufft: oh, it's a constant stream of fun on that topic here! :) | 18:07 |
| dstufft | we're lucky in that all we really need to define is what order versions sort in | 18:07 |
| dstufft | though we tried in PEP440 to be prescriptive about semantic versions | 18:07 |
| mordred | dstufft: what I did last night was make a PEP440 compliant version of semver | 18:08 |
| dstufft | mordred: neat | 18:08 |
| dstufft | mordred: are you gonna make it a website, or it's own PEP? | 18:08 |
| dstufft | mordred: you shuold send it to Nic Coghlan | 18:08 |
| dstufft | Nick | 18:08 |
| mordred | hrm. interesting idea - I hadn't thought of making it a pep | 18:08 |
| mordred | I was just going to add it as docs to pbr for us - but that's a great idea | 18:08 |
| mordred | I might have to remove oneof the snarky paragraphs though | 18:09 |
| dstufft | mordred: I think it'd be a nice PEP, were PEP440 defines the format and how to order it, and PEPXXX defines a semantic versioning for it that people can opt in to using | 18:10 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Add the semver documentation https://review.openstack.org/48855 | 18:11 |
| mordred | dstufft: ++ | 18:11 |
| mordred | dstufft: I'll start working on that | 18:11 |
| dstufft | mordred: do you have Nick's email? He's sort of the BDFL of packaging and might have insight | 18:11 |
| mordred | yah. | 18:11 |
| dstufft | cool | 18:11 |
| *** dkliban has joined #openstack-infra | 18:14 | |
| mordred | dstufft: btw - there's the new text in the context of adding it to pbr docs ^^^ | 18:22 |
| mordred | dhellmann: ^^ (I'd love a review from you on that) | 18:22 |
| mordred | sdague: (same with you, although I have not added any text around your dependency thoughts yet) | 18:22 |
| clarkb | this weather | 18:26 |
| clarkb | jog0 are flights having problems? | 18:27 |
| clarkb | I shouldnt have to strategize a drive to bend in september... should be "exciting" | 18:27 |
| *** SergeyLukjanov has joined #openstack-infra | 18:27 | |
| mordred | clarkb: wow. what sort of weather are you having? | 18:32 |
| clarkb | typhoon leftovers. lots of rain and wind | 18:33 |
| clarkb | and snow but not low enough for the passes thankfully | 18:34 |
| *** thedodd has joined #openstack-infra | 18:35 | |
| *** thedodd has quit IRC | 18:35 | |
| *** thedodd has joined #openstack-infra | 18:35 | |
| *** thedodd has quit IRC | 18:46 | |
| *** jpeeler has quit IRC | 18:49 | |
| lifeless | https://bugs.launchpad.net/oslo/+bug/1232868 | 18:54 |
| uvirtbot | Launchpad bug 1232868 in oslo "fresh cookiecut environment barfs" [Undecided,New] | 18:54 |
| mordred | lifeless: thanks | 18:55 |
| lifeless | mordred: added more info | 18:57 |
| lifeless | mordred: looks like it needs some - -> _ folding logic | 18:57 |
| mordred | lifeless: that's not a cookiecutter bug per-se (I might be wrong) - you need to do an initial commit | 18:57 |
| mordred | and YES - ti needs - _ folding logic | 18:57 |
| mordred | I believe I'm going to have to add an additional input var | 18:57 |
| mordred | so "repo_name" and "module_name" | 18:58 |
| * mordred started hacking on that | 18:58 | |
| lifeless | mordred: so if the init/add/commit is needed to make it usable, that seems like a bug in /something/ | 18:58 |
| mordred | lifeless: nod | 18:58 |
| mordred | lifeless: let me see if there is a way for me to add operational hooks | 18:58 |
| mordred | so that I can potentially do the git init/add/commit steps | 18:58 |
| lifeless | note that for me there is also git config --local user.email and user.name steps | 18:58 |
| mordred | point well taken | 18:58 |
| mordred | hrm | 18:59 |
| lifeless | so just doing the init/add/commit won't work | 18:59 |
| mordred | I'm not 100% sure how to accomodate that - perhaps printing a message saying "now you need to git init/commit before this will work" ? | 18:59 |
| lifeless | mordred: that would remove somme wtfing | 18:59 |
| mordred | so that at least the expectation is clear | 18:59 |
| mordred | yah. ok. I can work on that | 19:00 |
| lifeless | mordred: note that the current message saying 'have to have a parent' is super misleading | 19:00 |
