#foswiki 2014-03-21,Fri

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)

WhoWhatWhen
pharveywhat do you think of the github move sven
should the default + core be one repo
[00:00]
SvenDowideitSvenDowideit says yes [00:00]
pharveyI know you're a fan of the uber-fat repo :) [00:00]
gac410so keep the github/foswiki project, but build different structure under it? /me is getting enough info to be dangerous. [00:00]
SvenDowideitI do driveby commits to all sorts of projects
I like fat repo, but 99% of people do not
[00:00]
gac410We can't use server side checks for task status, tidy code, etc. if we move to github rather than our own repo. [00:01]
SvenDowideitso I would work to simplify the 99% [00:01]
pharveygac410: https://help.github.com/articles/creating-webhooks [00:01]
SvenDowideityou can, you need to write foswiki<->github integration plugins and webhooks [00:01]
pharveythere's a github API on cpan [00:01]
SvenDowideitthe other thing that should be as simple as possible, is allowing users to integrate their plugin repo into a running foswiki
plus
[00:02]
gac410yeah. I've written two local hooks pre-commit and commit-msg, that does the tidy and task checks locally, but we can't put server side hooks on github to block a push. [00:02]
SvenDowideitit would be possible to redo the tasks system to push&pull from gh issues
so you can use foswiki as a frontend, and GH issues as a store
with local cacheing
[00:02]
gac410I was thinking a webhook on github could update tasks web - post each push to a helper on f.o [00:03]
SvenDowideitso its all _possible_, and it might be worth it from a showoff how good foswiki is [00:03]
pharveyit's all possible; but it seems like the first step is to re-carve ourselves a default + core repo [00:03]
SvenDowideityup [00:03]
gac410The biggest reason to move to github is because we no longer have the developers with time / effort to do the foswiki.org hooks. I have little hope that we would be able to get full Tasks web / gh tracker integration. [00:04]
SvenDowideityup [00:05]
pharveyI'll try to give this some time over the weekend [00:05]
SvenDowideitI would probably move everything to GH, but that assumes that doing so will make more people contribute [00:05]
pharveythat's a good point [00:05]
gac410I think if we used github for the repo, and keep the Tasks web as-is. [00:05]
SvenDowideiti find it really useful to not have to work out how to contribute
i also prefer to do a quick commit to projects that use the GH PR's
as the issue and review system is thus known to me
[00:05]
gac410About the *only* wiki use I make of foswiki is the Tasks web. So for me, I'd be a developer that isn't a user if we do away with tasks. [00:06]
SvenDowideitif a project uses bugzilla, i tend to move on
and foswiki tasks is even less well known
yeah, i know, that was always my argument too
[00:06]
pharveydogfooding is important, but hrmmm. [00:07]
gac410github has a wiki too, maybe we don't even need to run a wiki :P [00:09]
pharveylol [00:09]
gac410Given our 1000's of tasks, would trying to migrate them into github issues make any sense. Or do we just "flush" the history and start over.
Seems like it supports a simple label tagging of issues. Nothing like how we manage issues.
[00:12]
.... (idle for 15mn)
Cervatorhey all - neat to hear about the idea of adopting some use of GitHub, i've long since moved there myself - have you considered a one repo per plugin approach? that may sound overwhelming at first glance, but Jenkins CI is making it work with hundreds of plugins
i've adopted something similar for my main hobby project these days, although both it and Jenkins are in Java with some solid IDE integration potential
[00:29]
gac410That seems to be the leaning. Or maybe the core + default extensions in one repo, and then everything else a repo per extension. [00:30]
Cervatori've got gradle tasks that'll fetch plugins and prep a local dev environment, likewise able to push back to multiple repos in one go, issues per plugin can get their own tracker, and so on
automating all the things is win either way - you can literally set up to run from IntelliJ with our project with a git clone and single Gradle task, that sets up all sorts of customizations locally to make everything easier :-)
need a plugin? one more gradle task and it is set up locally, no direct Git/GitHub interaction needed
[00:30]
gac410It's hard to justify moving away from the current tools. They do work pretty well. I know I don't have the time to rebuild our infrastructure. It "just works" [00:36]
Cervatoryep, definitely understandable, and i'm not sure if there are any ideal replacements for a Perl-based project, mainly familiar with Java tools [00:36]
gac410Moving to git is a no-brainer. I think that's generally accepted by the devs. github vs. our own repo ... leaning toward github I guess. [00:37]
Cervatori must admit i was one of the would-be contributors here who started down the path to set up to code but got lost along the way before i had a working local setup [00:37]
