14:06:04 <randomuser> #startmeeting Docs Project Meeting - Agenda: https://fedoraproject.org/wiki/Docs_Project_meetings
14:06:05 <zodbot> Meeting started Mon Nov 21 14:06:04 2016 UTC.  The chair is randomuser. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:06:05 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:06:05 <zodbot> The meeting name has been set to 'docs_project_meeting_-_agenda:_https://fedoraproject.org/wiki/docs_project_meetings'
14:06:05 <randomuser> #meetingname Fedora Docs
14:06:05 <randomuser> #topic Roll Call
14:06:05 <zodbot> The meeting name has been set to 'fedora_docs'
14:06:09 <bexelbie> .hello bex
14:06:10 <zodbot> bexelbie: bex 'Brian (bex) Exelbierd' <bex@pobox.com>
14:07:23 <randomuser> hello, bex
14:08:30 <bexelbie> hola
14:08:44 <bexelbie> I am finally done with most of my expenses and only have a few tickets left to close before I can focus again :)
14:08:47 <bexelbie> how are you?
14:09:33 <randomuser> super busy this time of year, Americans love their sales
14:10:09 <bexelbie> We even have several Black Friday sales in CZ/Europe
14:10:13 <bexelbie> No one really knows why :P
14:10:34 <randomuser> ha
14:10:36 <bexelbie> I also had to let someone know that CA Thanksgiving is at a different time from US Thanksgiving
14:11:20 <randomuser> yep, two entirely different holidays
14:11:34 <bexelbie> One thankful for a landing, one thankful for a miss :P
14:11:35 <bexelbie> j/k
14:12:10 <bexelbie> I will actually be traveling (personal) during US thanksgiving which is kind of messing with people's heads
14:12:43 <randomuser> that is atypical
14:13:14 <randomuser> well, since we're here
14:13:25 <randomuser> I mostly want to cover release prep tasks
14:13:38 <randomuser> #topic Where are with the Release Cycle and Docs?
14:14:05 <bexelbie> AIUI, November 22 is a go.  Do we have a publishing plan?
14:14:17 <randomuser> I know a few guides are prepped, so I'll publish them this evening
14:14:53 <randomuser> the release notes has one open PR, and it'll get published too
14:15:16 <randomuser> #info any guides that are branched for F25 will get published today
14:15:31 <randomuser> there's also some redirects to update
14:15:52 <randomuser> #info randomuser will update redirects to point to latest content
14:16:21 <randomuser> ... that's basically the whole plan
14:16:38 <bexelbie> cool!
14:17:19 * pbokoc 
14:17:47 <randomuser> hello, Petr
14:18:10 <pbokoc> randomuser, sysadmin guide, networking guide, selinux guide are ready, install guide will be done in the evening. Relnotes are mostly ready but there's an open pull request on them I think
14:18:44 <randomuser> yeah, i saw PR late last night
14:18:50 <pbokoc> also the virt getting started guide *might* be ready - I bumped productnumbers and everything, but I have no idea what state the whole guide is in overall. grundblom might know
14:19:35 <randomuser> iirc grundblom did some updates in this cycle, and it's not subject matter that will have changed
14:19:51 <randomuser> thanks for the bump, we'll ship it
14:20:20 <pbokoc> cool. Have fun publishing 6 guides :P
14:20:35 <pbokoc> I'll ping you when I'm done with the install guide
14:20:44 <randomuser> it's really not as bad as you might think once the setup is doen
14:21:09 <pbokoc> I've done it before and it still haunts me - and that was before you had to use a f18 VM to do it...
14:21:18 <randomuser> except that I'm afraid to script it b/c there always seems to be some extra place the version is declared in one of them...
14:22:04 <randomuser> #link https://fedorapeople.org/groups/schedule/f-25/f-25-docs-tasks.html
14:23:04 <randomuser> I should probably look at translations, too
14:24:02 <bexelbie> How long does it take translations to get pulled?
14:24:07 <randomuser> post-GA I think a service FAS user and script to cycle strings with Zanata will be my priority
14:24:22 <pbokoc> yeah, I updated zanata.xml in all books I was bumping for f25, too. One problem is that the lang map that was previously in the XML that mapped zanata's lang codes to publican's is deprecated, and now you're supposed to set those as aliases in each project
14:24:36 <pbokoc> I did that for relnotes but it took me forever, and I haven't found a way to copy that between projects
14:24:48 <randomuser> it's said that for a while, i thought the map still worked
14:25:15 <pbokoc> hm, it might - I haven't noticed the message before
14:25:33 <randomuser> bexelbie, it's manual review of each zanata project, then you know how long it takes to pull strings, then a publishing action for each language
14:25:46 <bexelbie> ahh, good times then
14:26:15 <pbokoc> also it doesn't really matter if we're going to get a new site for f26. We can just start using zanata's language codes. (And even if we won't, I suppose we could just set redirects like /fr/ -> /fr-FR/ etc.)
