| *** jamesmcarthur has joined #openstack-tc | 00:01 | |
| *** lbragstad has quit IRC | 00:48 | |
| openstackgerrit | Merged openstack/governance master: Remove Train goals placeholder https://review.opendev.org/664276 | 00:52 |
|---|---|---|
| *** jamesmcarthur has quit IRC | 00:52 | |
| *** ricolin has joined #openstack-tc | 01:10 | |
| *** jamesmcarthur has joined #openstack-tc | 01:35 | |
| *** jamesmcarthur has quit IRC | 01:37 | |
| *** diablo_rojo has joined #openstack-tc | 02:03 | |
| *** whoami-rajat has joined #openstack-tc | 02:09 | |
| *** diablo_rojo has quit IRC | 03:14 | |
| *** e0ne has joined #openstack-tc | 05:15 | |
| *** lpetrut has joined #openstack-tc | 05:28 | |
| *** e0ne has quit IRC | 05:42 | |
| *** lpetrut has quit IRC | 05:43 | |
| *** Luzi has joined #openstack-tc | 05:53 | |
| *** lpetrut has joined #openstack-tc | 06:46 | |
| *** tosky has joined #openstack-tc | 07:17 | |
| *** e0ne has joined #openstack-tc | 07:20 | |
| *** bsilverman has quit IRC | 07:32 | |
| *** jpich has joined #openstack-tc | 08:02 | |
| *** dtantsur|afk is now known as dtantsur | 08:02 | |
| evrardjp | o/ | 08:09 |
| asettle | o/ | 08:19 |
| ricolin | o/ | 08:30 |
| ttx | office hour! | 09:01 |
| ttx | live from Open Infra Days in Krakow, Poland | 09:02 |
| evrardjp | woot | 09:02 |
| ttx | Lots of SUSE green here | 09:02 |
| evrardjp | That's good to hear I guess :) | 09:03 |
| asettle | Yeah I saw the team over there | 09:12 |
| asettle | on LinkedIn | 09:12 |
| ricolin | nice | 09:20 |
| *** tobberydberg has quit IRC | 09:22 | |
| ttx | evrardjp: I learned that SUSE green hat swag is not very good in China | 09:31 |
| *** tobberydberg has joined #openstack-tc | 09:31 | |
| asettle | Why/ | 09:39 |
| asettle | ?* | 09:39 |
| ttx | asettle: hmm well... | 09:44 |
| ttx | "wearing a green hat" means your wife is cheating on you. | 09:46 |
| ttx | So apparently they are not very popular as booth swag | 09:46 |
| evrardjp | TIL | 09:46 |
| ttx | It's all over the Internets if you need more details :) | 09:47 |
| asettle | HAHA OH | 09:47 |
| asettle | That's funny | 09:47 |
| asettle | Someone should tell the marketing department | 09:47 |
| evrardjp | giving shoes or WATCHES is a no-no | 09:48 |
| ttx | Would be even funnier if the color was red, given some corporate dress code | 09:48 |
| evrardjp | whaaaaat | 09:48 |
| evrardjp | that sounds very sad | 09:48 |
| evrardjp | ttx: hahaha | 09:48 |
| ttx | asettle: I learned it from a SUSE guy... lesson apparently hard learned | 09:49 |
| evrardjp | ttx: Imagine if a red "O" also had a bad meaning. | 09:49 |
| asettle | Lol ttx oh dear | 09:49 |
| evrardjp | we want names now ttx | 09:49 |
| ttx | I'm not sure I'm allowed to disclose names. that was told in cofidence in akeynote room with 200 people | 09:50 |
| evrardjp | hahah | 09:50 |
| ttx | I'll just give enough hints | 09:50 |
| asettle | Ah yes, that hallow sanctity | 09:50 |
| ttx | I understood the mention, so the talk was in English, and it was given before now. | 09:51 |
| evrardjp | too many details :) | 09:51 |
| ttx | I know you're lazy | 09:51 |
| evrardjp | hahah | 09:51 |
| ttx | I said I would not give the name. I didn't say I don't want you to know | 09:51 |
| evrardjp | oh I got it when I asked for open infra days poland agenda. | 09:51 |
| evrardjp | hahaha | 09:52 |
| ttx | More seriously, I think we need some kind of briefing on major gaffes | 09:52 |
| evrardjp | ttx: that sounds a good idea. From the foundation I guess? | 09:54 |
| evrardjp | ttx: Horace could probably help? | 09:54 |
| ttx | yeah, I'm making a sidenote for that | 09:55 |
| evrardjp | thanks | 09:59 |
| asettle | tc-members - with the PDF goal, did we decide on services, clients, and libraries to be built into PDFs? Or just clients and services? | 10:00 |
| mugsie | I don't think we decided, did we? | 10:18 |