gac410What was the main difficulty? The subversion checkout? Foswiki / apache in general? [00:38]
Cervatorit was years ago, hard to remember for sure - i was even new to SVN back then
i think i just barely managed to check out the source, but then probably didn't get a working setup from that point
i still maintain a foswiki instance at work, although maintain might be generous - it could be so much more than it is but no time
i've been able to get the basics working on Apache, managed LDAP integration too, any more advanced Apache than that or running from source i didn't make it to (still haven't taken advantage of fastcgi or other setups like that)
tasks web ... was a little tricky as well ;-)
even tried to make my own at home/work for a while too
(issue tracker)
[00:39]
pharveywe do already have one repo per plugin, I personally like it but we haven't found a good workflow - partially due to the way foswiki developer filesystem layout is structured perhaps [00:42]
Cervatori'll say this - while the Gradle setup i've gotten used to feels like magic when it works, it can get a tad hairy when it doesn't :-) [00:43]
gac410There is a lot to tasks web, but importing 1000's of tasks vs. discarding history, ... tough call. [00:43]
Cervatoryeah - any chance to piecemeal them out to individual trackers per plugin to cut down on the apparent "weight" ? although that might make it harder to get a good overview
one of my nefarious plots is to tie plugin GitHub repo trackers into a single JIRA instance (one plugin repo == one "component" in JIRA) so you can have both sides, but ...
[00:44]
...... (idle for 28mn)
gac410Cervator: One issue is that issues are very frequently mis-categorized. For users to report, the need to figure out which extension / repo to report the problem. If each tracker is separate, it means moving tasks around. [01:12]
Cervatoryep, true that
could be done by synchronizing issues but then that's more logistics to worry about
[01:17]
...... (idle for 28mn)
gac410I've added some tables to vote / express opinions on migration to github. repo, tasks, etc. Foswiki:Development/MoveCodeRepositoryToGit Topic is getting a bit large. [01:45]
FoswikiBothttp://foswiki.org/Development/MoveCodeRepositoryToGit [ MoveCodeRepositoryToGit ] [01:45]
Cervatoron seeing the note there on pre-commit hooks - could that not be achieved with PR tests and/or pre-acceptance branches/forks? in a Java project i'd run a build on-commit and do any sort of analysis and tweaking at will, then only promote if all is well
admittedly that comment could be from 2011 ...
[01:50]
gac410pre-commit hooks are client or server side. anyone can install a client side hook but they are completely optional No way to enforce their use. [01:51]
Cervatorsort of like running Checkstyle on a java project - we can make it available but not force contributors to use it [01:52]
gac410Server side hooks, github doesn't seem to support. It needs to be server side to block a push [01:52]
Cervatoralthough then we can complain when a PR comes in that doesn't pass Checkstyle :-) build analytics can alert on that sort of stuff
there are a *ton* of webhooks, looked through them to see if anything similar might stick out ?
[01:53]
gac410yeah, it adds a lot of dev checking though. We have become very dependent upon the svn server side hooks to block commits. [01:53]
Cervatorunderstandable, at work Perforce triggers watch like hawks for malformed changelist descriptions
anyway - what i mean to suggest is that on GitHub when a PR is raised against a Foswiki repo you can automatically run a pull request builder type process to test the PR automatically, even post comments to the PR with what's wrong
[01:54]
gac410That means we need a controlling dev to check each commit. We don't do that today. Each dev is on their own, and we know that they can NOT commit if the code is not tidy, or if the task is in the wrong state, etc. [01:56]
Cervatortrue, although on GitHub that's nearly trivial :-) get PR, check comment from PR builder bot, press button, receive bacon [01:57]
gac410What's a PR? [01:57]
Cervatorpull request, a contributor puts a change in their fork of the project then packages it as a change (think of it as a patch perhaps) against the root repo (or any other fork) [01:58]
gac410I guess it's a completely different model of development. Who becomes the god that approves stuff. Core developers just commit against the repo. [01:58]
Cervatorit really is, getting used to Git and GitHub took me a long time and it really is different culture-wise
but then that's where you get that "social coding" aspect
[01:59]
gac410I suspect with the amount of dev activity lately, if we all had to wait for another dev to approve our PRs work would stop. [02:00]
Cervatorfor my hobby project we've got dozens of contributors and two of us check PRs and comment occasionally, not a very heavy burden
if you're very "trunk heavy" i could see that
