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


Channel log from 4 June 2012   (all times are UTC)

00:09awilliams has left IRC (awilliams!mistik1@unaffiliated/mistik1, Ping timeout: 245 seconds)
00:41vagrantc has joined IRC (vagrantc!~vagrant@c-76-115-60-19.hsd1.or.comcast.net)
00:41vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
00:43Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 250 seconds)
00:44shogunx has left IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com, Read error: No route to host)
00:48shogunx has joined IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com)
00:48* vagrantc returns to the confusing state of LDM_*DESKTOP and LDM_*SESSION
00:48awilliams has joined IRC (awilliams!mistik1@unaffiliated/mistik1)
00:57Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
01:21andygraybeal has joined IRC (andygraybeal!~andy@h167.93.213.151.dynamic.ip.windstream.net)
01:21Damianos has left IRC (Damianos!~Damianos@68-186-144-58.dhcp.kgpt.tn.charter.com, Quit: Damianos)
01:56shogunx has left IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com, Read error: No route to host)
02:01shogunx has joined IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com)
02:17stgraber has left IRC (stgraber!~stgraber@ubuntu/member/stgraber, Quit: kernel update)
02:20stgraber has joined IRC (stgraber!~stgraber@ubuntu/member/stgraber)
02:27* vagrantc schemes to implement LDM_SELECTED_SESSION
02:27
<vagrantc>
this would solve a lot of confusion.
02:28
because LDM_SESSION is used for two purposes ... setting an environment variable from lts.conf, and also being passed from the greeter as the selected session... and the dmrc processing can't tell the difference
02:29
or we could just deprecate LDM_SESSION.
02:29adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
02:33adrianorg_ has left IRC (adrianorg_!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Ping timeout: 252 seconds)
02:35F-GT has joined IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net)
02:37adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Ping timeout: 245 seconds)
03:01Parker955 is now known as Parker955_Away
03:08vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 245 seconds)
03:08andygraybeal has left IRC (andygraybeal!~andy@h167.93.213.151.dynamic.ip.windstream.net, Ping timeout: 246 seconds)
03:29F-GT has left IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net, Ping timeout: 252 seconds)
03:55sha__ has joined IRC (sha__!~sha@e177118226.adsl.alicedsl.de)
03:58sha has left IRC (sha!~sha@g231161112.adsl.alicedsl.de, Ping timeout: 245 seconds)
04:05Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
05:24bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
05:43alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
05:43shogunx has left IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com, Ping timeout: 252 seconds)
05:51
<vagrantc>
alkisg: ok, so ...
05:51
<alkisg>
Heya
05:52
<vagrantc>
alkisg: so my last patch pretty much gets the LDM_FORCE_SESSION and LDM_FORCE_DESKTOP working ... LDM_DESKTOP doesn't override LDM_SESSION, though, because there's no way to detect if the LDM_SESSION value is from selecting it in the menu,.
05:52* alkisg would vote for LDM_SESSION to take the meaning of LDM_DESKTOP, and to keep only those 2, LDM_SESSION and LDM_FORCE_SESSION, and a 3rd one passed by LDM, LDM_SELECTED_SESSION.
05:53
<alkisg>
It might cause some lts.conf upgrade problems, but I think it'll cause less confusion in the long term
05:53* vagrantc doesnt like breaking backwards compatibility
05:54
<alkisg>
02:29or we could just deprecate LDM_SESSION. ==> doesn't that break backwards compatibility anyway?
05:54
<vagrantc>
well, thenn it's a planned breakage, rather than a hidden change of purpose
05:56
<alkisg>
We could have a period where we support the old syntax too, i.e. test if LDM_SESSION refers to an Exec line instead of a .desktop file name
05:56bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
05:56shogunx has joined IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com)
05:57
<alkisg>
Overriding .dmrc but not forcing the session is kinda confusing on its own, as it is now
05:57
<vagrantc>
i like the LDM_DESKTOP better anyways
05:57bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
05:57
<alkisg>
cat .dmrc => Session=xxx, I think people will understand that more easily (or ls /usr/share/xsessions, always "session" in the names there)
05:59
I.e. let's just state that "LDM_SESSION may be a .desktop file name, or if not, a whole session command line to run", to support both meanings, but also that "the .desktop file name syntax is preferred"
05:59
...because it allows for checking if the session is actually available
06:00
We can keep it this way for a couple of years until people upgrade their lts.conf on their own timing
06:01
<vagrantc>
ok, that works too, i guess.
06:01
alkisg: so, check if it matches a session name?
06:02
<alkisg>
First check if it matches a session name, then if it doesn't contain spaces, check if `which $LDM_SESSION` exists, if not fall back to the default session, and if it does contain spaces, then don't check if it exists, just run it
06:03
While for LDM_FORCE_SESSION we can require that it's a .desktop name, it's a new variable
06:03
And, we can change the greeter to pass LDM_SELECTED_SESSION to the environment
06:04
<vagrantc>
so, if someone wants a custom session type, they need to define it on the server, too?
06:04
<alkisg>
Right
06:04
<vagrantc>
they need to actually set up a desktop file?
06:04
<alkisg>
Right
06:05
We could also implement an LDM_CUSTOM_SESSION, but do we want to?
06:06
<vagrantc>
hrm.
06:06
<alkisg>
I think nx and x2go support it, but they don't have ldminfod to help them
06:06
So they don't allow custom .desktop files, that's why they support it
06:08
The good thing with using a desktop file is that one can always have a menu item for it, (with the Name: from the desktop file), translated and all...
06:09
And it'll keep things cleaner in the future when we start supporting other DMs
06:14
<vagrantc>
so, i tried patching ldm to send only the selected session, but failed.
06:15
<alkisg>
vagrantc: I can try it later on today (gotta leave in an hour or so)
06:16
<vagrantc>
otherwise, desktop selection is basically working right
06:16
but sounds like you want to change it up a bit
06:17
<alkisg>
Yeah, I don't like the "_DESKTOP" naming, as what I'm seeing in the DM reads "Session"... and using the whole command line for session got me confused the first time I read the .dmrc handling
06:19ricotz has joined IRC (ricotz!~rico@p5B2AC3ED.dip.t-dialin.net)
06:19ricotz has joined IRC (ricotz!~rico@unaffiliated/ricotz)
06:20
<vagrantc>
seems like distinguishing LDM_SELECTED_SESSION from LDM_SESSION would really help a lot.
06:21
that way, the *only* time we write .dmrc is with LDM_SELECTED_SESSION
06:21
<alkisg>
True, we only want to write .dmrc when an LDM_SELECTED_SESSION is defined (and LDM_FORCE_SESSION doesn't override it)
06:22
Btw another reason why _DESKTOP might be confusing is the ubuntu package naming, people might think that LDM_DESKTOP="edubuntu-desktop" is valid because such a package exists
06:23
Let's sum up on what we want to do
06:23
LDM_SELECTED_SESSION is not an lts.conf variable, it's only set by the greeter
06:23
<vagrantc>
correct.