| lifeless | because I was like, how can I get this into gerrit so there is a parent, when I can't validate | 19:00 |
| mordred | lifeless: cookiecutter 0.7 (not yet released) supports the features we'll need for post-hooks | 19:01 |
| mordred | sigh. fail. | 19:02 |
| mordred | the main author is having to deal with life drama | 19:03 |
| mordred | which is why 0.7 isn't out yet | 19:03 |
| mordred | the life drama is perfectly fair | 19:03 |
| mordred | but definitely one of the reasons why projects with a single human gatekeeper are problematic | 19:03 |
| mordred | lifeless: for translation domains - do you know if it should be project name or module name? | 19:07 |
| mordred | like, python-swiftclient or swiftclient ? | 19:07 |
| lifeless | the former | 19:08 |
| lifeless | it's not related to python | 19:08 |
| lifeless | it's an arbitrary key into /usr/share/locale/<lang>/LC_MESSAGES/$domain.mo | 19:08 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/cookiecutter: Differentiate between repo and module name https://review.openstack.org/48857 | 19:09 |
| mordred | lifeless: let me know what you think of that ^^ | 19:09 |
| mordred | lifeless: I think it addresses the repo/module -/_ split, but I'm not sure if the default value/prompt text is clear enough | 19:10 |
| lifeless | Hard to tell without unaware guinea pigs | 19:11 |
| mordred | indeed | 19:11 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/cookiecutter: Add some clarifying text to README about git https://review.openstack.org/48858 | 19:12 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Add the semver documentation https://review.openstack.org/48855 | 19:16 |
| mordred | WOW | 19:17 |
| mordred | THE GATE IS SO BAD | 19:17 |
| mordred | lifeless: speaking of - did you see you had a real test failure on one of your pbr patches? | 19:19 |
| mordred | lifeless: thanks for those, btw | 19:19 |
| mordred | I want all of them to land | 19:19 |
| *** tjones has joined #openstack-infra | 19:28 | |
| *** jaypipes has quit IRC | 19:28 | |
| *** tjones has quit IRC | 19:32 | |
| lifeless | mordred: was one real? | 19:32 |
| mordred | lifeless: yes | 19:32 |
| mordred | I -1'd the code review on it | 19:32 |
| mordred | so it should be easy to spot | 19:32 |
| lifeless | thanks, shall look | 19:33 |
| lifeless | C was up all night with coughing/breathing issues | 19:33 |
| mordred | supporting python3 is hard | 19:33 |
| mordred | lifeless: ewww | 19:33 |
| lifeless | so today may be terrible | 19:33 |
| mordred | yah | 19:33 |
| lifeless | mum had her | 19:33 |
| lifeless | which means I have her now | 19:33 |
| dstufft | mordred: supporting python3 is easier if you just stop supporting python2 | 19:33 |
| dstufft | :D | 19:33 |
| mordred | dstufft: yes. yes it is | 19:35 |
| mordred | s/supporting python3 is hard/supporting python2 AND python3 is hard/ | 19:35 |
| lifeless | mordred: btw, putting nontrivial stuff in __init__ is a bit of an antipattern | 19:35 |
| mordred | lifeless: agree | 19:35 |
| mordred | lifeless: are we doing that in pbr? | 19:35 |
| lifeless | mordred: better to have $module.tests.support or $module.tests.base or something | 19:35 |
| mordred | ah. yes. | 19:35 |
| lifeless | mordred: cookiecutter and pbr | 19:35 |
| mordred | agree. I'll fix in both | 19:36 |
| notmyname | mordred: I can't seem to follow the conversation. something up with python-swiftclient? | 19:40 |
| mordred | notmyname: nope. sorry, was using you as a naming example in a question | 19:40 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Move base test case logic out of __init__.py https://review.openstack.org/48860 | 19:42 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/cookiecutter: Move base test class out of __init__.py https://review.openstack.org/48861 | 19:43 |
| *** ericw has joined #openstack-infra | 19:44 | |
| *** jaypipes has joined #openstack-infra | 19:45 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/cookiecutter: Remove vim modelines https://review.openstack.org/48862 | 19:45 |
| *** ericw has quit IRC | 19:48 | |
| *** thedodd has joined #openstack-infra | 19:49 | |