it was much worse for us when all the plugins were in the engine repo, then stuff really would pile up, and be unrelated no less
then we split them out into their own repos, put the plugin/module author or maintainer in charge so they can push their own code and accept PRs from others
[02:00]
gac410I don't think the project could live with the RM(s) needing to approve PRs The core devs are all trusted. So would we be all making our own PRs and approving our own pulls? [02:01]
Cervatorour engine repo shrunk dramatically after that, which means engine activity is relatively low
well, 2-3 of us do push straight to the engine from time to time, other times we make PRs specifically to review when we could otherwise just have pushed
[02:01]
gac410Foswiki engine is where much of the work has been. [02:02]
Cervatoryeah that probably would be tricky then [02:03]
gac410I can definitely see extensions becoming managed by their primary dev's Essentially our "Coordinate with author" setting on an extension. [02:03]
Cervatorfor us the few core devs are likewise trusted to run Checkstyle and such before doing direct pushes - but only 2-3 of us plus much smaller engine ... [02:03]
gac410We found even with the trusted devs, the only way to get perltidy consistent was to block the checkin [02:04]
Cervatori think i read about a smoke testing branch/repo somewhere, maybe that could work
core devs can push directly there, analytics run, if they all pass then auto-push to main/trunk
[02:05]
gac410We sort of have that with the nighly test runs, but they often don't pass currently :) [02:06]
Cervatorheh, heh
anyway - finally made it to the bottom of that proposal page in between other work
adopting Git in general probably comes down to also matching it with suitable codebase structure - with Git commits being across the whole codebase anything monolithic gets hairy fast, much easier to deal with that in SVN, Perforce, etc
i don't recall if there were ever any ideas about splitting the engine into smaller modules? something like that would be a lot more Git friendly, but of course the sheer effort to do so if it isn't trivial .... yeah :-)
that's been one of my biggest lessons converting from Perforce where i could have giant death-ray-yet-useful revision graphs per file to Git where each commit is the state of the whole repo at a point in time
very different way of working
also, i stay away from Git command line if i can avoid it :D
[02:06]
gac410100% command line here :)
Much of the core is probably too inter-twined to break up. The "default" extensions could be separate repos. But combining them probably makes sense from a release management perspective.
They are all managed as a unit.
[02:13]
Cervatorcould possibly be done at build-time ? but yeah, again quite understandable :-)
we might again have it easier as we didn't start *that* long ago (3ish years?) and had contributors and structure added in fairly early
on top of that it is a game and we're thinking about splitting more things out of engine like GUI, our entity system, physics, math, etc
probably easier to attract random GitHubbers that way too, we're past 300 forks
incidentally, i remember why there might also be some resistance to anything resembling a benevolent dictator approach even remotely :3
[02:22]
gac410:D
I've been the ReleaseManager for the last few releases and I agreed to handle 1.2, But I really truly do not want any day to day responsibility of reviewing commits.
And I'm reasonably certain others would not want me to do that either :)
[02:25]
Cervatori hear ya :-) i try to do that as project lead, at least for engine, but cannot keep up
we've got a solid lead architect who manages though
[02:27]
gac410I think 1.2 has taken too long and folks have drifted away. The foswiki camp lit some fires, but we need to keep the momentum. [02:28]
Cervatoryeah - agreed there, i kept thinking i'd wait to upgrade the work foswiki till 1.2 ... that was probably a couple years ago by now :3
on a somewhat related tangent i imagine a multi-repo / assembly approach could do wonders for the different "flavors" you could release / wrap foswiki in, if there are still different installers/VMs and such floating around
but again it comes down to putting together a bunch of automagic to make it all work with minimal effort
that's my personal passion, i'm a total logistics monkey, but the finite number of hours in a day disagrees with my desire to automate all the things!
[02:28]
gac410Our vm's and other installers have been one-off occasional contribs and had fallen way behind. I was embarrassed enough by some recent comments on the state of the 1.1. vmware vm that I spent a day and updated it.
Should be in good shape now.
[02:32]
Cervatorcool [02:32]
gac410Updated it to fcgi as well, so it should be pretty fast right OOTB [02:32]
Cervatoryeah that's the main thing hurting the foswiki at work, never having had time to optimize it -_-
yet it *still* beats the corporate portal, heh
[02:33]
gac410:)
Cervator: Do you dev on Windows or Linux or ...?