06:24
<alkisg>
LDM_SESSION keeps both its old meaning for compatibility, but that's deprecated and will only be around for a couple of years, and its new meaning, i.e. the xsession desktop file name
06:24
<vagrantc>
so much so, that it should obliterate the value if set
06:24
<alkisg>
Right
06:24
<vagrantc>
the LDM_SELECTED_SESSION
06:25
<alkisg>
The DM should show up the LDM_SESSION as the preselected one, if possible (I don't think there's code for it now), and the LDM_FORCE_SESSION as the only one available
06:25
.dmrc overrides LDM_SESSION but not LDM_FORCE_SESSION
06:26
<vagrantc>
if LDM_FORCE_SESSION is set to non-empty, it should just disble the menu.
06:26
<alkisg>
Or show the available sessions grayed out, I don't mind either wy
06:26
way
06:27
<vagrantc>
whatever's easiest to implement.
06:27
<alkisg>
LDM_SESSION handling is long, we first check .dmrc, then the .desktop file name, then the path if it doesn't have spaces, then run it if it has spaces
06:28
It's about the same with LDM_SELECTED_SESSION as ldminfod/ldm doesn't pass us the desktop file name
06:28
Finally, LDM_FORCE_SESSION checking is easy, we just check if the file name exists
06:28
That's about it?
06:28
<vagrantc>
what to do if LDM_FORCE_SESSION doesn't exist?
06:28
fail to log in? fall back to whatever?
06:29
<alkisg>
I think we should fail to login with a dialog
06:30
<vagrantc>
sounds good.
06:30
<alkisg>
How to handle the LDM_XSESSION meaning and priority?
06:31
Does it make sense for someone to define both LDM_XSESSION and LDM_SESSION?
06:31
<vagrantc>
it could.
06:31
<alkisg>
OK so we just prefix it as we do now
06:32
<vagrantc>
though really, *most* people shouldn't mess with LDM_XSESSION
06:32
it's nothing to do with the selected session, it's your selected /etc/X11/Xsession
06:32
which, in unusual cases, is valid to change ... some distros may not support other sessions, etc..
06:33
ah right, ldminfod has an xsession:/etc/X11/Xsession output ...
06:33
<alkisg>
We should also make ldminfod return the desktop file names to LDM, so that it's easier to check for missing sessions (LDM_SESSION, LDM_FORCE_SESSION) before login.
06:33
<vagrantc>
i.e. connecting to a different distro should have ldminfod tell you which LDM_XSESSION to use.
06:34
<alkisg>
Until we do that, LDM_SELECTED_SESSION can't contain a desktop file name, which would be the ideal case
06:34
<vagrantc>
alkisg: we could still log in, but just fail.
06:35
log in enough to check for the presence of the desktop file.
06:35
<alkisg>
Right, but we can't fix the UI to show the session as missing etc
06:36
<vagrantc>
hence LDM_FORCE_SESSION should just disable it.
06:36
<alkisg>
(before login)
06:36
<vagrantc>
alkisg: well, session-with-name should only return sessions that actually have desktop files.
06:36
<alkisg>
LDM_SESSION (the DEFAULT) should still show the session as preselected in the menu
06:37* vagrantc was going to just go ahead and upload earlier, but this sounds so much better.
06:37
<alkisg>
What I mean is: one sets LDM_SESSION=gnome-fallback. Before login, the greeter can't know if that session is available or not, so it can't have it preselected in the UI
06:37
<vagrantc>
right.
06:38
so we need another ldminfod entry ...
06:38
desktop-files: ... session-with-desktop: ?
06:38
<alkisg>
We can add it inside :: and in the parsing phase count the ::
06:38
(in case we deal with an older ldminfod)
06:38
<vagrantc>
so add it to session-with-name: ?
06:39
alkisg: this is where you want to maintain backwards compatibility
06:39
<alkisg>
Yeah
06:39
session-with-name:gnome-fallback:GNOME Classic (No effects):gnome-session --session=gnome-fallback
06:39
<vagrantc>
so old clients can connect to new servers.
06:40
all the ones i see have binaries with full paths
06:40
<alkisg>
we read 3 values out of this one, but if the last one is empty, we're dealing with an older ldminfod, so we swap 2 vars
06:40
<vagrantc>
alkisg: does the current ldm support adding another : ?
06:40
without failing badly?
06:41
<alkisg>
Meh, right, newer chroot in an older server, dunno, have to check
06:41
*older chroot in a newer server, whatever
06:41
<vagrantc>
alkisg: that's why we implemented session-with-name: instead of just adding to session
06:42
(well, actually people broke session: first)
06:42
<alkisg>
How long ago? Can we start using reusing session for this now?
06:42
<vagrantc>
alkisg: if older ldm can handle an arbitrary number of colons ... we can just keep appending, otherwise a new session-FOO:
06:43
ah, break compatibility with really old versions ? maybe.
06:44dobber has joined IRC (dobber!~dobber@213.169.45.222)
06:44
<alkisg>
session-with-name:GNOME Classic (No effects):gnome-session --session=gnome-fallback ==> I think we can append the desktop name at the end
06:44
Meh I meant to paste this: session-with-name:GNOME Classic (No effects):gnome-session --session=gnome-fallback
06:44
meh
06:44
Something wrong with my copy/paste : val = g_strsplit(lines[i], ":", 2);
06:45
<vagrantc>
december of 2008
06:45
<alkisg>
It only reads the first 2 colon-separated entries so we can add a 3rd one at the end
06:45
<vagrantc>
we did it right
06:46
<alkisg>
I hope no session command line contains : ...
06:46
<vagrantc>
eesh.
06:46
<alkisg>
Where are the session names translated? In the client? Nowhere?
06:47
Looking ahead in the future, we'd want to genereate .desktop files from init-ltsp.d
06:48
In the client: /usr/share/xsession/server1-gnome-session.desktop, with Exec='ltsp-launch-session <rest params>'
06:49
<vagrantc>
alkisg: yeah, i don't think we support translated session names.
06:51* vagrantc tests the spare : theory
06:52
<vagrantc>
alkisg: so we'd have to escape colons somehow...
06:53
if any of the session names, binary names, or desktop file names contain colons
06:53
hopefully, they just don't.
06:53
<alkisg>
vagrantc: do we do that already for the existing parts?
06:54
<vagrantc>
alkisg: no, but you scared me :P
06:54
"< alkisg> I hope no session command line contains : ..."
06:55
<alkisg>
Haha. Let's do a quick fix for now, and handle it all better in the new configuration system which will replace lts.conf, ldminfod and ltsp_config
06:55
I.e. it'll send the whole .desktop files to the clients
06:55
<vagrantc>
the quick fix, or the new idea?
06:56
<alkisg>
The new idea, later on, when we have pamssh etc
06:56
<vagrantc>
LTSP 6 :)
06:56
<alkisg>
Amen :)
06:58
<vagrantc>
alkisg: yup, ldm just seems to ignore the extra colon entries, just like you said :)
06:58
<alkisg>
Cool, so no compatibility problems there
06:58
<vagrantc>
so if we put the desktop filename on the end, it should work.
