Web   ·   Wiki   ·   Activities   ·   Blog   ·   Lists   ·   Chat   ·   Meeting   ·   Bugs   ·   Git   ·   Translate   ·   Archive   ·   People   ·   Donate

#sugar-meeting meeting, 2014-08-06 22:05:39

Minutes | Index | Today     Channels | Search | Join

All times shown according to UTC.

Time Nick Message
22:05 meeting Meeting started Wed Aug  6 22:05:39 2014 UTC. The chair is bernie. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:05 Useful Commands: #action #agreed #help #info #idea #link #topic #endmeeting
22:06 tch__ ok,
22:06 since the last meeting, we upgraded the VM to newer ubuntu, we revived the 2.5 instance and we migrated it to 2.5.1
22:07 the problem now, even after reading the translation wiki sections, I lack of context to determine if the instance is working sucessfully
22:07 s/successfully/properly
22:07 cjl I can enter words
22:08 tch__ also, I am a bit confused by some diferences between the fresh 2.5.1 instance I set at home, and this migrated one
22:08 cjl honey2 / cookiesearch.po completed, I don't see commit button though
22:08 tch__ ie,. I should be seeing VCS options in the project, but I dont
22:08 cjl, exactly,
22:09 cjl Did yiu import from old pootle backups in your fresh instance?
22:09 tch__ cjl, no, I added it as a new project following the procedure in pootle 2.5.1 docs
22:10 cjl, that means without the custom scripts, but I see now that the custom SL scripts does a lot of things,
22:11 cjl, I am a bit worried about my lack of information about the use cases for the instance,
22:12 cjl How can I help explain use cases?
22:13 Coder writes activity, generates POT once, never touches the git /po directory again (under old Pootle workflow).  The scripts did the rest.
22:14 tch__ i see,
22:14 cjl Translator works on words, submits strings, lang admin does commit to VCS
22:15 potupdater script synched the pootle-local git repo with git and generated new POT nightly (if needed)
22:16 As Pootle admin, I would do "Update from Templates to roll out new POT to all langs.
22:17 tch__ can you write down these steps and specify which tools you used for each?
22:17 cjl yes
22:17 bernie cool
22:17 tch__ is a bit confusing to tell what was done before through pootle itself and custom
22:18 bernie btw, since we're redoing a few things... wouldn't it make more sense if pootle did *not* write into all git repos like this?
22:18 cjl Yes, the added VCS stuff in Poolte is confusing
22:18 bernie if i was the maintainer of an activity, i'd want to *pull* from the pootle git repositories into mine.
22:18 tonyf <tonyf!~webchat@rev-18-85-44-69.sugarlabs.org> has joined #sugar-meeting
22:18 cjl bernie, idea is for lang admin to have control of push
22:18 bernie and i'd do it just before a release, for example.
22:19 cjl but I am not adverse to change
22:19 averse
22:19 bernie cjl: but the language admins can't release a new version of the activity bundle, can they?
22:19 cjl no
22:20 tch__ bernie, I actually research a bit about that, looking if it was possible for pootle to do PR in github, but I found nothing
22:20 cjl I would run through and do a lot for lang admins
22:20 bernie cjl: oh maybe they can make a custom bundle and put it on their deployment activity collection, for example
22:20 tch__: yeah it would be great that way
22:21 cjl do you eman like our lang_packs?
22:21 bernie tch__: changing topic: did you get a chance to extract the last state from the backups?
22:21 cjl: is it important to recover the list of projects, users and ACLs from the backup?
22:22 cjl yes
22:22 bernie tch__: if you could upgrade pootle and preserve the data, then it should be easy to do it again
22:22 tch__ bernie, not really, I focused on making it "work... to many things at once is too hard to handle
22:22 cjl Trying a recovery into your instance would be a great next step
22:22 tch__ bernie, that is what I did,
22:22 bernie tch__: the backup should have a directory with a dump of the mysql db... at least, all machines i setup do daily mysql dumps in their /backup dirs.
22:23 tch__ I mean, I used the db that was left in newpootle
22:23 working in that machine is a bit hard though, too many db copies running, different users owning different things, looks like it all attached with bandaids
22:23 bernie tch__: if the mysql dump isn't there, you could copy the entire /var/lib/mysql directory to the new server, but that's more of a "dirty" trick and i'd run a full dump/restore to make sure there's no corruption
