IRC chat logs for #ltsp on irc.freenode.net (webchat)


Channel log from 18 May 2017   (all times are UTC)

00:00book`_ has left IRC (book`_!~book`@68.ip-149-56-14.net, Quit: Leaving)
00:03book` has joined IRC (book`!~book`@68.ip-149-56-14.net)
05:19Statler has joined IRC (Statler!~Georg@p4FC1FBBB.dip0.t-ipconnect.de)
05:20ricotz has joined IRC (ricotz!~ricotz@p5B2A9C89.dip0.t-ipconnect.de)
05:20ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
05:37
<alkisg>
bennabiy, sbalneav: hi, I'm trying to clean up a bit the ltsp branches: https://code.launchpad.net/ltsp/+branches. Could you mark as "abandoned" your branches that are no longer in development?
05:38
stgraber, ogra_, I'm marking as abandoned some of your really older branches, if you see something that shouldn't be marked as such, ping me or revert it...
05:39
Essentially we should mark _all_ branches there as abandoned, as we've switched to git
06:12lucascastro has left IRC (lucascastro!~lucas@186.227.186.18, Remote host closed the connection)
06:12lucascastro has joined IRC (lucascastro!~lucas@186.227.186.18)
06:15
<alkisg>
!nbd-client
06:15
<ltsp>
nbd-client: To try mounting the NBD image from the client initramfs: nbd-client 192.168.67.1 -N /opt/ltsp/i386 /dev/nbd0
06:15
<alkisg>
!client-list
06:15
<ltsp>
client-list: to get a list of all nbd-clients (which sometimes is the same as ltsp clients), run: netstat -tn | sed -n 's/.*:10809 *\([0-9.]*\):.*/\1/p' | sort -Vu
06:32mikkel has joined IRC (mikkel!~mikkel@mail.dlvs.dk)
06:54ZAJDAN has joined IRC (ZAJDAN!4d30954b@gateway/web/freenode/ip.77.48.149.75)
06:54johnsmith has joined IRC (johnsmith!7ab0ed24@gateway/web/freenode/ip.122.176.237.36)
07:21wim1 has joined IRC (wim1!~Thunderbi@WEGC203029.UNI-GRAZ.AT)
07:29wim1 has left IRC (wim1!~Thunderbi@WEGC203029.UNI-GRAZ.AT, Quit: wim1)
07:36Statler has left IRC (Statler!~Georg@p4FC1FBBB.dip0.t-ipconnect.de, Remote host closed the connection)
07:41wim1 has joined IRC (wim1!~Thunderbi@WEGC203033.UNI-GRAZ.AT)
08:15Statler has joined IRC (Statler!~Georg@mail.lohn24.de)
08:23wim1 has left IRC (wim1!~Thunderbi@WEGC203033.UNI-GRAZ.AT, Quit: wim1)
08:41adrianorg has joined IRC (adrianorg!~adrianorg@187.115.104.129)
08:41
<ogra_>
alkisg, fine with me (they are not going away anyway, just getting tagged dead)
08:41
<alkisg>
Right
08:42
Cool, I marked as abandoned a whole lot of them, and only kept the ones from the more active devs currently, and I'll ping them to review those themselves
08:44adrianor1 has left IRC (adrianor1!~adrianorg@179.177.209.49.dynamic.adsl.gvt.net.br, Ping timeout: 240 seconds)
08:44* ogra_ saw the mail spam ;) good work
08:55ogra_ has left IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 240 seconds)
08:55ogra_ has joined IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de)
09:18johnsmith1 has joined IRC (johnsmith1!~Adium@122.176.237.36)
09:18ZAJDAN has left IRC (ZAJDAN!4d30954b@gateway/web/freenode/ip.77.48.149.75, Ping timeout: 260 seconds)
09:28johnsmith1 has left IRC (johnsmith1!~Adium@122.176.237.36, Quit: Leaving.)