06:58
or at least, ldm 2.2.9
06:59* vagrantc will test the version in squeeze, too ... 2.1.2
07:01
<vagrantc>
that goes back to 2010-04-19 ...
07:01
so about 2 years ...
07:03
looks like that version supports session-with-name, and seems to read the colons fine.
07:05
alkisg: so, if LDM_SESSION isn't a desktop file, should we bother to check for binary location (if full path) or that the binary is in $PATH?
07:06
alkisg: or do we just let Xsession handle that (i think it falls back to "default" behavior.
07:06
)
07:07
<alkisg>
(10:06:16 πμ) vagrantc: alkisg: or do we just let Xsession handle that (i think it falls back to "default" behavior. ==> ah if it does, then we needn't bother
07:08bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
07:08
<vagrantc>
and LDM_FORCE_SESSION is just desktop files...
07:08
<alkisg>
Right
07:09
Actually since we're looking to replace ldm I'm not sure if it's worth it to do the extra ldminfod bits, just in case someone will want to optimize the ldm ui...
07:10
So we can skip the colons stuff and postpone it for ltsp 6
07:11
(and there, solve it by downloading the whole session.desktop files)
07:11
<vagrantc>
alkisg: it would definitely provide much cleaner code, though.
07:11
<alkisg>
The new code yes, but we'll keep the old code too, so the mess will be there
07:12
<vagrantc>
alkisg: in order to support older servers?
07:12
<alkisg>
Yeah
07:12
<vagrantc>
like, say an release with a support cycle of 5 years
07:13
<alkisg>
LTSP 6 will be in a couple of years, right? If we'll massively change the configuration system there, we can break backwards compatibility a bit then
07:14
<vagrantc>
alkisg: doh!
07:14
<alkisg>
So if we won't have ldminfod then, why bother making it send the file names now...
07:14
<vagrantc>
alkisg: Xsession: unable to launch "/usr/bin/openbox-session:foo"
07:14
alkisg: so, at least with 2.1.2, it splits on the first colon ...
07:14
<alkisg>
Ouch. /me votes to skip the colons bit :)
07:14
<vagrantc>
yup
07:17
is there another location to put .desktop files that we should support ? as an admin, i prefer local customizations in /etc or /usr/local or something.
07:17
for those custom .desktop files
07:17* alkisg looks...
07:19
<vagrantc>
alkisg: still worth it to implement LDM_SELECTED_SESSION and LDM_FORCE_SESSION and improve LDM_SESSION ?
07:19
<alkisg>
Yes, I think so
07:19
<vagrantc>
newer version also breaks with the extra colons.
07:20* vagrantc kills that idea dead.
07:20
<vagrantc>
although it actually wouldn't be too hard to generate another line...
07:20
<alkisg>
What does ldm return for LDM_SESSION if the user hasn't selected anything and there's no lts.conf?
07:20
<vagrantc>
i think "default"
07:21
by hasn't got an lts.conf, i presume you mean no LDM_SESSION settings?
07:21
<alkisg>
Yes
07:22
So, in case we don't want to touch ldm much and don't want to introduce LDM_SELECTED_SESSION, the downside is...
07:22
...that if LDM_SESSION is defined in lts.conf, then we set the user's .dmrc, while we shouldn't. Is there any other downside?
07:23
<vagrantc>
that's the main one.
07:23
<alkisg>
That can be mitigated by only writing LDM_SESSION if it's _not_ a desktop file name, and it is an Exec line of a desktop file name
07:23
So users that adopt the new LDM_SESSION=<desktop file name> scheme won't be affected by the problem
07:24
...while for the other ones, the problem did already exists, so they shouldn't mind :P
07:24
I.e. I think we can get away with only modifying the .dmrc processing script, nothing else
07:24
And only have 2 variables, LDM_SESSION and LDM_FORCE_SESSION
07:25
(10:23:19 πμ) alkisg: That can be mitigated by only writing LDM_SESSION if it's _not_ a desktop file name, and it is an Exec line of a desktop file name ==> I meant writing to .dmrc
07:25
<vagrantc>
so LDM_SESSION can be set from lts.conf/the environment ... and from selecting in the LDM greeter menu ...
07:25sha__ has left IRC (sha__!~sha@e177118226.adsl.alicedsl.de, Remote host closed the connection)
07:26
<alkisg>
Yes, and if it's only saved if it matches an Exec: line
07:27
<vagrantc>
the only sessions available currently all come from .desktop files ... so that should help.
07:27
alkisg: i think that's how it currently behaves.
07:28
alkisg: so the other problem, when LDM_SESSION is set, selecting from the menu will change the session once, write to the .dmrc, but the next time it doesn't respect the change.
07:28
set from lts.conf
07:28
<alkisg>
.dmrc takes precedence over LDM_SESSION
07:28
(or, should)
07:29
<vagrantc>
i don't think it does.
07:29
<alkisg>
Let's again sum up what we want:
07:29
Keep only 2 variables, LDM_SESSION and LDM_FORCE_SESSION. LDM_SESSION can be a session file name set from lts.conf, or an Exec: line set from ldm or lts.conf, or a custom line set from lts.conf.
07:30
The order is: LDM_FORCE_SESSION > .dmrc > LDM_SESSION
07:30
And we only write LDM_SESSION to .dmrc if it matches an Exec: line
07:30
Good enough? I think we can implement that by only modifying the dmrc handling script
07:31
And in the documentation we mention that using a whole command line is deprecated, but still supported for backwards compatibility
07:32
<vagrantc>
that doesn't allow you to ever change your .dmrc through LDM
07:32* alkisg just understood what vagrantc was saying, thinking...
07:33
<vagrantc>
if .dmrc overrides LDM_SESSION, you can't change your default, except by manually editing .dmrc
07:33
<alkisg>
LDM_FORCE_SESSION > LDM_SESSION (Exec:) > .dmrc > LDM_SESSION (desktop or other)
07:34
I was thinking about that, but didn't clear it up well in my head before
07:34* vagrantc still wants LDM to export LDM_SELECTED_SESSION
07:34
<vagrantc>
it would be so much clearer.
07:38toscalix has joined IRC (toscalix!~toscalix@212-166-224-166.red-acceso.airtel.net)
07:40
<vagrantc>
alkisg: i might almost have this...
07:41
<alkisg>
void __attribute__ ((visibility("default"))) ==> me hates all those stupid C modifiers
07:41
<muppis>
:D
07:42
<alkisg>
vagrantc: plugin.c, line 122: setenv("LDM_SESSION", session, 1); ==> setenv("LDM_SELECTED_SESSION", session, 1);
07:42
I think that's the only change required in ldm
07:43
<vagrantc>
alkisg: already tried that ... it then sets LDM_SELECTED_SESSION=default if nothing is selected
07:44
<alkisg>
vagrantc: that's OK though, we don't mind
07:44
It's the same as if the user has selected default manually, in which case we'd use his .dmrc, or fall back to LDM_SESSION
07:45
(or fall back to the X default session if LDM_SESSION is undefined)
07:47
So, LDM_FORCE_SESSION > LDM_SELECTED_SESSION (!=default) > .dmrc > LDM_SESSION > default X session
07:48
And we only save .dmrc if LDM_SELECTED_SESSION != default
07:48
<vagrantc>
right.