| evrardjp | I can't remember that but if it's a community goal, it makes sense to apply everywhere. | 10:21 |
| evrardjp | Else it's like the move to python-openstackclient ... ahem ahem. | 10:22 |
| *** tdasilva_ has quit IRC | 10:25 | |
| asettle | Lol evrardjp | 10:26 |
| asettle | Yeah, you voted on it being a community goal :p | 10:26 |
| evrardjp | I did, with the hope it would apply everywhere. I didn't think of trimming down the goal. Or maybe I didn't understand your question here fully ? | 10:32 |
| evrardjp | (most likely) | 10:32 |
| *** stephenfin has joined #openstack-tc | 10:42 | |
| asettle | tc-members - I've got stephenfin in here regarding the PDF goal. We need to clarify if we're generating PDF builds for services, clients, and libarires | 10:43 |
| asettle | I have a faint memory it was only for the services | 10:43 |
| asettle | But the community goal I wrote was unclear, sadly | 10:44 |
| asettle | Understandably evrardjp thinks it should go across all three | 10:44 |
| asettle | But as stephenfin noted, that's a lot of data and a lot more things can go wrong | 10:44 |
| asettle | We need to have a clear understanding soon, as it's important we kick start this goal as we're already going through the release process. | 10:45 |
| evrardjp | not sure what you mean by through the release process? | 10:48 |
| evrardjp | the goal says across the project repositories, so I expected all of them for consistency. | 10:50 |
| evrardjp | but we can track this in storyboard I guess? | 10:51 |
| evrardjp | (like start with the services if you prefer?) | 10:51 |
| asettle | evrardjp, we're already working on Train | 10:52 |
| evrardjp | it's not against you or the goal itself, is that we used a definition to reject some goals, it doesn't make sense to come back over the definition. | 10:52 |
| evrardjp | at least for me | 10:52 |
| asettle | evrardjp, that's fine - but the definition (to me and stephenfin at least) it's not particularly clear | 10:52 |
| asettle | And that's on me | 10:52 |
| evrardjp | this is why it's a train goal :) | 10:52 |
| evrardjp | not sure, I seem confused | 10:52 |
| evrardjp | what's the problem? | 10:53 |
| asettle | stephenfin, ?? | 10:53 |
| evrardjp | you are afraid it's not possible to finish on time? | 10:53 |
| asettle | I'll wait for stephenfin to clarify, as it is not me that is working on the initial steps and questioned how far this is reaching | 10:53 |
| stephenfin | The issue is that building PDFs is time-consuming and LaTeX is flaky | 10:54 |
| stephenfin | and I'm wondering whether there is value in building PDFs for something like oslo.db | 10:55 |
| stephenfin | Compared to building the nova or novaclient docs | 10:55 |
| jroll | it doesn't hurt to build PDFs for libraries, but I think focusing on services and clients first is best | 10:56 |
| stephenfin | Hmm, well it looks like we might be going to do it for all because that's actually the easiest thing to implement. See https://review.opendev.org/664555 | 10:59 |
| stephenfin | Apologies in advance. I don't work with Ansible that often | 10:59 |
| evrardjp | I agree with the pragmatic approach, as said above, and as jroll highlights :) | 11:00 |
| evrardjp | stephenfin: haha no worries. I think it's also easier to have a global approach for the tooling, but then I guess we need to ensure things are right in the projects themselves? | 11:00 |
| asettle | stephenfin, I'll add some Ansible folks to your patch | 11:00 |
| stephenfin | cool, ta | 11:03 |
| * stephenfin -> 🏊 | 11:03 | |
| jroll | stephenfin: ++, cool | 11:07 |
| mugsie | I think that having the PDF support for libraries could be useful for the disconnected / limited web access use cases (e.g. China) | 11:39 |
| mugsie | which (from what I can remember) was one of the driving forces for the goal? | 11:40 |
| asettle | mugsie, other than my own selfishness, ABSOLUTELY | 11:46 |
| *** ricolin_ has joined #openstack-tc | 12:27 | |