[02:33]
Cervatormostly Windows, but the Foswiki at work is hosted on RedHat (and i tinkered with one in Fedora at home for a while) [02:35]
gac410Recent had some comments on another project that a perltidy git hook just won't work due to the state of git on windows. [02:35]
Cervatorwish i had more opportunity to dev on Linux, but hard enough to find time to do hobby-work, let alone relearn dev on a different base OS :-)
and hmm
i don't think i've gotten advanced enough with Git to hit any OS specific issues, outside of the ever-present line endings of doom
[02:35]
gac410:) [02:36]
Cervatormost of our magic is present in Jenkins [02:36]
gac410Oh okay. [02:36]
Cervatorincluding an overly elaborate release promotion system i'm totally finishing any day now
...
Git is scary enough as it is, been trying to keep it simple :-)
oh, incidentally, i'm RasmusPraestholm on Foswiki, didn't even think about having switched to my usual nick who knows how long ago
[02:36]
gac410:)
Yeah sometimes it's really tough to associate irc nicks with wikinames.
[02:39]
Cervatoryep!
i think i put myself in an IRC table on the site a loooong time ago, if it is still there somewhere
[02:40]
gac410no idea. Usually I just right-click the nick in pidgin to get the "Real Name" ... but your's doesn't help :)
One more question ... thoughts on hooks that actually modify the commit. Everything I've heard says that's a bad practice.
[02:42]
Cervatoryeah i'm online through Thunderbird's IRC plugin, not really support for a lot of IRC fanciness ... :-)
well, don't modify the commit, just add more on top if needed
yeah you don't want to edit existing commits or rebase code that has been pushed to any sort of shared repo
[02:43]
gac410Another project wanst to use githook-perltidy to tidy the file during the commit, All we intend to do is reject non-tidy. [02:45]
Cervatorwhat build tools are actually involved, if any? been too long for me to remember plus it's perl :-)
just hooks like that on commit?
[02:46]
gac410No idea. I think it's everyone on their own. [02:47]
Cervatorhehe
i know Jenkins isn't exactly built for Perl, but it sounds like some people are using it: https://wiki.jenkins-ci.org/display/JENKINS/Perl+Projects
and i'm sure there are other CI servers that might support Perl better
[02:47]
pharveyI certainly use it [02:47]
Cervatoroh good :-)
rather than rely on hooks you might be able to involve a CI server with full-blown code analytics in it, allowing some sort of auto-promotion if the analysis is good
that would also help standardize the process and integrate with GitHub better
[02:48]
pharveyI only recently decided to pick up jenkins, after using it for a while at my previous workplace [02:50]
gac410I think that would be really tough to enforce. I'm just not a gui guy. My experiences with eclipse have been really negative. I'd have to upgrade my pc for one thing. [02:51]
pharveymy current setup does tag and push the tag if the build passes
failing builds aren't tagged, not sure why I set it up that way upon reflection
[02:51]
gac410Are you using it for foswiki? [02:52]
pharveyno [02:52]
CervatorEclipse? no Eclipse involved, unless you want it :-)
we have devs working in IntelliJ, Eclipse, NetBeans, text editors - in the end it all goes to GitHub and Jenkins
[02:52]
pharveyTravisCI is free and seems popular with open source perl bits
but it's non-free in the open source sense
more reliance on .com infrastructure :/
[02:53]
CervatorTravis has some solid GitHub integrations too
trying to set them all up myself with Jenkins, but getting there
[02:54]
pharveyHaving worked on a big perl codebase that did perl the "CPAN way" (Build.PL and dist.ini style modules)
It would be awesome if FW was a little more cpan-ified
sigh
first things' first... carve up the repo properly :}
[02:54]
Cervatorhear hear :-) [02:55]
pharveydon't know if Babar/gmc would care for me putting java crap on the f.o server :P
Jenkins Web UI doesn't seem terribly strong from a web-facing security standpoint
I'd feel nervous about exposing it on the web
[02:56]
Cervatorwe've had ours on the web for about 2.5 years so far no problem, but admittedly we're still low key and stealthy [02:57]
pharveythere are quite a few public-facing jenkins instances [02:58]
gac410Watching our server logs .. .and dealing with the wiki spam. we are under constant attack. [02:58]