07:49
<alkisg>
OK, on to implementing it?
07:50* vagrantc has been.
07:59
<vagrantc>
alkisg: going to try this: http://paste.debian.net/172759/
08:01rkwesk_ has joined IRC (rkwesk_!~rkwesk@adsluser1289.att.sch.gr)
08:02F-GT has joined IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net)
08:06* vagrantc builds and tests
08:06
<alkisg>
vagrantc: ...how do I apply a patch? :P
08:07
<vagrantc>
alkisg: patch -p0 < patch ?
08:07
<alkisg>
Got it, was trying with a random number, -p2
08:07
<vagrantc>
alkisg: that's the level of subdirs
08:07
<alkisg>
Ah
08:08
<vagrantc>
alkisg: depending on thee patch, it's usually 0 or 1
08:08
i.e. if it includes ldm-trunk/src/foo.c or just src/foo.c
08:09
<rkwesk_>
Now that some are using ltsp without chroot how is epoptes-client -c going to be setup having only thin client?
08:09
<alkisg>
vagrantc: in the code you pasted you don't take LDM_SESSION into account
08:10
rkwesk_: with ltsp-pnp we don't have "plain thin client" setups, only setups that support both thin and fat clients
08:10
<vagrantc>
alkisg: if LDM_SESSION is passed through lts.conf, we don't bother with it
08:11
alkisg: i.e. no save to .dmrc ...
08:11
<alkisg>
vagrantc: we do need to do the desktop -> command line change
08:12
<vagrantc>
alkisg: if LDM_SESSION=desktopfile
08:12
right ...
08:12
<alkisg>
rkwesk_: so you can just run `epoptes-client -c localhost` on your server
08:12
<rkwesk_>
yes, ok but epoptes-client -c fails to fetch certificate although /etc/hosts has the ip with server
08:12
<alkisg>
rkwesk_: although it's not needed as the keys are already there
08:13
<rkwesk_>
ah, so no need to do epoptes-client -c at all
08:13
<alkisg>
Right
08:13
<rkwesk_>
that's one more welcome simplification!!
08:16
another thing i would like 2 raise is restricting the client from opening console or looking at other client's home
08:17
<alkisg>
rkwesk_: I think it's better to teach them to protect their private files with right click => properties
08:18
rkwesk_: linux is designed such as /home/username is readable, to get the user's .face and other stuff
08:18
If you really need it, you can execute chmod 700 /home/*
08:18
<rkwesk_>
ok, i just am thinking ahead to mischief makers in lab
08:18
<alkisg>
About restricting the client, I don't think that would work either, I prefer security at a lower level and not at the GUI level
08:19
E.g. someone can press alt+f2 or navigate to a dir an execute a file from nautilus or write nautilus scripts that launch programs or create a .desktop launcher etc etc
08:19
<rkwesk_>
i agree. an inexpierenced sysadmin needs help here
08:19
<alkisg>
Why?
08:20
What can students do to cause harm?
08:20
<rkwesk_>
if in lab and someones alters another's work
08:20
<alkisg>
He doesn't have write rights by default
08:21
So he won't be able to unless the other user allows him by right clicking the document and changing its attributes to writeable by all
08:21
<rkwesk_>
so each user is in seperate group?
08:21
<alkisg>
Of course
08:21
And there are no write rights to groups either
08:22
So even if they had the same group, they still wouldn't be able to modify other users's files
08:22
(unless the other user gave write access to some groups on purpose)
08:22
<rkwesk_>
ok. where does the teacher hide docs he/she does not want read or even that it exists
08:23
<alkisg>
Right click on documents, create "private" folder, right click on folder, remove read rights
08:24
<rkwesk_>
so one needs to teach an inexpierenced teacher/sysadmin how and why
08:25
<alkisg>
Yes, same as in windows XP or before
08:25
They had to right click and remove [v] Everyone from the access list
08:26
Their private data like emails is already private, in .* folders
08:26
<rkwesk_>
what about having a private folder created by default? Then we go ahead and explain why
08:27
<alkisg>
The "explain why" part is the big part, while making a folder private is the easy part
08:27
<rkwesk_>
the hidden folders are also hidden from the owner at first
08:29
yes i agree. so many r afraid of technology
08:29
<alkisg>
In the new sch-scripts we'll have a "create shared folders" functionality. We can probably add a "create private folders" UI there.
08:30
<rkwesk_>
anything that simplifies is a positive
08:30
<alkisg>
But I think it's best if both the teacher and the students learn about file rights, if the class is old enough to have students that might cause mischiefs
08:31* alkisg would prefer it if all Documents, Desktop, Videos etc folders were private by default, and only the Public folder remained readable by others
08:31
<rkwesk_>
agreed. i can forsee some eyes rolling, however.
08:32
<alkisg>
But that's an XDG decision to make, I'm not sure if we should default to that.
08:33
<rkwesk_>
best to not reinvent away from common setup
08:34
just a few simple additions like a private place for anyone who wants it
08:34Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com)
08:35
<rkwesk_>
then we can explain the role of root
08:35
and why no one is root unless a specific action is taken
08:38
apart from the discussion in Athens on this new no chroot ltsp are other places going to hold such workshops?
08:38tol has joined IRC (tol!~tol@lvps178-77-99-218.dedicated.hosteurope.de)
08:39
<rkwesk_>
anyone here not in Greece interested to hold such an event?
08:40toscalix has left IRC (toscalix!~toscalix@212-166-224-166.red-acceso.airtel.net, Remote host closed the connection)
08:40
<alkisg>
rkwesk_: not many people use the new setup yet, it's very new, not available by default in 12.04 . I imagine more people will start using it gradually.
08:40
In Greece we'll have more workshops and seminars etc over time
08:41
<rkwesk_>
i will be happy 2 help
08:41
<vlt>
tol: Hello.
08:43rkwesk_ has left IRC (rkwesk_!~rkwesk@adsluser1289.att.sch.gr, Quit: Leaving)
08:52toscalix has joined IRC (toscalix!~toscalix@62-87-105-165.red-acceso.airtel.net)
08:54ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de)
09:04
<vagrantc>
alkisg: ok, i think we have it ... very similar to your commit a couple ago ... one more test case and then i'll push
09:04
<alkisg>
vagrantc: if you want upload the file to pastebin for me to take a quick look in case we forgot something
09:04
Also, isn't it time to head to bed? :)
09:05
<vagrantc>
it's a bit past time
09:05
it's a pastebin pasttime
09:06
alkisg: http://paste.debian.net/172766/
09:06
<alkisg>
vagrantc: we should also update the comments on top
09:07
<vagrantc>
alkisg: yes, probably so.
09:08
oh, i haven't tested LDM_FORCE_SESSION
09:09
though LDM_FORCE_SESSION will just fall back to the defaults if it's not possible to set anything.