| *** ricolin has quit IRC | 12:29 | |
| *** ijolliffe has joined #openstack-tc | 12:31 | |
| fungi | i agree that if there's really not much work necessary to be able to generate pdfs from sphinx, then we should just add it to the job which generates sphinx docs and have it make pdfs at the same time | 12:35 |
| fungi | plus whatever (hopefully minimal) plumbing projects need in their sphinx configs? | 12:36 |
| evrardjp | fungi: so, just changing the current jobs to do both for train and above | 12:39 |
| evrardjp | is that what you mean? | 12:39 |
| fungi | if that's what's involved? i haven't really looked | 12:41 |
| fungi | i mean, presumably there's at least some work necessary in each repository to support pdf generation during documentation builds, or else it wouldn't have been a community goal | 12:42 |
| fungi | from this it looks like some work is needed within the documentation source files to make the result more pdf-friendly: https://specs.openstack.org/openstack/docs-specs/specs/ocata/build-pdf-from-rst-guides.html#work-items | 12:45 |
| fungi | oh, wait, that's requirements for the theme itself | 12:46 |
| fungi | so maybe it is just adding related deps (latex?) to doc/requirements.txt and a directive or several in doc/source/conf.py and then adjusting the definitions of the shared doc build jobs in openstack/openstack-zuul-jobs repo? | 12:47 |
| fungi | oh, right, latex deps wouldn't go in doc/requirements.txt because they're not python. i guess bindep.txt with test and docs profiles added | 12:49 |
| evrardjp | fungi: I think stephenfin has written a first patch for changing openstack-zuul-jobs to add dependencies and to run the pdf generation command | 12:50 |
| fungi | cool | 12:50 |
| evrardjp | fungi: yes I would say that bindep would be a common location, but bindep are for the project themselves though, so it would mean a "test" profile -- or optionally, include this in the job directly | 12:51 |
| fungi | though if the deps are only encoded in the job then that doesn't help folks who are wanting to build pdf docs locally... *cough* suse *cough* | 12:51 |
| *** jamesmcarthur has joined #openstack-tc | 12:51 | |
| evrardjp | but I haven't digged into that yet either, just quickly browsed | 12:51 |
| evrardjp | fungi: indeed :) | 12:51 |
| evrardjp | but isn't that already the case with translations though ? | 12:51 |
| evrardjp | I mean if the boat has sailed... :p | 12:52 |
| fungi | oh, is it? | 12:52 |
| evrardjp | I guess I will have homework to double check? | 12:52 |
| evrardjp | :p | 12:52 |
| fungi | you can't build non-default languages without installing things not specified in the repo? | 12:52 |
| evrardjp | I think I might be confused, and things might be just fetching from outside (so not self contained, but buildable) | 12:53 |
| evrardjp | I have to double check | 12:53 |
| evrardjp | please ignore my comments in the meantime :) | 12:53 |
| fungi | we ought to eventually encourage projects to get extended documentation build dependencies into doc/requirements.txt (for python deps) and their bindep.txt with a doc profile (for system packaged deps) | 12:53 |
| evrardjp | yup that sounds a good plan | 12:54 |
| evrardjp | as long as the bindep is having the right profile, I think it's the right path forward | 12:54 |
| evrardjp | fungi: I was wrong, that ship hasn't sailed. It's well contained. What I thought was not contained is shipped into a python dep, which should be in doc/requirements anyway, so we are good with that approach | 12:58 |
| *** jamesmcarthur has quit IRC | 13:02 | |
| *** lbragstad has joined #openstack-tc | 13:03 | |
| fungi | so presumably the idea is that if you checkout the stable/train branch and `bindep docs` tells you that you've got the right packages installed then you should be able to `tox -e docs` and get not only html but also pdfs | 13:04 |