| *** dcramer_ has quit IRC | 19:52 | |
| *** ekarlso has quit IRC | 20:00 | |
| openstackgerrit | A change was merged to openstack-dev/cookiecutter: Differentiate between repo and module name https://review.openstack.org/48857 | 20:01 |
| lifeless | mordred: copy-pasting the stdout/stderr capturing and timeout glue everywhere seems poor. | 20:01 |
| mordred | lifeless: agree. thoughts? | 20:01 |
| lifeless | mordred: can I suggest a single fixture in oslo somewhere that encapsulates all of that | 20:01 |
| lifeless | e.g. oslo.testsupport.StandardFixtures() | 20:02 |
| lifeless | then the copy-paste is just | 20:02 |
| mordred | lifeless: yes. actually, I've been meaning to do more with the fixtures modules in oslo, break it out into its own library - and to move some thigns in to fixtures themselves | 20:02 |
| lifeless | self.useFixture(StandardFixtures()) | 20:02 |
| mordred | yah | 20:02 |
| mordred | I've actually been thinking even that an oslo BaseTestCase class might not be terrible - which would then use an oslo fixture thing | 20:04 |
| mordred | what do you think | 20:04 |
| mordred | ? | 20:04 |
| lifeless | base classes are bad | 20:04 |
| openstackgerrit | A change was merged to openstack-dev/cookiecutter: Add some clarifying text to README about git https://review.openstack.org/48858 | 20:04 |
| openstackgerrit | A change was merged to openstack-dev/cookiecutter: Move base test class out of __init__.py https://review.openstack.org/48861 | 20:04 |
| openstackgerrit | A change was merged to openstack-dev/cookiecutter: Remove vim modelines https://review.openstack.org/48862 | 20:04 |
| mordred | awesome | 20:05 |
| mordred | you know we have one in every project pretty much though | 20:05 |
| lifeless | http://rbtcollins.wordpress.com/2010/05/10/maintainable-pyunit-test-suites/ | 20:05 |
| lifeless | yes, and in most of them it's a code smell IMO | 20:05 |
| mordred | but where do you put self.useFixture(StandardFixtures()) ? | 20:06 |
| mordred | in every test case? and then assume that the 1200 devs out there are going to remember to do it? | 20:06 |
| lifeless | so that genuinely common to everyone policy is about the only time it makes sense :) | 20:07 |
| lifeless | but | 20:07 |
| mordred | that was all I was thinking of putting into the base test class :) | 20:07 |
| lifeless | the problem is that folk then want to add helpers | 20:07 |
| lifeless | and assertions to the base | 20:07 |
| mordred | ah. no. I'm not interested in that | 20:07 |
| lifeless | the difference between | 20:07 |
| lifeless | 'inherit from X' | 20:07 |
| lifeless | and 'def setUp(self):\nself.useFixture(StandardFixtures())' | 20:07 |
| lifeless | is 3 lines, one of which is blank | 20:08 |
| mordred | but you have to inherit from _something_ | 20:08 |
| mordred | and it's 3 lines per test class which can be forgotten | 20:08 |
| lifeless | sure | 20:08 |
| lifeless | which is why | 20:08 |
| lifeless | 09:07 < lifeless> so that genuinely common to everyone policy is about the only time it makes sense :) | 20:08 |
| mordred | nod | 20:08 |
| *** ekarlso has joined #openstack-infra | 20:09 | |
| lifeless | I think the cookiecutter having a base class is an ok compromise | 20:09 |
| lifeless | there's not enticement to follow the antipattern openstack-wide that way | 20:09 |
| *** fbo is now known as fbo_away | 20:10 | |
| dstufft | mordred: use py.test, no more base calsses :V | 20:10 |
| dstufft | autouse fixtures ftw | 20:11 |
| mordred | dstufft: too much magic, not enough explicit | 20:11 |
| mordred | I find py.test test suites very difficult to follow | 20:11 |
| dstufft | hard to follow? | 20:11 |
| dstufft | really? :V | 20:11 |
| mordred | yes. truly | 20:11 |
| mordred | the first time I tried to make a patch to tox, I had NO idea how the test suite worked | 20:12 |
| dstufft | what trips you up? I find them way easier because there's almost no boilerplate | 20:12 |
| lifeless | dstufft: py.test is way too clever | 20:12 |
| mordred | there's no boilerplate | 20:12 |
| mordred | dstufft: that's the problem | 20:12 |
| dstufft | .. | 20:12 |