09:10
<alkisg>
It seems OK, but testing is the best way to tell :)
09:10* alkisg is off, bb
09:10alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
09:28toscalix has left IRC (toscalix!~toscalix@62-87-105-165.red-acceso.airtel.net, Remote host closed the connection)
09:47andygraybeal has joined IRC (andygraybeal!~andy@h167.93.213.151.dynamic.ip.windstream.net)
09:49khildin has joined IRC (khildin!~khildin@ip-83-134-215-68.dsl.scarlet.be)
10:09Trixboxer has joined IRC (Trixboxer!~Trixboxer@115.124.115.71)
10:10vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Ping timeout: 252 seconds)
10:56tol_ has joined IRC (tol_!~tol@lvps178-77-99-218.dedicated.hosteurope.de)
11:13bieb1 has joined IRC (bieb1!~hbieber@173.226.16.130)
11:38ogra_ has left IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 260 seconds)
11:39ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de)
11:50Parker955_Away is now known as Parker955
11:55Parker955 is now known as Parker955_Away
12:07yalu has left IRC (yalu!~yalu@192.24-200-80.adsl-dyn.isp.belgacom.be, Ping timeout: 248 seconds)
12:15khildin_ has joined IRC (khildin_!~khildin@ip-83-134-213-196.dsl.scarlet.be)
12:18khildin has left IRC (khildin!~khildin@ip-83-134-215-68.dsl.scarlet.be, Ping timeout: 252 seconds)
12:19toscalix has joined IRC (toscalix!~toscalix@240.216.219.87.dynamic.jazztel.es)
12:32mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)
12:39khildin_ has left IRC (khildin_!~khildin@ip-83-134-213-196.dsl.scarlet.be, Ping timeout: 240 seconds)
12:43adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
12:44litlebuda has left IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt, Remote host closed the connection)
12:49dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
12:51khildin_ has joined IRC (khildin_!~khildin@ip-80-236-225-27.dsl.scarlet.be)
13:10elias_a has left IRC (elias_a!elias@hilla.kapsi.fi, Read error: Operation timed out)
13:12shamino has left IRC (shamino!shamino@hilla.kapsi.fi, Ping timeout: 276 seconds)
13:13elias_a has joined IRC (elias_a!elias@hilla.kapsi.fi)
13:13shamino has joined IRC (shamino!shamino@hilla.kapsi.fi)
14:44alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
15:07daya has joined IRC (daya!~daya@unaffiliated/daya)
15:11darkpixel_ has joined IRC (darkpixel_!~darkpixel@curetheitch/staff/darkpixel)
15:14staffencasa has joined IRC (staffencasa!~staffenca@128.193.8.220)
15:19alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
15:23F-GT has left IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net, Ping timeout: 250 seconds)
15:37tol has left IRC (tol!~tol@lvps178-77-99-218.dedicated.hosteurope.de, Quit: leaving)
15:37tol_ has left IRC (tol_!~tol@lvps178-77-99-218.dedicated.hosteurope.de, Quit: leaving)
15:39dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
15:40garymc has joined IRC (garymc!~chatzilla@81.138.225.164)
15:41tol has joined IRC (tol!~tol@lvps178-77-99-218.dedicated.hosteurope.de)
15:41
<garymc>
Anone know How I can block all but a few websites for my LTSP users? I can do it via router but then my computer is blocked even though I told the router to only screen the LTSP server
15:42
Need to get this done cos some people in the office are taking the piss and sitting on Facebook and eating up our bandwidth
15:49
<knipwim>
use a local /etc/host file on each client?
15:55mikkel has joined IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk)
15:57
<garymc>
knipwim is there a more universal way?
15:58
<knipwim>
hmm, i though it was pretty universal, you can maintain it on the server, and download it at client boot time
15:59
you can probably do it in iptables, some kind of host based blocking, but i wouldn't know how
15:59map7 has left IRC (map7!~map7@teksup41.lnk.telstra.net, Ping timeout: 240 seconds)
16:00
<dead_inside>
garymc are you using thin or fat clients? if your using thin you should only need to edit /etc/hosts on the server, if your using fats then you only need to edit /etc/hosts on the image and make sure they dont have any dns servers
16:01
<garymc>
im using FAT
16:01
dont remeber how to do all that :(
16:07darkpixel_ has left IRC (darkpixel_!~darkpixel@curetheitch/staff/darkpixel, Ping timeout: 244 seconds)
16:07bieb1 has left IRC (bieb1!~hbieber@173.226.16.130, Max SendQ exceeded)
16:08DIoX|DaZ has left IRC (DIoX|DaZ!~KaKa@server.civicclub.lt, Ping timeout: 244 seconds)
16:08
<||cw>
garymc: you should be able to do it on your router, if it's ignoring your exclusions you either didn't apply it right or it's a router bug
16:08DIoX|DaZ has joined IRC (DIoX|DaZ!~KaKa@server.civicclub.lt)
16:09
<garymc>
yeah must be a router bug
16:09khildin_ has left IRC (khildin_!~khildin@ip-80-236-225-27.dsl.scarlet.be, Quit: I'm gone, bye bye)
16:09
<garymc>
BT hg2700
16:09
nightmare
16:10
im gopnna try setup a newer router
16:10bieb has joined IRC (bieb!~hbieber@173.226.16.130)
16:18garymc has left IRC (garymc!~chatzilla@81.138.225.164, Ping timeout: 244 seconds)
16:18vagrantc has joined IRC (vagrantc!~vagrant@c-76-115-60-19.hsd1.or.comcast.net)
16:18vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
16:19knipwim has left IRC (knipwim!~wim@ip4da83870.direct-adsl.nl, Ping timeout: 244 seconds)
16:19knipwim has joined IRC (knipwim!~wim@ip4da83870.direct-adsl.nl)
16:20* vagrantc ponders setting ssh -o CheckHostIP=no
16:20
<vagrantc>
in ldm by default...
16:21toscalix has left IRC (toscalix!~toscalix@240.216.219.87.dynamic.jazztel.es, *.net *.split)
16:21Trixboxer has left IRC (Trixboxer!~Trixboxer@115.124.115.71, *.net *.split)
16:21Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, *.net *.split)
16:21SmallR2002 has left IRC (SmallR2002!~quassel@c-98-253-173-240.hsd1.il.comcast.net, *.net *.split)
16:21Parker955_Away has left IRC (Parker955_Away!~parker@74.112.203.151, *.net *.split)
16:21
<vagrantc>
although it didn't actually cause problems now that it checks against hostname ... it just created a /root/.ssh/known_hosts with an entry for the ip address ...
16:27
rebooting the thin client would be enough to fix that
16:27
if you did load-balancing by the same name it could cause problems...
16:31map7 has joined IRC (map7!~map7@teksup41.lnk.telstra.net)
16:31toscalix has joined IRC (toscalix!~toscalix@240.216.219.87.dynamic.jazztel.es)
16:31Trixboxer has joined IRC (Trixboxer!~Trixboxer@115.124.115.71)
16:31Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com)
16:31SmallR2002 has joined IRC (SmallR2002!~quassel@c-98-253-173-240.hsd1.il.comcast.net)
16:31Parker955_Away has joined IRC (Parker955_Away!~parker@74.112.203.151)
16:32toscalix_ has joined IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es)
16:34toscalix has left IRC (toscalix!~toscalix@240.216.219.87.dynamic.jazztel.es, Remote host closed the connection)
16:34
<vagrantc>
ok, i think i'll just tag ldm 2.2.10 as is
16:34
knipwim, stgraber: any desired last minute features in ldm before i tag?