14:26:31 <randomuser> last time I checked the API didn't provide the right info for determining a language's readiness... but i'll check again after this cycle
14:26:38 <pbokoc> wait, no, that wouldn't work, never mind me
14:28:08 * bexelbie is so hoping for using Zanata language codes
14:28:13 <randomuser> if anyone wanted to pull and commit translations in advance it would be helpful
14:28:39 <bexelbie> randomuser, afaik from talking to Alex there is no api for readiness as it isn't a status in Zanata (if I understood him correctly)
14:29:00 <bexelbie> I suspect we need to build something side loaded to manage this.  It sounds like there may be consensus to set a % translation to check against and then just keep on building
14:29:10 <bexelbie> I think we are also going to need to think about caching, possibly
14:29:46 <randomuser> caching, you say?
14:31:40 <randomuser> i'm not sure where caching would be a good fit here
14:31:48 <bexelbie> the pulls are so slow, depending on how we choose to finally do the translation builds, we may want to cache the files
14:32:09 * randomuser nods
14:32:35 <randomuser> pulling translations and publishing need not be a linear sequence
14:33:32 <bexelbie> I like my processes like I like my lines, straight
14:33:33 <bexelbie> :P
14:33:33 <randomuser> historically we've committed contributed translations into the release branch; afaik that's typical a upstream procedure
14:34:05 <randomuser> if the publishing bit only needs to rely on having available committed strings, it won't be so bad
14:34:18 <bexelbie> randomuser, I hate to think of the clone times for those repos :P
14:34:28 <bexelbie> not all of our users are on high-speed connections
14:34:35 <randomuser> hm
14:35:47 <randomuser> bexelbie, otoh, I'd hate to think of relying on a 'third party' as the sole source of a significant portion of our contributed content
14:36:13 <bexelbie> third-party? zanata? low-bandwidth users? a second git repo?
14:36:17 <bexelbie> what do you mean?
14:36:59 <randomuser> one generally expects an upstream project to have committed everything they ship
14:37:23 <randomuser> although, your leading questions suggest you have a solution in mind :)
14:37:45 <bexelbie> well, I consider the Zanata instance to be part of Fedora - if this isn't the case, we probably need to rethink a lot of other users of it
14:38:08 <bexelbie> I was thinking we could cache the translated files to an second repo so they are easy to get to, but don't create clone/clogging issues in the content repo
14:38:22 <bexelbie> Theoretically we republish for translation more often than we republish for new content
14:38:50 <randomuser> I consider fedora's ownership of zanata to be slightly more than of bugzilla
14:39:09 <randomuser> it's a dedicated instance that not-fedora operates
14:39:41 <bexelbie> Then it sounds like I think we should have a cache and you think we should have a copy so we both win :P
14:39:58 <randomuser> ...presumably the application architecture leads to a separate instance and not a contextual view of a shared instance
14:40:21 <bexelbie> probably? (not sure I followed all of that, sorry)
14:40:37 <randomuser> bexelbie, well, you say caching and I think, oh, varnish or something
14:40:54 <randomuser> the cache should go behind the API endpoint, not in front
14:41:13 <bexelbie> Ahh, I mean caching files for later processing so I don't have to get them from the source location.  I am not speaking at all about web technologies
14:41:48 * randomuser nods
14:42:07 <randomuser> both good ideas then :)
14:42:34 <randomuser> let's stay in touch as we work out a solution
14:42:37 <bexelbie> yes
14:42:45 <bexelbie> and work we shall ... very hard ... after turkey-day
14:43:34 <randomuser> #chair bexelbie pbokoc
14:43:34 <zodbot> Current chairs: bexelbie pbokoc randomuser
14:44:02 <randomuser> I need to begin my commute momentarily, anything else we should cover while I'm here?
14:45:32 * bexelbie is good
14:45:36 <bexelbie> safe driving
14:45:52 <randomuser> no promises
14:46:12 <randomuser> okay, see you around everyone
14:46:20 <randomuser> #endmeeting