| dstufft | you want boilerplate? | 20:12 |
| mordred | as a person who has not read _everything_ there is to know about py.test | 20:12 |
| mordred | it's black magic | 20:12 |
| mordred | it's not clear what's going on | 20:12 |
| mordred | and I can't folow from the code itself | 20:12 |
| lifeless | dstufft: even the magical assert interception squicks me | 20:12 |
| mordred | "where does the X in the arg list come from? what's calling it?" | 20:12 |
| dstufft | lifeless: oh man I hate self.assertFoo | 20:13 |
| dstufft | I won't use them | 20:13 |
| dstufft | I use bare asserts even in non py.test | 20:13 |
| mordred | and herein we reach impass | 20:13 |
| mordred | dstufft: please don't use base asserts in non py.test | 20:13 |
| mordred | it's fine in py.test - they're supported there | 20:13 |
| lifeless | dstufft: you'll want to to fit in the style with projects to contribute too | 20:13 |
| mordred | they're terrible in non py.test | 20:14 |
| lifeless | dstufft: or your reviewers may have words for you | 20:14 |
| *** ericw has joined #openstack-infra | 20:14 | |
| mordred | in any case - things with magical side effects are confusing to newcomers | 20:14 |
| dstufft | I got better things to do then try to remember which assertFoo I want in any particular case | 20:14 |
| lifeless | there's no breadcrumbs | 20:14 |
| mordred | and we have a TON of those around here. they have a hard enough time figuring out the explicit stuff :) | 20:15 |
| lifeless | dstufft: sure, I'm not trying to modify your personal preference; just saying that in different environments there is value to consistency | 20:15 |
| lifeless | dstufft: and btw, assertThat, the last assertion you'll ever need :P | 20:15 |
| mordred | py.test to me seems like powertools, which is great in a lot of scenarios | 20:16 |
| mordred | like "if I use this, I can get a ton more done with less typing, and I'm smart, so it's ok" | 20:16 |
| dstufft | lifeless: sure, and I just don't submit patches to places that make me have to remember which assertWhatever I need to use (or look them up), maybe that's rude of me IDK, I got more things on my plate then I can get reasoanbly get done so if it annoys me I don't do it :D | 20:17 |
| lifeless | this is like the LISP argument :P | 20:17 |
| lifeless | dstufft: you must have a headache contributing to openstack then :) | 20:17 |
| dstufft | lifeless: I don't think I have any openstack patches :) | 20:17 |
| dstufft | granted I probably would if I worked directly on openstack at work (and I don't) | 20:18 |
| dstufft | and i'd probably have to put up with assertFoo in that case, but that's got $$ | 20:18 |
| mordred | dstufft: well, most of the people at nebula hate openstack dev anyway - so you're probably safe :) | 20:18 |
| mordred | they used to yell at me a lot | 20:19 |
| *** ericw has quit IRC | 20:19 | |
| dstufft | heh | 20:20 |
| dstufft | I'm somewhat isolated from openstack directly, I spend the bulk of my time working on our pieces | 20:20 |
| mordred | well, I'm THRILLED with the contribution your work makes to OpenStack personally | 20:21 |
| dstufft | my "contribution" to openstack seems to be whining about packaging | 20:21 |
| dstufft | ;) | 20:21 |
| ekarlso | openstack dev, what's bad about that ? | 20:22 |
| mordred | dstufft: you might also run pypi and hack on pip and listen to us :) | 20:25 |
| mordred | dstufft: in any case, if you ever want an ATC exception to come to the summit for free, you just let us know | 20:25 |
| dstufft | mordred: :) | 20:25 |
| *** salv-orlando has joined #openstack-infra | 20:26 | |
| dstufft | mordred: gotta go for a bit | 20:26 |
| *** ekarlso has quit IRC | 20:28 | |
| *** ekarlso has joined #openstack-infra | 20:28 | |
| mordred | sdague: I have 9 patches up for pbr. several of them I've rechecked several times. Of that, in the last day, I have two passing test runs | 20:32 |
| mordred | none of the fails are real | 20:32 |
| *** ekarlso has quit IRC | 20:38 | |
| *** AlexF has joined #openstack-infra | 20:44 | |
| sdague | mordred: how many of them are elastic-recheck finding | 20:44 |
| sdague | ? | 20:44 |