22:24 cjl newpootle was never advertised as active
22:24 bernie tch__: no, wait... newpootle had a really old snapshot of pootle. maybe one year old
22:24 cjl getting db from housetree pootle instance would be impt
22:24 bernie tch__: the last state of pootle is only in the daily backups of the old pootle vm
22:25 cjl: we have that in the backups, but the db files might be corrupted, who knows. the disk was dying after all.
22:25 tch__ there is another issue, what we call old what version of pootle was?
22:25 bernie tch__: err... dunno. older than 2.5, i think
22:25 tch__: it might require a more complicated migration
22:25 tch__: but your predecessor could do it so it's a good sign
22:26 tch__ are you sure these were in mysql already? I thik the previous guy who worked on this had to migrated to different versions to finally migrate to 2.5
22:26 bernie, I was actually thinking the opposite
22:26 bernie tch__: possibly. i remember he had all sort of trouble with migrating the data :-(
22:26 tch__ bernie, to have a fresh one working and importing projects back to it
22:26 cjl In worst case, I'd take a loss of work t ohave a working Pootle instance
22:26 bernie tch__: but then again, he also didn't have much experience with django, mysql and all that.
22:27 tch__: othet than the projects, there are users with access rights, passwords, etc.
22:27 tch__: maybe not enough users to care.
22:27 tch__ i don't think django is the main issue,  but to understand pootle itself and then our particular way of using it
22:27 bernie cjl: wouldn't the actual translation work be in the git repositories?
22:27 cjl: what's in the db?
22:28 cjl bernie only if committed, but most hopefully is.
22:28 submitted, not not *committed* would be lost.
22:28 bernie tch__: getting the git repositories from the backups should be very easy...
22:29 tch__ cjl, maybe we can export it using the offline optoins and import it back?
22:29 cjl PO filesimport well (usually)
22:29 tch__ bernie, the thing is that we need to change repositories too,
22:31 cjl, going back to the pot files
22:31 cjl yeah
22:31 tch__ cjl, we needed to use cli scripts to regenerate them?
22:31 or was cron job
22:31 cjl we did, not sure what other Pootle instances did
22:32 cron job
22:32 code is in git "pootle-helpers"
22:32 tch__ cjl, ok
22:32 cjl, another topic,
22:32 cjl, the current projects organization
22:32 cjl https://git.sugarlabs.org/pootle-helpers
22:33 tch__ cjl, we also need the custom scripts to organize the FS to make pootle think we have fewer projects right?
22:33 cjl maybe only small change from old pootle to new pootl3e
22:34 tch__ to have glucose, honey, etc
22:34 cjl scripts created symlinks per lang per PO
22:35 Not sure what issue is, Pootle handles multiple projects fine.
22:35 tch__ cjl, yeah, but real projects are not organized like that
22:35 ie., sugar is different project fro sugar-toolkit
22:36 but we use these scripts to make it look like is one project
22:36 cjl can have multiple PO files per project, we did set each PO up pointing to it's own repo
22:36 Honey was essentially arbitrary grouping
22:36 tch__ my question is, do we need to preserve that organization ?
22:37 bernie yes
22:37 sorry, responding to cjl
22:37 cjl It means a lot to developers
22:37 tch__ I am just thinking about the cost it has in terms of maintainability to depend on those scripts
22:38 just to have it organized in that way
22:38 lets say (hypothetically) that pootle can cover a basic workflow now
22:38 cjl It wsn't too bad in the past
22:38 tch__ like
22:39 updating POT, making commits and pushing..
22:39 out of the box
22:39 wouldn't it be easier to use it as it is?
22:40 cjl, yeah, but in the past we had maintainers :)
22:40 haha
22:40 cjl one project per PO?
22:40 tch__ one project per project
22:40 sugar as pootle project, sugar-toolkit as another, TA as another, etc
22:41 cjl hmm
22:41 bernie i think it makes more sense this way, yes
22:41 tch__ I am just thinking aloud,
22:41 I still don't know if pootle can cover these things out of the box,
22:41 bashintosh sorry folks to drop in but, do you really need to have Pootle doing all these things, all at once, for all these users? I mean, wouldn't it be easier to start (almost) from scratch and re-architect the whole thing with a bit of maintainability in mind? At least, it would be time spent at the beginning but then much easier for others to come. Just my 2 cents, I don't yet know enough about all this