09:31johnsmith has left IRC (johnsmith!7ab0ed24@gateway/web/freenode/ip.122.176.237.36, Ping timeout: 260 seconds)
09:52wim1 has joined IRC (wim1!~Thunderbi@WEGC33.UNI-GRAZ.AT)
10:03markus_e92 has left IRC (markus_e92!~markus_e9@80-121-123-207.adsl.highway.telekom.at, Ping timeout: 246 seconds)
10:07markus_e92 has joined IRC (markus_e92!~markus_e9@62-46-100-243.adsl.highway.telekom.at)
10:56markus_e92 has left IRC (markus_e92!~markus_e9@62-46-100-243.adsl.highway.telekom.at, Ping timeout: 240 seconds)
10:57markus_e92 has joined IRC (markus_e92!~markus_e9@62-46-100-243.adsl.highway.telekom.at)
12:53johnsmith has joined IRC (johnsmith!~Adium@122.176.237.36)
12:55mikkel has left IRC (mikkel!~mikkel@mail.dlvs.dk, Quit: Leaving)
13:26
<sbalneav>
alkisg: k, will do today
13:37donkey_ has joined IRC (donkey_!614ea0a5@gateway/web/freenode/ip.97.78.160.165)
13:41
<donkey_>
im having an issue authenticating from within the terminal
13:42
i need to install a dpkg but when i try to sudo it keeps saying that the password is incorrect
14:07
<alkisg>
donkey_: are you doing that on a client?
14:07
Why would you install something on the client? It would be lost on reboot...
14:07
Install packages on the server, or in the chroot...
14:09lucascastro has left IRC (lucascastro!~lucas@186.227.186.18, Remote host closed the connection)
14:09lucascastro has joined IRC (lucascastro!~lucas@186.227.186.18)
14:10
<donkey_>
it has to be installed via gui
14:10
which is really dumb
14:11
i suppose i could add a gui to the server
14:11
<alkisg>
Or run ssh -X from a client to the server
14:11
If you install it to the client, it will be lost on reboot
14:11
Or use ltsp-pnp
14:11
!ltsp-pnp
14:11
<ltsp>
ltsp-pnp: ltsp-pnp is an alternative (upstream) method to maintain LTSP installations for thin and fat clients that doesn't involve chroots: https://help.ubuntu.com/community/UbuntuLTSP/ltsp-pnp
14:13
<donkey_>
as an alternative
14:13
all i really need is a pdf viewer that supports editable forms and buttons
14:13
<alkisg>
Install the acrobat reader package in the chroot
14:14
<donkey_>
it installs version 9 which doesnt support the forms
14:14
did that already
14:14
<alkisg>
And which one does?
14:14
What are you trying to install?
14:14
<donkey_>
foxit requires the gui
14:14
so that is out
14:14
<alkisg>
foxit has a linux version?
14:15
<donkey_>
yea
14:15
but you have to install it via a script that loads some dumb gui installer
14:17
various web browser extensions have failed to work also
14:39
<alkisg>
donkey_: foxit installer is per user, based on wine?
14:40
Or does it need sudo access?
14:40
I don't see a linux download in their site...
14:41
Actually... https://www.foxitsoftware.com/products/pdf-reader/comparison.php
14:41
says linux doesn't support form
14:41
Acrobat 9 does support forms, maybe your form is more advanced
14:46lucascastro has left IRC (lucascastro!~lucas@186.227.186.18, Read error: Connection reset by peer)
14:47ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
14:47lucascastro has joined IRC (lucascastro!~lucas@186.227.186.18)
14:55
<donkey_>
yea when i tried acrobat reader the form didnt include a button needed to submit the data to a remote site
14:56
i'll try evince or okular and see what happens
14:58wim1 has left IRC (wim1!~Thunderbi@WEGC33.UNI-GRAZ.AT, Quit: wim1)
15:00
<bennabiy>
alkisg: As far as any there from me, you can mark them all. If I need to do current work, I will start a new one. As far as I know I do not have any pending merge reviews etc...
15:09
<alkisg>
bennabiy: I left them in case you need to delete them, instead of mark them as abandoned
15:09
I can't delete branches that belong to others, and some of them might need deletion...
15:09
<bennabiy>
alkisg: I will delete
15:10
<alkisg>
I couldn't rely on other devs cleaning up theirs, as they've been away for too long, but of us, let's each one do his own
15:10
Cool
15:10
<bennabiy>
alkisg: I just need to remember how ;) been a long time sime I worked with bzr, or would I do it through launchpad directly?