| mordred | sdague: at least half | 20:44 |
| mordred | sdague: but it's the same ones over and over again | 20:44 |
| mordred | sdague: it's amazing how painful it is | 20:45 |
| sdague | yep | 20:45 |
| sdague | we've got a good dozen races in there | 20:45 |
| *** ericw has joined #openstack-infra | 20:45 | |
| sdague | yeh, I've got a fix to something that I need to see the neutron fail to be useful. I'm on my 5th recheck, 3 passes, 2 unrelated fails | 20:49 |
| sdague | I actually think the idle time test swarm would provide some very interesting data about how bad things are, especially with elastic-recheck to help us semi-automatically classify them | 20:49 |
| *** ericw has quit IRC | 20:50 | |
| *** AlexF has quit IRC | 21:14 | |
| *** ericw has joined #openstack-infra | 21:16 | |
| *** ericw has quit IRC | 21:21 | |
| *** dims has joined #openstack-infra | 21:27 | |
| *** ekarlso has joined #openstack-infra | 21:28 | |
| openstackgerrit | A change was merged to openstack-infra/devstack-gate: Add pypi-mirror to the list of pre-cloned repos https://review.openstack.org/48805 | 21:40 |
| *** tjones has joined #openstack-infra | 21:44 | |
| *** ianw has joined #openstack-infra | 21:46 | |
| mordred | sdague: ++ | 21:49 |
| mordred | O M G | 21:50 |
| mordred | the d-g patch passed! | 21:50 |
| *** Ryan_Lane has joined #openstack-infra | 22:02 | |
| *** tjones has quit IRC | 22:03 | |
| *** dkranz has quit IRC | 22:04 | |
| jog0 | clarkb: just arrvied in Bend | 22:08 |
| *** SergeyLukjanov has quit IRC | 22:12 | |
| *** ekarlso has quit IRC | 22:14 | |
| *** ekarlso has joined #openstack-infra | 22:14 | |
| *** Ryan_Lane has quit IRC | 22:21 | |
| *** Ryan_Lane has joined #openstack-infra | 22:23 | |
| *** Ryan_Lane has quit IRC | 22:24 | |
| mordred | jog0: congrats | 22:27 |
| clarkb | jog0 me too. just checked in. thinking of getting food at $brewpub in a bit | 22:27 |
| mordred | clarkb: it's guam where you lived, yeah? | 22:28 |
| jog0 | clarkb: what hotel are you at? | 22:28 |
| jog0 | count me in for food | 22:28 |
| mordred | clarkb: or am I spacecake? | 22:28 |
| clarkb | pheonix inn on franklin | 22:28 |
| clarkb | mordred pohnpei, near guam | 22:29 |
| jog0 | mordred: the two are orthogonal | 22:29 |
| jog0 | clarkb: same, I am on the second floor | 22:29 |
| mordred | jog0: ++ | 22:29 |
| BobBall | Random question that I assume ppl in here will know... Do I need to run tempest tests in a special way to get them executing in parallel? | 22:30 |
| jog0 | clarkb: let me know when you are ready and lets meet in the lobby | 22:30 |
| *** ArxCruz has joined #openstack-infra | 22:30 | |
| clarkb | will do | 22:31 |
| clarkb | BobBall we use testr --parallel to run them | 22:32 |
| BobBall | looks like I need to upgrade testr then :) | 22:34 |
| BobBall | thanks | 22:35 |
| *** mrda has joined #openstack-infra | 22:37 | |
| mordred | BobBall: I _think_ the tox envs in the tempest tree will execute parallel or non-parallel as appropriate | 22:38 |
| BobBall | hmmmm - ok... Maybe it's a bug with the way I'm setting things up then | 22:39 |
| BobBall | I'm pretty sure they aren't running in parallel | 22:39 |
| BobBall | I'll dig | 22:39 |
| BobBall | thanks | 22:39 |
| clarkb | jog0: meet in the lobby at 4? | 22:48 |
| jog0 | clarkb: sounds good | 22:49 |
| *** slong has joined #openstack-infra | 22:50 | |
| mordred | jog0, clarkb: if you want to review pbr patches, I will approve any beer you attempt to expense... | 22:51 |
| *** mrda has quit IRC | 22:53 | |
| *** boris-42 has quit IRC | 22:57 | |
| *** cody-somerville has joined #openstack-infra | 22:59 | |
| *** vipul is now known as vipul-away | 23:02 | |
| mordred | lifeless: did you ever make any headway on ValueError: I/O operation on closed file | 23:13 |
| mordred | in pip? | 23:13 |
| mordred | lifeless: http://logs.openstack.org/03/48803/1/check/check-pbr-devstack-vm-rawinstall/da93fc2/console.html | 23:13 |
| mordred | lifeless: I'm hitting it in my patch to try to get wheels to work | 23:13 |
| *** thedodd has quit IRC | 23:18 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Use wheels for installation https://review.openstack.org/48803 | 23:19 |