16:34* vagrantc waits for the the, stop, stop, hold the press moment
16:35
<stgraber>
vagrantc: nope, go ahead
16:37litlebuda has joined IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt)
16:37
<vagrantc>
do we have any c++ code in ldm?
16:39
stgraber, mgariepy : i've been trying to enable hardened build flags for ldm, and resolved most of them, but i still get this lintian warning: W: ldm: hardening-no-fortify-functions usr/lib/ldm/libssh.so.0.0.0
16:40
i did have CPPFLAGS imported, but it didn't seem to take effect for that one.
16:43troels has joined IRC (troels!~troels@0x573c2c98.bynqu1.dynamic.dsl.tele.dk)
16:45
<stgraber>
it's all in C
16:45
not sure why the hardening doesn't work for libssh
16:47* vagrantc disables CPPFLAGS to see if it works for other stuff.
16:48toscalix_ has left IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es, Remote host closed the connection)
16:48toscalix_ has joined IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es)
16:48* vagrantc is pretty much cargo-culting the hardening stuff
16:49
<knipwim>
vagrantc: have you seen the "bug report" in the dev mailing list
16:49
from floris bos
16:50
<vagrantc>
knipwim: oh yeah ...
16:50
wasn't clear to me weather those were worth applying.
16:50staffencasa_ has joined IRC (staffencasa_!~staffenca@128.193.8.220)
16:50
<knipwim>
i can't really judge them, if they should be included or not
16:51
<vagrantc>
i'll apply them and test them. :)
16:53staffencasa has left IRC (staffencasa!~staffenca@128.193.8.220, Ping timeout: 244 seconds)
16:53
<vagrantc>
stgraber: so, other binaries, ldm, ldmgtkgreet, ldm-dialog pick up the CPPFLAGS, but for some reason libssh doesn't
16:54staffencasa_ has left IRC (staffencasa_!~staffenca@128.193.8.220, Client Quit)
16:54staffencasa has joined IRC (staffencasa!~staffenca@128.193.8.220)
16:55
<vagrantc>
stgraber: for the full lintian warning: http://paste.debian.net/172835/
16:59troels has left IRC (troels!~troels@0x573c2c98.bynqu1.dynamic.dsl.tele.dk, Quit: Leaving)
17:01litlebuda has left IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt, Remote host closed the connection)
17:04komunista has joined IRC (komunista!~slavko@adsl-195-168-242-147.dynamic.nextra.sk)
17:08* vagrantc wonders about
17:08
<vagrantc>
greeter.c: In function 'handle_command':
17:08
greeter.c:308:5: warning: 'g_strncasecmp' is deprecated (declared at /usr/include/glib-2.0/glib/gstrfuncs.h:175) [-Wdeprecated-declarations]
17:10
<mgariepy>
stgraber, http://paste.ubuntu.com/1023513/ do you think that should be applied to 12.04 ?
17:11
i have a variable name issue with getltscfg-cluster and client/share/ltsp/init-ltsp.d/04-server
17:11
<stgraber>
mgariepy: yeah, that's definitely a bug, no reason not to fix that in 12.04
17:19
<mgariepy>
stgraber, should i add some patch for the bug ?
17:20
like need-sru ?
17:21
<stgraber>
mgariepy: just give me the bug number
17:21
<mgariepy>
https://bugs.launchpad.net/ltsp/+bug/1008618
17:48toscalix_ has left IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es, Remote host closed the connection)
17:48toscalix_ has joined IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es)
17:49bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
17:51toscalix_ has left IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es, Remote host closed the connection)
17:52toscalix_ has joined IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es)
17:56toscalix_ has left IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es, Excess Flood)
17:59toscalix_ has joined IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es)
18:19andygraybeal has left IRC (andygraybeal!~andy@h167.93.213.151.dynamic.ip.windstream.net, Ping timeout: 246 seconds)
18:38alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
18:38Trixboxer has left IRC (Trixboxer!~Trixboxer@115.124.115.71, Quit: "Achievement is not the end, its the beginning of new journey !!!")
18:57Parker955_Away is now known as Parker955
18:57toscalix_ has left IRC (toscalix_!~toscalix@240.216.219.87.dynamic.jazztel.es, Remote host closed the connection)
19:14jordanm has left IRC (jordanm!~jordanm@li111-251.members.linode.com, Changing host)
19:14jordanm has joined IRC (jordanm!~jordanm@pdpc/supporter/active/jordanm)
19:17Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, Read error: Connection reset by peer)
19:26
<vagrantc>
alkisg: oh-ho!
19:26
<alkisg>
Ahoy there
19:27
<vagrantc>
so, i'm trying to disable the display of the session selection when LDM_FORCE_SESSION is set ...
19:27
but just floundering.
19:27
<alkisg>
Whoah, digging into C? :P
19:27andygraybeal has joined IRC (andygraybeal!~andy@h167.93.213.151.dynamic.ip.windstream.net)
19:28
<vagrantc>
http://paste.debian.net/172868/
19:28
<alkisg>
vagrantc: I'll check the code a bit and see if I can understand anything (/me hasn't used glib yet)... did you try the ML patches? That guy seemed like he knew what he was talking about...
19:30
<vagrantc>
alkisg: yeah, those patches seem to at least not break anything. :)
19:30||cw has left IRC (||cw!~chris@phpgroupware/cw, Remote host closed the connection)
19:31
<alkisg>
vagrantc: + if (getenv('LDM_FORCE_SESSION') != NULL)
19:31
Not ==
19:31
Err
19:31
Sorry
19:32
<vagrantc>
alkisg: but i only want it to include the menu if LDM_FORCE_SESSION is not set
19:32
<alkisg>
vagrantc: double quotes
19:32
"LDM_FORCE_SESSION"
19:32
<vagrantc>
doh!
19:32
<alkisg>
Single quotes are for characters only
19:32
(single characters, I mean)
19:32
<vagrantc>
so used to using them mostly interchangeably
19:34
<alkisg>
https://bugs.launchpad.net/ltsp/+bug/282884 => Lns also asked for an option to disable the language menu too
19:34||cw has joined IRC (||cw!~chris@gateway.wilsonmfg.com)
19:34||cw has joined IRC (||cw!~chris@phpgroupware/cw)
19:35
<vagrantc>
alkisg: yes, i vaguely recalled that
19:40highvoltage has left IRC (highvoltage!~highvolta@sateda.stgraber.org, Changing host)
19:40highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)
19:41* alkisg wonders how easy it would be, if the user enters a wrong username/password, to not restart ldm...
19:44ricotz has left IRC (ricotz!~rico@unaffiliated/ricotz, Quit: Ex-Chat)
19:44
<vagrantc>
just restart the greeter?