15:10
<alkisg>
Launchpad directly
15:10
<bennabiy>
I think there is only 1 I need to keep for ppa sake
15:11
but I will look at it
15:11
<alkisg>
ppas and recipes will need git now
15:11
You can't mix bzr and git recipes
15:11
donkey_: anyways, finding an appropriate pdf reader isn't an ltsp task, but a generic linux task; you might be able to get some help on your #distro channel, or you might even need a windows box for that...
15:12
Once you find one, if you're not able to install it to a chroot, then it's somewhat an ltsp task
15:12
<donkey_>
my question wasnt necessarily the pdf reader
15:12
but why when i try to sudo in the terminal it rejects my password
15:13
<alkisg>
!sudo
15:13
<ltsp>
I do not know about 'sudo', but I do know about these similar topics: 'sudoers', 'fat-sudo'
15:13
<alkisg>
!sudoers
15:13
<ltsp>
sudoers: Not recommented for security reasons: RCFILE_01="echo USER ALL=NOPASSWD: /path/to/program >> /etc/sudoers". USER and /path/to/program can also be ALL.
15:13
<alkisg>
!fat-sudo
15:13
<ltsp>
fat-sudo: to allow members of the sudo group to execute "sudo" in fat clients without a password prompt, put this in lts.conf: RCFILE_01="echo '%sudo ALL=NOPASSWD: ALL' >> /etc/sudoers"
15:13
<donkey_>
i cant perform any admin tasks from within the terminal
15:13
<alkisg>
!LDM_PASSWORD_HASH
15:13
<ltsp>
LDM_PASSWORD_HASH: LDM_PASSWORD_HASH=True in lts.conf saves the password hash to /etc/shadow on login, so that the users can unlock the screensaver etc. If they happen to change their password though, that only takes effect until logout.
15:13
<alkisg>
With those, you can use sudo
15:13
And then after you use sudo on the client, you'll realize it was not what you're looking for :)
15:13
As all changes happen in RAM, and are lost on reboot
15:14
That's why we keep it "disabled" by default, to protect the users from doing something that doesn't make much sense
15:42
You maintain your linux installation on the server, not on the clients
15:58
bennabiy: you go there: https://code.launchpad.net/ltsp/+branches and click on the branches you no longer want, and then delete...
15:59
<bennabiy>
alkisg: in general, the password was to get screen unlocking working, but I guess others use it for things that don't make sense
16:00
<alkisg>
Yeah, although screensavers should detect when there's no password defined, and not prompt for one
16:00
gnome-screensaver does that; mate ,not
16:03
<bennabiy>
yes
16:06
deleted
16:13
So how will recipes work with git repo now?
16:17kjackal_ has joined IRC (kjackal_!~quassel@ppp-94-66-56-119.home.otenet.gr)
16:23
<alkisg>
bennabiy: there are 2 kinds of recipes, one for bzr and one for git branches
16:23
You just need to create git recipes instead of bzr recipes
16:24
<bennabiy>
great
16:24
I do plan on doing more with ltsp as soon as some of my other projects get to a stable place
16:24
Unless a bug lands on my front door and demands my attention...
16:25
<alkisg>
bennabiy: see an example recipe that I've created today: https://code.launchpad.net/~alkisg/+recipe/ltsp+debian-packaging
16:26
<bennabiy>
where is documentation on the nest-part portion? why debian debian etc?
16:35lucascastro has left IRC (lucascastro!~lucas@186.227.186.18, Remote host closed the connection)
16:35lucascastro has joined IRC (lucascastro!~lucas@186.227.186.18)
16:48johnsmith has left IRC (johnsmith!~Adium@122.176.237.36, Quit: Leaving.)
16:56lucascastro has left IRC (lucascastro!~lucas@186.227.186.18, Remote host closed the connection)
16:56lucascastro has joined IRC (lucascastro!~lucas@186.227.186.18)
17:02
<alkisg>
bennabiy: google launchpad recipe documentation, first hit there :)
17:02
<bennabiy>
alkisg: thanks
17:03
alkisg: did you ever commit the fix for https://bugs.launchpad.net/ltsp/+bug/1610304
17:04
<alkisg>
bennabiy: no I didn't, I think someone had reported a negative side on it
17:04
<bennabiy>
really? It is not linked to that bug.
17:04
<alkisg>
Yes he only discussed it here in irc, but I don't remember exactly
17:05
Are you using that patch? Does it work for you?