| fungi | er, well, more commonly you'd probably run bindep as `tox -e bindep docs` so you get bindep from pypi | 13:04 |
| evrardjp | I am not that fond of ^ | 13:11 |
| evrardjp | because bindep doesn't actively install packages for all the distros, it just list the missing, right? | 13:12 |
| stephenfin | fungi: Are you referring to what I was doing? | 13:12 |
| evrardjp | so it means that installing the bindep python through tox is fine, but installing the bindep missing distro packages sounds kinda weird | 13:12 |
| evrardjp | (through tox) | 13:12 |
| stephenfin | Because if so, I've purposefully avoided messing with individual projects' tox and/or bindep files | 13:13 |
| evrardjp | stephenfin: but then you won't be contained, and won't be able to build locally your pdfs | 13:13 |
| evrardjp | well | 13:13 |
| evrardjp | you can, but you need to know the receipe :) | 13:13 |
| evrardjp | it's not as easy as tox -e docspdf or tox -e docs | 13:14 |
| stephenfin | Yeah, I'm happy to package the script up somewhere external (like openstack-doc-tools or openstackdocstheme) | 13:14 |
| stephenfin | But touching every single project is an impossible task and, despite appearances, I'm no masochist :) | 13:14 |
| evrardjp | stephenfin: yes I see that the translation is already a bin in openstackdocstheme, so it could make sense to ship it there | 13:14 |
| evrardjp | stephenfin: oh really? | 13:14 |
| evrardjp | :p | 13:14 |
| evrardjp | stephenfin: I think it's fine, for a community goal, as you're not alone | 13:15 |
| evrardjp | shouldn't this be spread out around the community? (but I agree it's still a massive amount of work) | 13:15 |
| *** mriedem has joined #openstack-tc | 13:16 | |
| *** tdasilva has joined #openstack-tc | 13:17 | |
| fungi | evrardjp: bindep doesn't install anything, it just tells you what system packages you're missing for a particular profile defined in the local bindep.txt file. an alternative is to just look in or grep the bindep.txt file to get a list of packages and then check whether you have them installed | 13:18 |
| evrardjp | fungi: yes exactly | 13:18 |
| evrardjp | fungi: so it means you would have to zypper install / dnf install/ apt install | 13:19 |
| fungi | stephenfin: i don't get why it's a community goal if it doesn't require changes to projects | 13:19 |
| evrardjp | or whatever, which is not known in advance | 13:19 |
| evrardjp | so it means those scripts would have to know what's the cli required to install packages on $currentrunningsystem | 13:19 |
| fungi | evrardjp: right, bindep is meant as a resource for folks manually running things on their local system, though it does have the ability to generate a bare package list suitable for feeding into most package managers | 13:20 |
| stephenfin | fungi: tbh, I'm not TC and don't really know what a community goal entails. This was an idea dhellmann suggested and it made sense to me | 13:20 |
| evrardjp | fungi: that's what I meant :) | 13:20 |
| stephenfin | I'd personally rather have a 'pdfdocs' tox target in each project but I also have no interest in convincing projects that I'm not core on to do it :) | 13:20 |
| evrardjp | stephenfin: but that's the thing | 13:20 |
| evrardjp | you don't need to convince anymore | 13:21 |
| evrardjp | :p | 13:21 |
| fungi | why not just build pdfs when building the html? | 13:21 |
| fungi | does it take significantly more time/resources? | 13:21 |
| evrardjp | if <binary dependency> is available, build pdf? | 13:21 |
| evrardjp | where I suppose binary dependency would be latex | 13:21 |
| stephenfin | It does take some time, yes, and the dependencies are not insignificant | 13:21 |
| stephenfin | I counted at least a gig of deps when trying to build in a VM | 13:22 |
| stephenfin | also, it's not even possible in some environments | 13:22 |
| fungi | oh, fun. is that a gig more deps than for building html docs, or are a lot of them shared dependencies? | 13:22 |