19:44
alkisg: i seem to recall some concern about values being selected from the previous session
19:45
i don't really remember the details
19:45
<alkisg>
vagrantc: no need to restart the greeter, just ask for the username/password a second time
19:45bieb has left IRC (bieb!~hbieber@173.226.16.130)
19:46
<alkisg>
OK if there's such a concern, we could restart the greeter, but I don't see any reason to restart X just for a wrong password...
19:47
<vagrantc>
cuz we too dumb.
19:48
alkisg: oh, i think it's because there's no way to know what's wrong, hence scotty's libpam/libnss modules
19:49
alkisg: the double-quotes did it!
19:49* vagrantc checks to make sure it works properly when LDM_FORCE_SESSION is not set.
19:49
<alkisg>
vagrantc: if the user logged in, we restart X when ldm finishes, if he didn't login, no need to restart it, afaik
19:49
<vagrantc>
if so, we can also easily solve LNS's wishlist request, too.
19:50
alright. cool.
19:50* vagrantc did something.
19:53* vagrantc wonders about variable names for the menu-disabling features
19:53
<alkisg>
...cool, time to add "C programming" in your resume ;)
19:54
<vagrantc>
i've done other such trivial patches before :)
19:54
<alkisg>
vagrantc: isn't LDM_FORCE_SESSION enough? Do you think another variable is needed?
19:54
(and, LDM_FORCE_LANGUAGE...)
19:55
<vagrantc>
LDM_FORCE_LANGUAGE, LDM_FORCE_BACKEND ?
19:55alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection)
19:55
<alkisg>
LDM_FORCE_BACKEND already exists
19:55
<vagrantc>
oh.
19:55* vagrantc should learn about grep
19:55alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
19:57
<alkisg>
vagrantc: I think that ldm exits(1) on failure, and exits(0) when the user managed to log in - I think we can use that from the shell script to respawn it without restarting X
19:58* vagrantc wonders where the backend setting code is
20:06vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi)
20:11
<alkisg>
Meh making ldm restart only after successful login was easy. Making the Ubuntu logo go away without restarting X is the hard part! Nah, I don't want to invest on LDM that much... :-/
20:12
(i.e. the logo was there inside the user session)
20:12
(on the second login)
20:13
<vagrantc>
alkisg: as in you think we should push, or it's too buggy?
20:13
oh, i see.
20:14* vagrantc notices LDM_DEFAULT_BACKEND
20:14
<vagrantc>
hmmm...
20:15
<alkisg>
There's a bug report related to the backend menu item being untranslatable because it's not in the gtk template file etc
20:15
https://bugs.launchpad.net/ltsp/+bug/967475
20:15
<vagrantc>
apparently, LDM_FORCE_BACKEND does remove it from the menu
20:16* vagrantc test LDM_FORCE_LANGUAGE
20:16
<vagrantc>
and then i think i'll upload.
20:17
<alkisg>
vagrantc: I think we should also push the ML patches, the "use timeout box before initialization" sounds a bit scary
20:18
<vagrantc>
yes, in progress
20:30daya has left IRC (daya!~daya@unaffiliated/daya, Quit: Leaving)
20:39
<vagrantc>
stgraber: does this commit have a LP bug associated to it?
20:39
1434 Stéphane Graber 2012-04-30
20:39
ldminfod: If Hidden=true or TryExec points to an invalid target, hide the entry.
20:40mikkel has left IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk, Quit: Leaving)
20:40
<stgraber>
vagrantc: yes
20:40
vagrantc: https://bugs.launchpad.net/ubuntu/+source/ldm/+bug/991745
20:40
<vagrantc>
stgraber: thanks.
20:41
wanted to include it in the changelog
20:45
alkisg: so, apparently there's a per-package substvars file.
20:46
alkisg: i.e. debian/ltsp-client-core.substvars
20:46
<alkisg>
vagrantc: and, example content?
20:46
<vagrantc>
alkisg: so now it's just a matter of figuring out when to insert stuff into them.
20:46
<alkisg>
misc:ltsp=package1,package2 etc?
20:47
<vagrantc>
cat ltsp-server.substvars
20:47
misc:Depends=debconf (>= 0.5) | debconf-2.0
20:47
<alkisg>
Ah autogenerated right
20:47
<vagrantc>
i think is: value: foo | bar, baz
20:48
so we could have: ltsp-server.substvars contain ubuntu: x, y, x \n debian: x, z
20:48
or probably more like ubuntu-depends, debian-recommends
20:49
and tweak it somewhere in debian/rules
20:50
or simply put in ltsp-depends: and change it from debian/rules
20:51mmetzger has left IRC (mmetzger!~mmetzger@99-71-214-196.lightspeed.mdldtx.sbcglobal.net, Ping timeout: 265 seconds)
20:53mmetzger has joined IRC (mmetzger!~mmetzger@99-71-214-196.lightspeed.mdldtx.sbcglobal.net)
20:58
<alkisg>
vagrantc: http://raphaelhertzog.com/2010/09/27/different-dependencies-between-debian-and-ubuntu-but-common-source-package/
20:58
Haha this is fun: dpkg-vendor --derives-from Ubuntu && echo yes
21:02
<vagrantc>
alkisg: that looks easy for a single source package that builds a single binary ...
21:03
alkisg: but having to invole the dh_gencontrol multiple times with different packages seems tricky...
21:03
invoke
21:12
alkisg: ok, i think i found the proper fix to your debuild -tc problem.
21:13
<alkisg>
vagrantc: bzr2301+pkg1129 built fine on launchpad though :)
21:14
<vagrantc>
alkisg: did you work around it, or did it just start working?
21:14
<alkisg>
Didn't you say you did something?
21:14
It started working, I didn't do anything...
21:14
<vagrantc>
alkisg: i did something that shouldn't have fixed it, and didn't fix it for me.
21:15
<alkisg>
vagrantc: maybe it doesn't fix it for debuild -b -tc, I haven't tried, but it worked on launchpad, whereas previous packaging versions didn't work
21:15
<vagrantc>
huh.
21:15
<alkisg>
So I no longer need to maintain an ltsp-packaging launchpad branch just for that, I use yours instead :P
21:15adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
21:16
<alkisg>
Click on the recipe, and in 8 hours the new ltsp package is built, awesome :D
21:16adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
21:17
<vagrantc>
that's a bit ... slow
21:20adrianorg_ has joined IRC (adrianorg_!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
21:20adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
21:33Parker955 is now known as Parker955_Away
21:43
<alkisg>
Yeah it is slow, it was supposed to be a joke... sometimes it builds fast, on e.g. 20 minutes, sometimes super slow, 20+ hours...
21:44adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
21:44adrianorg_ has left IRC (adrianorg_!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
21:50komunista has left IRC (komunista!~slavko@adsl-195-168-242-147.dynamic.nextra.sk, Quit: Leaving.)
21:55dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...)
21:55adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
21:55adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
22:07adrianorg_ has joined IRC (adrianorg_!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
22:07adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Write error: Connection reset by peer)
22:09adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
22:09adrianorg_ has left IRC (adrianorg_!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
22:10
<alkisg>
vagrantc: maybe this?