17:05
<bennabiy>
I need to test it and get back to you
17:05
I just found it
17:05
I have run into that issue as well
17:05
in a couple labs (we use LDAP)
17:06
but since most of our clients are thin, it is not an issue on a day to day basis, but will be once I roll out fat clients as replacements
17:07
<alkisg>
OK, I made a comment to request feedback, and I'll commit it once you verify it's ok
17:07
<bennabiy>
alkisg: if you can find the negative, I can look into it as well
17:08
<alkisg>
Nah I don't think I remember enough to locate it in the logs...
17:08
<bennabiy>
alkisg: ok
17:09
alkisg: I will try to test it soon and get back to you (of course, it is not a show stopper for you if it takes a while ;)
17:09
<alkisg>
Nice
17:54vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
17:56
<vagrantc>
alkisg: thanks for all the launchpad housecleaning
18:16drymis has joined IRC (drymis!615ab632@gateway/web/freenode/ip.97.90.182.50)
18:17
<alkisg>
vagrantc: now it's time to set up the translations!
18:17
So, some possible solutions are...
18:18
1) manual uploads of our .pot to launchpad => nah
18:19
2) bzr mirrors of our branches for automatic import. Then, (2a) manual export of .po as .tar.gz when we want them, or (2b) automatic export to a third bzr branch on launchpad. After the export, we'd copy the .po to our main git branch when we want
18:20
3) to maintain only a second bzr branch, that only has the .pot/.po files inside it. Launchpad would then automatically import/export to that one, like it did in the ltsp branch in the past. And again when we want we'd copy the .po to our main git branch
18:20
Which one do you prefer? I'm between (2b) and (3)...
18:24
vagrantc: and another thing to discuss, is maintaining your ltsp+debian packaging directly on launchpad; if you had a branch there "disabled-upstreamed-patches" with the upstreamed patches removed from debian/patches/series, we could use it for daily builds
18:24
So you'd have "dmaster", "jessie-backports", and "disabled-upstreamed-patches" or however else you'd like to call that; that would only have 1 file changed from dmaster
18:27
<vagrantc>
alkisg: i'm inclined to 2b, presuming i can then git merge --squash the updates
18:27
<alkisg>
(currently I'm maintaining a copy with just that...)
18:27
<vagrantc>
alkisg: the difference could be as simple as "rm debian/patches/series"
18:27
alkisg: you could actually leave the patches there...
18:27
<alkisg>
Sure, although some of the patches are nice to have
18:28
<vagrantc>
alkisg: ah, some of the patches i've never pushed upstream?
18:28
<alkisg>
Yes
18:28
It would be nice if you actually pushed them upstream though :)
18:28
<vagrantc>
well, maybe we should consider if they're possible to push
18:28
i've been a bit cautious with some of those
18:28
but some are probably appropriate
18:28
<alkisg>
Also, it would be nice to have patches/upstream dir in ldm too, like you do in ltsp
18:30
vagrantc: is my idea of maintaining a "disabled-upstreamed-patches" branch feasible, or it results in merge conflicts?
18:39kjackal_ has left IRC (kjackal_!~quassel@ppp-94-66-56-119.home.otenet.gr, Ping timeout: 240 seconds)
18:41Statler has left IRC (Statler!~Georg@mail.lohn24.de, Remote host closed the connection)
18:52
<vagrantc>
alkisg: it's prone to lots of pointless merge conflicts
18:52
<alkisg>
OK
18:52Parker955_Away has left IRC (Parker955_Away!~parker@2607:5300:60:5f16::ac59:f0a2, Ping timeout: 246 seconds)
18:52
<vagrantc>
alkisg: if it's a branch that you are ok with rebasing all the time, should be fine, though
18:52
<alkisg>
vagrantc: to minimize the changes needed for translations etc, I'll (1) rename the old ltsp-trunk branch to ltsp-trunk-old, and then (2) setup ltsp-trunk as a bzr mirror of the git one
18:52* vagrantc would name it ltsp-translations
18:53
<vagrantc>
to make it clear
18:53
<alkisg>
That would require me changing all the series though
18:53
I can do that; hopefully I won't do any mistakes in launchpad there...