| lifeless | mordred: yes, it was a cascade error | 23:22 |
| lifeless | mordred: there was an earlier error that was reported, and once fixed it was all good | 23:22 |
| lifeless | mordred: the earlier error was due to the use of pypi.o.o rather than my local mirror | 23:23 |
| dstufft | mordred: ping | 23:23 |
| dstufft | mordred: how hard would it be to run the openstack tests (just once is alll I really need) against a different index server then pypi.o.o | 23:24 |
| mordred | lifeless: yup. same here. thanks - found the real error (I think) | 23:24 |
| mordred | dstufft: hrm. which ones, the "build the mirror" tests, or the "run tempest" ones? | 23:25 |
| mordred | dstufft: (both I think I can trigger easily, but want to know which thing you want) | 23:25 |
| dstufft | mordred: whichever would stress the index server the most | 23:25 |
| mordred | ah. that would be the build-the-mirror test most likely | 23:25 |
| dstufft | mordred: I don't know if you know or not, I'm writing a PyPI 2.0, and sometime in the near future i'm going to have a preview.pypi.python.org/simple/ with a whole new codebase that I'm going to try and get people to install against it some to make sure I didn't miss anything | 23:26 |
| mordred | dstufft: yes. I'm very excited about that | 23:26 |
| mordred | dstufft: I could probably even just set up a non-voting job to hammer it constantly and give you a nice stream of feedback :) | 23:26 |
| dstufft | figure the openstack tests would be a decent data point | 23:27 |
| mordred | we are evil | 23:27 |
| dstufft | :D | 23:27 |
| dstufft | it won't be behind a CDN or anything for now, and I need to get a *.pypi.python.org SSL certificate still | 23:27 |
| dstufft | but I almost have the /simple/* api finished | 23:27 |
| mordred | neat! | 23:27 |
| mordred | dstufft: is this replacement something that will be actually runnable elsewhere too? | 23:27 |
| dstufft | It's working locally, I need to write tests and documentation and finish a few things like Cache-Control headers | 23:27 |
| dstufft | mordred: ideally yes | 23:27 |
| dstufft | that's the goal anyways | 23:28 |
| mordred | dstufft: neat. I mean, tbh, I still am not sure we really need to do that, since the real purpose of our mirror is just to protect against transient network barf and static files in an apache does fine for that | 23:28 |
| * mordred wants to make sure we don't go _too_ far in running an alternate universe over here | 23:28 | |
| dstufft | right now the only pypi.p.o centric stuff is Fastly integration and that's conditioned on a config change | 23:29 |
| dstufft | s/change/variable/ | 23:29 |
| mordred | ossum | 23:29 |
| dstufft | (Fastly specific because varnish doesn't have surrogate-keys) | 23:29 |
| * mordred should go poke devpi again and see if it works yet | 23:29 | |
| dstufft | Part of my work with Warehouse (PyPI 2.0) is actually documenting the fucking API | 23:30 |
| dstufft | because documentation is a cool thing to have | 23:30 |
| mordred | yup | 23:30 |
| mordred | also, I'm excited | 23:30 |
| mordred | because once warehouse is there | 23:30 |
| mordred | you can start adding NEW api | 23:30 |
| dstufft | yes :D | 23:30 |
| mordred | like "o hai, I'm a person, and I want to upload this thing" | 23:30 |
| dstufft | the upload API is conceivably a thing that could happen sooner rather than later | 23:31 |
| dstufft | since it's distinct from the download API | 23:31 |
| dstufft | and isn't as big of a problem | 23:31 |
| dstufft | I'm hopefully going to be able to get a PEP to depercate/break ``python setup.py upload`` completely | 23:31 |
| dstufft | and make twine the official uploader | 23:31 |
| mordred | ooh. nice | 23:32 |
| mordred | you know, people are going to be angered | 23:32 |
| dstufft | Mostly because sending credentials in plaintext is bad | 23:32 |
| mordred | because of $crazy | 23:32 |
| * mordred eats dinner now | 23:32 | |
| dstufft | of course they are, I'm getting good at pissing people off tho :D | 23:32 |
| dstufft | mordred: have fun | 23:32 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!