22:10
Depends: ${${dist}:depends}
22:10
dpkg-gencontrol -Vdist=$detected_dist
22:10
And in package.substvars:
22:10
ubuntu:depends=packages
22:10
debian:depends=packages
22:10
<vagrantc>
alkisg: no, the problem is on a per-package basis
22:11
alkisg: not ubuntu vs. debian
22:11
<alkisg>
vagrantc: you can have multiple package.substvars
22:11* vagrantc will look into it after finishing up ldm 2.2.10-1 upload
22:11
<alkisg>
ltsp-client-core.substvars, ltsp-client.substvars etc...
22:12
Sure
22:12
<vagrantc>
alkisg: the instructions talk about calling dh_gencontrol -- -Vfoo:Depends="foo | bar, baz"
22:13
which i don't see a sane way to do that per-package ... but maybe you don't have to specify -V if the package.substvars files have the right stuff in them.
22:13
<alkisg>
That's to define additional vars, but you can also have substvars files, at least that's what man deb-substvars says
22:13* vagrantc hopes so
22:13adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
22:13* vagrantc ignores irc
22:14adrianorg__ has joined IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
22:16risca has joined IRC (risca!~risca@m90-129-58-193.cust.tele2.se)
22:17F-GT has joined IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net)
22:17
<vagrantc>
oh, i think this upload should fix a debian bug too ...
22:17* vagrantc stops the presses
22:18adrianorg_ has joined IRC (adrianorg_!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br)
22:18adrianorg__ has left IRC (adrianorg__!~adrianorg@189.58.224.221.dynamic.adsl.gvt.net.br, Read error: Connection reset by peer)
22:18risca has left IRC (risca!~risca@m90-129-58-193.cust.tele2.se, Client Quit)
22:21
<vagrantc>
or maybe nnot ... i'll just ask them to confirm if it's fixed.
22:30F-GT has left IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net, Ping timeout: 240 seconds)
22:30
<vagrantc>
ldm 2.2.10 tagged, pushed and uploaded.
22:40
<alkisg>
Nice work :)
22:40
4.10 Variable substitutions: debian/substvars
22:40
The debian/substvars file is usually generated and modified dynamically by debian/rules targets, in which case it must be removed by the clean target.
22:41
<vagrantc>
alkisg: where's that from?
22:41
<alkisg>
http://www.debian.org/doc/debian-policy/ch-source.html
22:42
<vagrantc>
sounded like policy.
22:42
<alkisg>
So we could easily create different debian/ltsp-xx.substvars files from debian rules (contents depending on the distro), but I think that'll also work with static files as well
22:43
<vagrantc>
alkisg: we need to append, not just create, in case one of the other hooks already added to it.
22:43
<alkisg>
(what I proposed at first)
22:43
<vagrantc>
it won't work with static files if they get removed on the clean target.
22:45* highvoltage notices new ldm with session stuff uploaded to debian :)
22:46
<alkisg>
I wonder if we could also put them as X-ubuntu-depends: lines in debian/control, so that they're all together in the same file, and extract them with a small script, and append them to substvars.
22:47
<vagrantc>
alkisg: there's a certain elegance there ...
22:47
alkisg: although at that point you might just want to use control.in and do some sed-magic on it.
22:48
<alkisg>
Ah, true, that would be easier
22:48
<vagrantc>
you also can't technically modify debian/control at build time, at least in Debian.
22:48
if our diff is fairly small, i'd like to go with substvars.
22:48
or at least try it out.
22:48
<alkisg>
With X-ubuntu-depends you don't need to modify the control file
22:49
<vagrantc>
alkisg: and it will just get ignored in the generated package?
22:49
<alkisg>
I think so, don't X-* vars get ignored?
22:49
<vagrantc>
alkisg: people grumble about arbitrary cruft in the control files.
22:49
<alkisg>
OK
22:50
Is control.in something widely used?
22:50
<vagrantc>
alkisg: X-* get ignored by certain tools ... but there's nothing like experimentation.
22:50Parker955_Away is now known as Parker955
22:50
<vagrantc>
alkisg: control.in or something similar is definitely used fairly often, although it's implemented a hundred different ways.
22:51DIoX|DaZ has left IRC (DIoX|DaZ!~KaKa@server.civicclub.lt, Ping timeout: 240 seconds)
22:53DIoX|DaZ has joined IRC (DIoX|DaZ!~KaKa@server.civicclub.lt)
22:53
<alkisg>
vagrantc: http://www.debian.org/doc/debian-policy/ch-controlfields.html, last paragraph
23:00
<vagrantc>
alkisg: so, in theory, we could add X-Ubuntu-Depends and X-Debian-Depends, have debian/rules grab those values and stick them in the appropriate substvars files?
23:00
alkisg: and it wouldn't show up anywhere else?
23:00
<alkisg>
vagrantc: right, and what I like about it is that it's very easy to see the differences
23:00
Yup
23:00
<vagrantc>
it is a most excellent idea if we can get it to work! :)
23:02* alkisg tests if lintian or other tools produce warnings about X-Debian-Depends:...
23:05
<alkisg>
All seem fine
23:06
<vagrantc>
grep-dctrl -F Package -X ltsp-server --show-field X-Debian-Depends control
23:06
will get us the values
23:06
<alkisg>
Whoah, nice one
23:08
<vagrantc>
grep-dctrl -F Package -X ltsp-server --show-field X-Debian-Depends --no-field-names control
23:09
<alkisg>
(correct output, and): grep-dctrl: control:72: expected a colon
23:09
A space instead of an empty line
23:11
<vagrantc>
yeah, i'm getting that too
23:11
<alkisg>
vagrantc: grep -n ' $' control
23:11
3 lines need fixing
23:11
<vagrantc>
oh.
23:12
<alkisg>
(we should do that in ltsp sources as well some time)
23:13* alkisg didn't know about grep-dctrl, the solution now seems even more elegant :)
23:13* vagrantc is good for something
23:13
<vagrantc>
alkisg: this is a really good idea.
23:15* alkisg hopes that soon Ubuntu will use the Debian LTSP package as is
23:16
<vagrantc>
alkisg: if we can resolve the differences in desired dependencies, what are the remaining issues?
23:17
<alkisg>
ltsp cluster, ltsp live and the ltsp-client-builder differences
23:17
But I think stgraber wants to extract ltsp-live to a separate package,
23:17
and the rest should be easy enough to merge
23:17
<vagrantc>
probably ab good idea with ltsp-cluster, too
23:18
<alkisg>
Yeah I'd like that too
23:18
<vagrantc>
we've dirverged quite a bit with ltsp-client-builder, last i looked.
23:18
that could arguably be it's own source package as well, though.
23:24* alkisg ==> pumpkin, very good work these days :)
23:25alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
23:26JesseC has left IRC (JesseC!~JesseCWor@wsip-98-175-20-126.br.br.cox.net, Ping timeout: 250 seconds)
23:45JesseC has joined IRC (JesseC!~JesseCWor@wsip-98-175-20-126.br.br.cox.net)