18:53
Series are what define translatable projects
18:53
<vagrantc>
we seem to be having two concurrent conversations and i'm not sure what's responding to what :)
18:54
the whole -trunk thing has always been a bother
18:54
<alkisg>
Eh, you took 20 minutes to respond so I moved on to translations :D
18:55
So, this is the "series": https://launchpad.net/ltsp/ltsp-trunk
18:55
<vagrantc>
alkisg: in general, i've made a habit of putting patches pulled from upstream into debian/patches/upstream/, if i didn't do that, it's a mistake on my part
18:55
<alkisg>
Yeah in ldm the paches are in the multiseat/ dir
18:56
I'll rename that series to ltsp-translations then. Or do you want that to be plain ltsp?
18:57* vagrantc should've put them into debian/patches/upstream/multiseat/ then
18:57
<vagrantc>
alkisg: i'd prefer "ltsp-translations" as "ltsp" might make people thing it's something to manually push to
18:57
or pull from
18:57
<alkisg>
Series doesn't contain code, it's not pushable
18:57
The branch will be ltsp-translations, sure
18:57
<vagrantc>
i don't understand what series is really, in this context
18:58
<alkisg>
Go there: https://launchpad.net/ltsp/ltsp-trunk/+edit
18:58
<vagrantc>
i mean, i followed the launchpad url, but it doesn't really explain much to me :)
18:58
<alkisg>
Well, when we want to declare that we have a translation project, we declare a series
18:58
I'm not sure what else a series manages, other than translations and milestones
18:59
We can also use series milestones in bug reports...
18:59
So we'll have 3 branches. ltsp, ltsp-bzr-mirror, and ltsp-exported-translations
19:00
How do you like to name the last 2?
19:00
<vagrantc>
ltsp is git, ltsp-bzr-mirror is bzr, and ltsp-exported-translations is bzr?
19:00
<alkisg>
Yes
19:00
<vagrantc>
can we then mirror ltsp-exported-translations to git? :)
19:01
<alkisg>
Not sure... I think only <other> => bzr is supported, not <other> => git
19:01
Let me check
19:01
<vagrantc>
ok, it's easy enough to handle with git-remote-bzr
19:02
but if launchpad can make it just happen, even better :)
19:03
<alkisg>
OK, what about the names?
19:03
ltsp-bzr-mirror? ltsp-translations? ltsp-exported-translations?
19:03
(same for ldm and ltspfs, of course...)
19:03
It can be named just ltsp
19:03
It's not pushable since it's a mirror
19:04
So we can have ltsp (git), ltsp (bzr mirror), ltsp-translations (exported from rozetta)
19:04lucascastro has left IRC (lucascastro!~lucas@186.227.186.18, Read error: Connection reset by peer)
19:07
<vagrantc>
alkisg: that sounds good
19:07
the mirror polls periodically?
19:07
<alkisg>
OK, I'm starting by renaming the series from ltsp-trunk to plain ltsp etc
19:07
Yes, every 5 hours or so
19:13
https://code.launchpad.net/~ltsp-upstream/ltsp/ltsp
19:14
"This is a bzr mirror of the ltsp git branch, to automatically import the translations to launchpad."
19:15
Series: https://launchpad.net/ltsp/ltsp
19:16
Translations: https://translations.launchpad.net/ltsp/ltsp/+translations
19:21
Exported translations: https://code.launchpad.net/~ltsp-upstream/ltsp/ltsp-translations
19:24
<vagrantc>
huh, now code.launchpad.net/ltsp displays differently
19:28
<alkisg>
vagrantc: differently, how? only ltsp-trunk series was renamed to ltsp there...
19:28
I think it's ok; I'll do the same with the other series too
19:28
Although we're not using ltspfs translations, I'll just rename the series
19:28
<vagrantc>
alkisg: now it just shows the branches for ltsp, and links to the various other branches
19:29
alkisg: before it showed all the "other branches"
19:29
<alkisg>
vagrantc: maybe you were looking at https://code.launchpad.net/ltsp ?
19:29
<vagrantc>
yes, that's exactly what i'm looking at that's changed
19:30
not a big deal
19:30
but before ltsp, ldm, ltspfs, etc. were all projects of the ltsp project
19:31
<alkisg>
Ah you mean since we switched to git?