Cervatorwhat you can do is run a primary instance behind a firewall and proxy one for public display, publishing jobs to it [02:58]
pharveyperhaps I'm being nervous about nothing [02:58]
gac410pharvey. I added some "vote" tables at the bottom of the MoveCodeRepoToGit [02:59]
@(#$# system ran out of X resoures again. This is really annoying Back in a bit. [03:04]
***ChanServ sets mode: +o gac410 [03:05]
pharveyI keep writing build and system scripts in perl, even though I work on a python product
pharvey wonders if he's going to hell
at least I'm using Pod::Usage :)
[03:12]
Cervatori'm trying to avert a powershell-based apocalypse at work but it isn't going well :-( /hugs Groovy [03:13]
pharveyoooh powershell
actually powershell has surprised the hell out of me sometimes
[03:13]
Cervatorit has its uses for Windows, yeah - but i'm trying to get away from Windows :D [03:14]
pharveydo you still have to drop back to cmd.exe to do mklink? [03:14]
Cervatordunno, i'm managing to not write any myself, it is just being used by others more commonly [03:14]
pharveyah, there's a powershell module which gives you a mklink command
lame
[03:15]
Cervatorhaving all kinds of fun dealing with PS profiles, bit-counts, differing environment variables, and all that -_- [03:15]
pharveyI've given myself a 30 line limit for bash.sh scripts. If it gets bigger I see if it makes sense to perlify it [03:15]
Cervatoras opposed to telling Jenkins "Here, simply copy this Groovy where you need it then run it!" [03:15]
pharveydid I mention SvenDowideit infected me with the Docker disease? [03:16]
Cervatori'm not managing to keep up here quite well enough to notice that sort of stuff :3
but it looks tasty, much love for automation tools of all sorts
[03:17]
pharveyI'm really just using it as a lightweight vagrant
not actually deploying anything with it
[03:17]
Cervatorbrb [03:18]
SvenDowideitmeh
on windies, i'm using the msys bash and tools that come with git
[03:22]
................. (idle for 1h20mn)
***gac410 has left [04:42]
...................... (idle for 1h49mn)
ChanServ sets mode: +o CDot [06:31]
.......................................... (idle for 3h29mn)
ChanServ sets mode: +o Lynnwood [10:00]
.... (idle for 19mn)
ChanServ sets mode: +o pharvey [10:19]
ChanServ sets mode: +o pharvey
JulianLevens has left
[10:25]
............................. (idle for 2h22mn)
GuilainCraaahhh a query search make me crazy !
hi everyone
in one search (query type) i've : %SEARCH{" d2n(fields[name='Datededbutdepriodesansmission'].value) > d2n('%DATE%') "
and in my dataform i've that : %META:FIELD{name="Datededbutdepriodesansmission" attributes="" title="Date de début de période sans mission" value="4 Avr 2014"}%
the problem, is the search doesn't work or "today < 4 Avr 2014"
any help should appreciate :)
[12:47]
i'm follow a track ! my date are localised in my dataforms
(french localisation)
[13:01]
jastside note: a shorter way to write fields[name='foo'].value in query syntax: foo [13:09]
***ChanServ sets mode: +o gac410 [13:15]
GuilainCtrue jast, it's an old code which i'm upgrade, thanks for tip
is it possible in search query to call spreadsheetplugin ?
[13:24]
jastI don't think so [13:25]
GuilainCfor doing something like this : %CALC{$SUBSTRING(fields[name='foo'],7,4)}%
ok...
[13:26]
gac410hm.. inside-out left-right, I think you can expand the %CALC as a value inside a search [13:26]
GuilainCno idea, for the moment, so delay to an other time
the objective is to opera on fields
in the search
[13:26]
jastright, the constant part you can treat with %CALC [13:28]
GuilainCmy idea, was to transform formfield storage date form this format DD/MM/AAAA to an acceptable format for comparison like AAAA/MM/DD [13:28]
jastthe part queried from each topic... nope [13:28]
GuilainCwith this kind of piece of code %CALC{$FORMATTIME($TIME($SUBSTRING(14/07/1992,7,4)/$SUBSTRING(14/07/1992,4,2)/$SUBSTRING(14/07/1992,1,2)),$day $mon $year)}%
oki thanks jast & gac410 for your help
I forget date comparison, for the moment, and add a status for my search ;)
badly development, but enough efficient :)
[13:28]
CDotGuilainC: do you have a locale defined? [13:38]
GuilainCeuh... what does it mean ? in configure ?
the question is what in data form, date is in french localisation ? DD/MM/AAAA
?