22:41 bundle..
22:41 tch__ but I do know what the git integration works like that
22:41 bernie tch__: i'm a bit outdated. have sugar, sugar-base and sugar-artwork been merged now?
22:41 tch__ out of the box
22:42 bashintosh, I agree, always we can cover a decent workflow
22:42 bernie, yes, sugar-base is gone
22:42 bernie, sugar-artwork is still there
22:43 cjl but no strings :-)
22:43 bernie bashintosh: makes sense, yes. i think tch agrees too.
22:43 cjl: so it doesn't matter for pootle :)
22:43 cjl: then which projects would have to be split up for pootle?
22:43 cjl I'm not going to argue against any organization that can be made t owork in a maintainable fashion
22:44 Fructose and multiple Honeys, at least
22:45 tch__ amost of the projects I think
22:45 ie., sugar and sugar-toolkit are just one project in current pootle
22:45 but ar different projects in git
22:46 cjl We might be able t odrop Dextrose (no AC anymore)
22:47 tch__ cjl, think it would be possible to specify which projects are essential? in case we need to start from scratch
22:47 I know sugar is :), but what about others, like terminology
22:47 (keep in mind I lack context)
22:47 cjl yes, I could list esssentials
22:48 tch__ cjl, great, that makes two deliverables now :)
22:48 bernie guys sorry i have to leave now
22:48 i'm hosting an event, people will come soon.
22:48 tch__ bernie, ok, thanks for joining bernie
22:49 bashintosh bernie: see you!
22:49 bernie bye! thanks for taking care of the dog (pootle)
22:49 bashintosh, tch__: maybe the two of you should spend some time together, going over the current pootle setup.
22:49 better have two maintainers for pootle, since it's so tricky
22:50 tch__ bernie, yes,
22:50 bernie bye!
22:50 leaves
22:50 tch__ thats why I am thinking that, if the new pootle can cover a decent workflow, we should it as it is
22:50 bashintosh bernie: ciao
22:50 tch__ as much as possible
22:51 bashintosh tch__: are you documenting all the steps you're doing somewhere?
22:51 cjl ciao bernie
22:51 tch__ bashintosh, more or less, making the current instance to work was more of a try and error thing
22:52 walterbender cjl, ping
22:52 cjl Let me work on these edeliverables, especially what the pootle-helpers did for us.
22:53 walterbender: pong
22:53 tch__ cjl, thanks Chris, what I will do is see how much the new pootle can really do..
22:53 walterbender cjl, I happen to be in DC tomorrow.
22:53 any chance to meet for lunch?
22:54 cjl I have something around lunchtime, but maybe after?
22:54 walterbender I have a 3pm meeting at the World Bank
22:54 but maybe dinner?
22:54 cjl Dinner would work
22:55 walterbender cjl, email your phone number
22:56 tch__ bashintosh, once we get a better idea of what path to follow, I will definitely need a hand to set and document everything :)
22:56 bashintosh tch__: OK, it looks like a very custom edit of the original Pootle especially because there's a lot of "legacy" things that you guys are trying to get back on track.
22:56 tch__: you're the boss, tell me what bolts to turn :)
22:57 tch__ bashintosh, cjl I think thats it for today..
22:57 bernie, bashintosh cjl thanks a lot for coming..
22:57 cjl yep, let me work on some descriptions of old workflow
22:58 tch__ cjl, many thanks
22:58 bashintosh tch__: cjl bernie pleasure to be here!
22:59 tch__ #endmeeting
23:00 maybe bernie has a control haha
23:01 bashintosh tch__: lol
23:06 tonyf has quit IRC
23:40 satellit__e <satellit__e!~satellit@2601:8:8800:62a:468a:5bff:fe34:b748> has joined #sugar-meeting
00:50 CanoeBerry <CanoeBerry!CanoeBerry@c-98-216-39-142.hsd1.ma.comcast.net> has joined #sugar-meeting
00:51 CanoeBerry_ has quit IRC
02:34 walterbender has quit IRC
02:45 curiousguy13 has quit IRC
02:57 curiousguy13 <curiousguy13!~curiousgu@> has joined #sugar-meeting
03:16 tch__ has quit IRC
03:53 CanoeBerry has quit IRC
04:01 curiousguy13 has quit IRC
04:28 reubencaron_ <reubencaron_!~reubencar@activation.laptop.org> has joined #sugar-meeting
04:36 reubencaron has quit IRC
06:51 CanoeBerry <CanoeBerry!CanoeBerry@c-98-216-39-142.hsd1.ma.comcast.net> has joined #sugar-meeting
06:59 walterbender <walterbender!~walter@pool-108-48-0-210.washdc.fios.verizon.net> has joined #sugar-meeting
07:38 walterbender has quit IRC

Minutes | Index | Today     Channels | Search | Join

Powered by ilbot/Modified.