19:31
<vagrantc>
right
19:31
<alkisg>
OK you got me confused, I thought we were talking about the series and translations
19:31
<vagrantc>
well, since i switched to git ... yesterday? ... and now today it looks different
19:31
sorry
19:31
<alkisg>
Yes when you set the default to git yesterday, that's the new look
19:31
And you need to click on the bazaar branches to see the other ones
19:32
https://code.launchpad.net/ltsp/+branches
19:32
<vagrantc>
there's no correlary to status=mature|abandoned|etc. for git repositories
19:32
<alkisg>
The interface for git is more bare, yeah
19:32
Also, if you want, delete your pending review at https://code.launchpad.net/ltsp/+activereviews, I don't have access to that one
19:33
...and delete/mark as abandoned your branches that you no longer need at https://code.launchpad.net/ltsp/+branches
19:41
<vagrantc>
abandoned most branches
19:47
<alkisg>
vagrantc: I've set up the translation settings so that rosetta only reads our .pot files and not our .po files
19:48
I hope that way it won't get confused and do multiple commits
19:48
The down side is that we shouldn't manually merge ltsp/ldm translations anymore, we should direct people to launchpad
19:48
(I don't think anyone has manually sent translations in the last years, right?)
19:53
vagrantc: go to https://launchpad.net/ltsp and check the series
19:53
ltspfs is clean, the other two have milestones and version etc
19:54
Noone updates the milestones and versions
19:54
Should I delete them so that they look like ltspfs?
19:55JerryT has left IRC (JerryT!~jerry@wsip-70-165-106-163.om.om.cox.net, Ping timeout: 272 seconds)
19:56
<vagrantc>
we *can't* commit translations manually now?
19:56
since we haven't been using milestones or versions (beyond tagging versions), i don't see any reason to keep them
19:57
<alkisg>
vagrantc: well, we can, but I told launchpad not to import them automatically
19:57
I think that it was pushing e.g. a po file, and then in the next import it was seeing that it was modified,
19:57
<vagrantc>
alkisg: multiple commits should be a non-issue now, as we can do git merge --squash and only end up with one commit in git.
19:57
<alkisg>
so then it was pushing it again etc
19:58
<vagrantc>
extra busywork for launchpad for no real reason, i guess.
19:58
<alkisg>
Do we want that extra headache? Did you receive any translations by mail in the last years?... yeah, that
19:58
<vagrantc>
i think i've gotten one or two
19:58
but maybe that was longer ago
19:59
<alkisg>
Ah btw I've seen that in the git tree you've deleted an ar.po that was in the bzr tree
19:59
Was that done by mistake?
19:59
<vagrantc>
might have been... dunno
20:00
alkisg: which projecT?
20:00
alkisg: i see an ar.po in my git checkout
20:00
of ltsp
20:00
<alkisg>
I may not remember the name correctly, let me find it...
20:01
Ah maybe that was in the ltsp+debian-packaging...
20:01
<vagrantc>
ldm doesn't have ar.po
20:01
<alkisg>
am.po, in ltsp-debian-packaging
20:02
<vagrantc>
huh
20:02
not sure what happened there.
20:03
<alkisg>
It's the complexity of having the code alongside with the packaging :P :D
20:03
<vagrantc>
alkisg: am.po was introduced after 5.5.9
20:04
<alkisg>
Ah, ok, all is fine then
20:04
Although you don't have translations freeze in stretch yet, right?
20:05
Anyways. So I think I've updated everything wrt translations and series and imports; if you see something wrong do ping
20:05
<vagrantc>
cool!
20:05
alkisg: so, i guess it's not clear to me what happens when the histories diverge
20:05
alkisg: what's the workflow for importing new translations?
20:06
<alkisg>
I've updated 4 greek strings in ltsp translations as a test case
20:06
I export launchpad to export them to ltsp-translations
20:06
Then I'll pull locally and merge them to git and push them
20:06
<vagrantc>
and then we'll see if anything explodes? :)
20:06
<alkisg>
Personally I wouldn't bother with the history, I would just use cp *.po
20:07
<vagrantc>
that's where we're differing
20:07
<alkisg>
Do you really insist on trying to preserve the worthless rozetta commits?
20:07
"launchpad translations updates"?
20:07
<vagrantc>
not the full commits, but i think it's worth noting them as squashed commits, maybe.