[13:40]
CDotactually, i just checked and it won't make any difference. Only English month names are recognised.
Foswiki::Time::parseDate assumes "Apr" not "Avr"
[13:40]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Time [13:41]
GuilainCit exactly that,
for summary,
for my user, i use JSCalendarPlugin
which define date, in dataform in DD Month YYYY
where Month is french one : Fev instead of Feb ; Apr instead of Avr
knowing that,
[13:42]
CDot... which is localised to France, yes, I understand. However Foswiki::Tine does *not* parse localised dates, only those in English. [13:43]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Tine [13:43]
GuilainCi conclude that d2n() in search query doesn't undertstand french date, so can't make any comparison [13:43]
CDotthat's correct. As I said (twice), only English month names are recognised. [13:44]
GuilainCwell, I know why it doesn't work ;)
precisely, thank for precision
[13:44]
CDotsure. You can use ISO format dates (generally a better idea anyway)
2014-04-04
[13:44]
GuilainCfor computer is a better idea, but for user, is not ;) [13:45]
CDotsure. Please feel free to open a Task to get support for localised month names. [13:45]
GuilainCfor the moment, i've done a new "state" which is the meaning of the time comparison... it's enough
do you think is more a core Task ? or Specific to JSCalendarPlugin
which localise date ?
[13:45]
CDotno, it's a core problem. [13:46]
GuilainCoki, i will take my coffe break in this aftenoon, to open a task, it will be my first one.. [13:47]
CDotthe inadequate code is in Foswiki::Time, which is a core module. [13:47]
GuilainCthanks all for your support. [13:47]
CDotgood idea to paste this discussion into the task, for reference. [13:47]
GuilainCoki
just between you and me ? or from the explanation of my problem ?
[13:49]
CDotfrom your explanation. It helps to have some background to problems. [13:52]
................. (idle for 1h22mn)
***ChanServ sets mode: +o pharvey [15:14]
gac410Anyone around familiar with Centos 5?
Foswiki:Support/Question1429 having trouble with getting "version" updated
[15:17]
FoswikiBothttp://foswiki.org/Support/Question1429 [ perl modules needed ] [15:18]
***ChanServ sets mode: +o pharvey [15:32]
jastno, but that message looks like it's an internal issue in the old version package, probably it deals poorly with functions it doesn't know about yet [15:37]
gac410Ah. Yeah maybe the install still put in an old version. I didn't think of that possibility,. [15:38]
.... (idle for 15mn)
***ChanServ sets mode: +o pharvey [15:53]
..... (idle for 23mn)
ChanServ sets mode: +o pharvey [16:16]
.... (idle for 19mn)
ChanServ sets mode: +o pharvey [16:35]
..... (idle for 23mn)
ChanServ sets mode: +o pharvey [16:58]
GuilainCCDot: done, hope to have well done my first "task" http://foswiki.org/Tasks/Item12812 [17:09]
CDotGuilainC: looks fine. Will be even better when you attach the patch that fixes it :-) [17:10]
GuilainC:) i lovely to done that... but you don't know how badly i code...
suppose that the parse of date, need more sub-tree in treatment...
for each different month spelling
[17:12]
***ChanServ sets mode: +o pharvey [17:19]
CDotGuilainC: no - there should be existing support somewhere in the locale module
just a case of finding it :-)
[17:24]
GuilainCthen developing is more and more easy ;) i will take a look this week-end [17:29]
.... (idle for 16mn)
gac410Started a new page Foswiki:Development.GitConversionTasks Lets try to gather a complete list of all the things we need to touch for a conversion. [17:45]
FoswikiBothttp://foswiki.org/Development.GitConversionTasks [ GitConversionTasks ] [17:45]
***ChanServ sets mode: +o pharvey [17:54]
...... (idle for 27mn)
ChanServ sets mode: +o pharvey [18:21]
........... (idle for 54mn)
gac410Hy all, .. any translators around. Digging through the details of git migration, maybe github. There might be an alternative to Pootle - Weblate.
It has good git & github integration including cross-branch translation and provides a pre-defined github service for pushing changes back into the translation engine.
Might be worth a closer look by the translation teams.
http://weblate.org/
Seems to have similar requirements too. python, django, database backend, ...
[19:15]
.......... (idle for 45mn)
ping gmc - you around [20:06]

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)