| stephenfin | Fedora 30 is missing the xindy package. I've been meaning to get around to filing a bug but haven't done so | 13:22 |
| stephenfin | Yeah, you've to pull in texlive which is huge | 13:22 |
| fungi | but yeah, if it's a full texlive environment it will be massive, agreed | 13:23 |
| fungi | i saw in the openstack-manuals change it also required inkscape? | 13:23 |
| stephenfin | It did? | 13:23 |
| fungi | i guess for some svg postprocessing? | 13:23 |
| evrardjp | stephenfin: my point was that if we have a standard shell script we ship into openstackdocstheme, we could use it to build the pdf too, if the binary for latex is available. If not available, and pdf is requested, then exit with an error? | 13:23 |
| evrardjp | so that thing would then be "make sure you have something in your tox to use that one" ? | 13:24 |
| fungi | stephenfin: https://review.opendev.org/#/c/396943/9/bindep.txt | 13:24 |
| stephenfin | Oh, indeed. Must be for SVGs, aye | 13:24 |
| *** jamesmcarthur has joined #openstack-tc | 13:26 | |
| *** tdasilva has quit IRC | 13:54 | |
| *** tdasilva has joined #openstack-tc | 13:55 | |
| *** zaneb has quit IRC | 14:16 | |
| *** zaneb has joined #openstack-tc | 14:16 | |
| *** diablo_rojo has joined #openstack-tc | 15:02 | |
| *** lpetrut has quit IRC | 15:12 | |
| dhellmann | fungi , evrardjp : the work stephenfin is doing will set up the build and ignore failures, so that project teams can then ensure they *can* build PDFs and that those PDFs look good and are organized well. *that* work is the community part of the goal | 15:16 |
| evrardjp | I see. | 15:17 |
| evrardjp | it's about the ability, not the fact to do it everywhere | 15:17 |
| dhellmann | and we don't want this in the default "docs" env in tox because that requires every developer to install texlive | 15:17 |
| dhellmann | some teams (nova) already know their docs do not build as pdfs successfully | 15:17 |
| evrardjp | dhellmann: which is why I proposed to make that optional based on the presence of the latex toolkit | 15:17 |
| *** Luzi has quit IRC | 15:18 | |
| dhellmann | the other reason not to edit the tox.ini is that is a pain in the ass and no way to get work done | 15:18 |
| evrardjp | but I guess it's my understanding of the goal which was incorrect. I need to re-read that goal | 15:18 |
| dhellmann | there are hundreds of them | 15:18 |
| evrardjp | dhellmann: yeah. that's indeed a pain. | 15:18 |
| fungi | makes sense | 15:20 |
| fungi | so doable as a separate tox testenv and maybe a dedicated bindep profile | 15:21 |
| fungi | so we could verify dependencies are installed by checking `tox -e bindep pdf` and then build the pdfs with `tox -e pdf` (or something along those lines) | 15:22 |
| fungi | the goal was missing any real implementation plan, hence the questions | 15:22 |
| *** tdasilva has quit IRC | 15:31 | |
| *** tdasilva has joined #openstack-tc | 15:31 | |
| *** e0ne has quit IRC | 15:36 | |
| dhellmann | yeah, we should update the goal document with those details | 15:45 |
| dhellmann | it was not intended to be a requirement that teams update their repos to allow local building of PDFs, so we should make sure to make that optional | 15:45 |
| dhellmann | asettle , stephenfin : dunno which of you wants to do ^ | 15:46 |
| stephenfin | asettle? :) | 15:46 |
| * stephenfin is digging through tokenizer issues | 15:47 | |
| asettle | Hello yes tis I | 15:47 |
| asettle | oh yeah probably me then | 15:47 |
| asettle | It's on the list | 15:47 |
| fungi | so just to reiterate what dhellmann has said above: the community part of the goal is that project teams ensure they *can* build PDFs and that those PDFs look good and are organized well, but it's not a requirement that teams update their repos to allow local building of PDFs | 15:56 |