20:08
<alkisg>
Of course I"ll have a commit message
20:08
"Importing the latest translations from launchpad"
20:08
But I don't think the bzr tree has any more info than that
20:08
Rozetta has more info internally, who translated what
20:08
But it's not exposed in bzr trees
20:09
<vagrantc>
we had some sort of system for extracting the translation-credits or something... will that get lost?
20:09
you'll also need to copy the .pot files
20:09
<alkisg>
I don't know what that was
20:09
We update the .pot files at "string freezes" directly in the git tree
20:09
Launchpad isn't involved at that step
20:09
<vagrantc>
there's a translator_credits fuction
20:10
<alkisg>
Then we git push, then launchpad mirrors to bzr, and it then shows them to translators
20:10
<vagrantc>
it appears to be in the .po files as a translation string hack
20:11
alkisg: yeah, i get that
20:11
translator-credits
20:11
and jammcq gets credit for everything, somehow :)
20:11
<alkisg>
Eh, are translators supposed to translate "translator-credits" themselves, or is launchpad supposed to do that automatically?
20:11* vagrantc shrugs
20:11
<vagrantc>
i *thought* launchpad added those
20:12
<alkisg>
msgid "translator-credits" msgstr ""
20:12
"Launchpad Contributions:\n" " Alkis Georgopoulos https://launchpad.net/~alkisg\n" " Fotis Tsamis https://launchpad.net/~ftsamis\n" " Jim McQuillan https://launchpad.net/~jam-mcquil"
20:12
So yes, that's the case
20:12JerryT has joined IRC (JerryT!~jerry@wsip-70-165-106-163.om.om.cox.net)
20:12
<alkisg>
That will still work now
20:12
The person that imported the .po to launchpad gets credits; it's a launchpad "feature"
20:13
We'd need to remove jammcq somehow from there
20:13
<vagrantc>
i'm a little nervous if rosetta's view and git's view of the .po files diverge somehow
20:13
<alkisg>
We write .pot to git
20:13
We pull .po from rozetta and write them to git
20:13
Everything's one way, so it's synced fine
20:14
We don't have any two-way synching there...
20:14
<vagrantc>
the old way we had rosetta updating the .pot as well or soemthing?
20:14
<alkisg>
Rozetta doesn't read .po files at all
20:14
No, the old way had rozetta read the .po files too
20:14
And try to merge the online translations with the ones from the .po files
20:14
<vagrantc>
well, let's try this and see how it goes
20:15
<alkisg>
That's error prone; it might also have caused the jammcq issue :)
20:15
<vagrantc>
how can we manually fix the jammcq issue? :)
20:16
<alkisg>
Let's see how it goes first, and we'll worry about that later
20:16
<vagrantc>
well, at the time we started using rozetta, i was getting manually submitted tranlations
20:16
<alkisg>
I'm a bit worried about how launchpad will sync from the bzr mirror to the exported translations tree
20:17
I hope there won't be any conflicts there
20:18
So now, I'll keep an eye on https://code.launchpad.net/~ltsp-upstream/ltsp/ltsp-translations, I hope I'll see the new greek translations committed in a few hours/days...
20:19
Ideally, rozetta will do a push --overwrite there, i.e. it will only have a single commit after the bzr mirror head
20:19
<vagrantc>
it *is* spelled rosetta. heh.
20:20* vagrantc was thinking it had everything to do with the rosetta stone ...
20:21
<vagrantc>
alkisg: yes, this is one case where overwrite makes the most sense
20:21
well, i would guess it might have multiple commits
20:21
if translations were made on different days and no other changes
20:22
unless it always bases the changes on the latest mirrored branch ... then i guess it'd only be one commit.
20:22* vagrantc still thinks git merge --squash will be easier than cp *
20:22
<vagrantc>
unless it's broken
20:23
<alkisg>
git merge from bzr source?
20:24
<vagrantc>
that's how i've committed to ltsp for the last several years
20:24
i never commit with bzr
20:24
git-remote-bzr is used as a git backend that uses bzr repositories as a remote
20:24
<alkisg>
I mean, isn't there an extra command needed to convert first?
20:25
Ah, it's a backend, not a command
20:25
<vagrantc>
handled essentially transparently
20:25
<alkisg>
If you give me the command, I don't mind doing it that way, although cp sounds a much easier way to merge that single commit where we don't even want to keep the commit message or the author...
20:26
It sounds to me like an overcomplicated method to copy files and ignore the commit message and author
20:26
<vagrantc>
will burn that bridge after we've crossed it
20:26
<alkisg>
So, the translations are exported daily; we can test all that tomorrow
20:27
<vagrantc>
i find using tools designed to merge changes far less error prone than commandline globbing myself
20:27
alkisg: i suspect the end result *should* be virtually identical, so it shouldn't matter too much which way
20:28
alkisg: in fact, you can stick to cp and i can stick to using git merge
20:28
should be compatible
20:28
<alkisg>
np from me; if it's just a couple of commands I can easily put them in my how-to
20:28
Btw, there's a one-time import functionality: https://translations.launchpad.net/ltsp/ldm/+request-bzr-import
20:29
...we can use it if we ever want to sync .po files (not .pot) again
20:29
i.e. manually received translations
20:29
<vagrantc>
great, that solves my only worry
20:29
just a matter of remembering how to do it
20:29
<alkisg>
At that point we'd need to push them to git, and invoke the one-time import
20:29
<vagrantc>
push to git, wait for bzr to sync, and then request?
20:30
<alkisg>
Right
20:30* alkisg missed the extra step there :)
20:30
<alkisg>
Oh well I hope we'll remember all that infrastructure tomorrow :D
20:30
For now, ltsp in launchpad is clean! Yey!
20:30
<vagrantc>
and hopefully before too long, it'll support translations from git natively and we won't have to remember much
20:31
<alkisg>
Let's hope so
20:32
Or if github cooperates with some opensource translations site, in a very seamless way, we can switch there :D
20:33
vagrantc: in a few days I'll start committing some things that are in sch-scripts now, to ltsp
20:33
(as part of the internationalization effort)
20:33
<vagrantc>
yay
20:34
<alkisg>
sch-scripts has code (1) for greeks => it'll remain in a small sch-scripts project, (2) for all => it'll become ltsp manager, (3) and some code that really belongs to ltsp and not to ltsp-manager
20:34
So I'll start with (3) first...
20:35
OK that's all for today, /me waves goodnight :)
20:39
<vagrantc>
alkisg: thanks!
20:40
since ltspfs contains no translations, we can probably kill the bzr mirror
20:47
alkisg: well, seems like merging is a no-go
20:48
alkisg: somewhere in the translation from bzr to git back to bzr and exported to bzr, and then imported to git again ... the history got messed up so merging won't be possible really
20:48
alkisg: at least, not easily
20:49
although git diff ..translations/master | patch -p1
20:49
git commit -a
20:49
might basically work
20:50
though you might need the occasional git add mixed in
20:56
<bennabiy>
what is it that prevents LTSP from being hosted on github?
20:56
I prefer it to launchpad
20:57
or gitlab or the like
20:59ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
21:01
<vagrantc>
bennabiy: personally, there's enough controversy over the terms of service on github as it relates to GPL code
21:02
bennabiy: and various other license incompatibilities
21:02
<bennabiy>
I wondered if that was the case
21:02
<vagrantc>
bennabiy: licensing aside... we're already using launchpad for various services and it would be less of a transition
21:03ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
21:04
<vagrantc>
and while the current state of translation support isn't great with git on launchpad, i think it will get better over time
21:04
i don't think bzr is sticking around, and git is the obvious alternative for launchpad to support
21:04
<bennabiy>
git is to bzr what zfs is to btrfs
21:05
<vagrantc>
21:06
heh
21:06* bennabiy missed vagrantc's comment in the snowstorm
21:06* vagrantc suggests not cleaning a keyboard while using irc
21:06
<bennabiy>
heh
21:07
<vagrantc>
though that came out impressively uneventful, considering
21:07
<bennabiy>
true, these are the days we live in
22:19Freejack has left IRC (Freejack!~quassel@unaffiliated/freejack, Read error: Connection reset by peer)
22:25Freejack has joined IRC (Freejack!~quassel@unaffiliated/freejack)
22:46adrianorg has left IRC (adrianorg!~adrianorg@187.115.104.129, Ping timeout: 246 seconds)
23:38vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)