| fungi | so basically the job should not depend on in-repository declaration of new dependencies or a new tox env (i suppose that would be future additions to the relevant section of our pti, maybe as a distant future cycle goal), but the job should still run and produce pdfs for all projects covered by the goal | 15:58 |
| *** tdasilva_ has joined #openstack-tc | 15:58 | |
| *** tdasilva has quit IRC | 16:00 | |
| evrardjp | I am not really sure to understand the success criteria | 16:02 |
| evrardjp | I guess I will see | 16:02 |
| clarkb | not to further complicate things but pandoc can make pdfs and is only 26MB large | 16:05 |
| clarkb | (though that is compressed tarball form) | 16:05 |
| clarkb | if we are really concerned about the size of the dep chain maybe we should look at pandoc? | 16:05 |
| clarkb | (granted the reason tex is huge is it is full of "make things pretty" tooling) | 16:07 |
| fungi | curious if pandoc was investigated as an option and, if so, what ruled it out | 16:12 |
| *** tdasilva has joined #openstack-tc | 16:18 | |
| *** tdasilva_ has quit IRC | 16:20 | |
| *** jamesmcarthur has quit IRC | 16:32 | |
| *** jpich has quit IRC | 16:33 | |
| *** whoami-rajat has quit IRC | 16:38 | |
| *** dtantsur is now known as dtantsur|afk | 16:51 | |
| *** whoami-rajat has joined #openstack-tc | 17:05 | |
| *** jamesmcarthur has joined #openstack-tc | 17:10 | |
| *** dtruong has quit IRC | 17:24 | |
| *** jamesmcarthur has quit IRC | 17:26 | |
| *** jamesmcarthur has joined #openstack-tc | 17:26 | |
| *** tosky has quit IRC | 17:30 | |
| *** ijolliffe has quit IRC | 17:42 | |
| *** dtruong has joined #openstack-tc | 17:50 | |
| fungi | tc-members: i've replaced the health tracker page, feel free to edit the initial prose. if you need access to any of the old data it can be had in the "history" accessed from the "page" drop-down up in the top navbar | 18:07 |
| fungi | https://wiki.openstack.org/wiki/OpenStack_health_tracker | 18:07 |
| *** ricolin_ has quit IRC | 18:38 | |
| *** jamesmcarthur has quit IRC | 18:40 | |
| *** tdasilva has quit IRC | 18:45 | |
| *** dtruong has quit IRC | 18:58 | |
| *** dtruong has joined #openstack-tc | 18:59 | |
| *** jamesmcarthur has joined #openstack-tc | 18:59 | |
| dhellmann | fungi , clarkb : the docs theme includes styling support for sphinx's pdf generation, so I don't think we really looked past the fact that it's possible to do with sphinx | 19:01 |
| *** jamesmcarthur has quit IRC | 19:01 | |
| *** jamesmcarthur has joined #openstack-tc | 19:03 | |
| fungi | got it, but for sphinx specifically you need a full texlive environment | 19:04 |
| fungi | (for sphinx's pdf generation feature) | 19:04 |
| *** jamesmcarthur has quit IRC | 19:05 | |
| *** jamesmcarthur has joined #openstack-tc | 19:07 | |
| dhellmann | yes, right, sphinx generates latex and latex generates the pdf | 19:14 |
| *** jamesmcarthur has quit IRC | 19:15 | |
| *** diablo_rojo has quit IRC | 19:34 | |
| *** jamesmcarthur has joined #openstack-tc | 19:34 | |
| *** jamesmcarthur has quit IRC | 20:03 | |
| *** Sundar has joined #openstack-tc | 20:04 | |
| *** tdasilva has joined #openstack-tc | 20:08 | |
| *** jamesmcarthur has joined #openstack-tc | 20:25 | |
| *** jamesmcarthur has quit IRC | 20:28 | |
| *** jamesmcarthur has joined #openstack-tc | 20:53 | |
| *** jamesmcarthur has quit IRC | 21:04 | |
| *** diablo_rojo has joined #openstack-tc | 21:40 | |
| *** diablo_rojo has quit IRC | 21:49 | |
| *** diablo_rojo has joined #openstack-tc | 22:05 | |
| *** whoami-rajat has quit IRC | 22:55 | |
| *** mriedem has quit IRC | 23:06 | |
| *** lbragstad has quit IRC | 23:41 | |
| *** ianychoi has quit IRC | 23:48 | |
| *** ianychoi has joined #openstack-tc | 23:48 | |
| *** diablo_rojo has quit IRC | 23:53 | |
| *** tdasilva_ has joined #openstack-tc | 23:54 | |
| *** tdasilva has quit IRC | 23:57 | |
| *** jamesmcarthur has joined #openstack-tc | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!