Profile Management Tools/Irclog.5.08.2013

From AppArmor
Jump to: navigation, search
[00:32] <cboltz> kshitij8: are you ready for the meeting? ;-)
[00:33] <kshitij8> yes :)
[00:33] <cboltz> sbeattie: do you know where John is?
[00:35] <cboltz> hmm, let's start the meeting anyway ;-)
[00:35] <cboltz> kshitij8: can you start with the usual last week / this week report?
[00:35] <sarnold> john's on vacation :)
[00:35] <kshitij8> yess. though I hope he gets back soon I need my mail answered.
[00:35] <cboltz> maybe also some lines about the summit ;-)
[00:36] <cboltz> sarnold: you just volunteered to replace him in this meeting then ;-)
[00:36] <sarnold> cboltz: I'll try :)
[00:36] <kshitij8> well the start of week saw some fairly big commits and the equally huge reviews ;)
[00:37] <kshitij8> i was able to get the regexparser working and wrote some testcases during my summit.
[00:38] <kshitij8> continuing on the sameline Im fixing bugs and stuff from the review. once all testing stuff passes I'll simply wrap it into logprof tool
[00:38] <kshitij8> that wont take longer the few minutes i suspect.
[00:39] <kshitij8> besides, I also did a bit of refractoring and plan on splitting code further.
[00:41] <kshitij8> from the summit, well we had a some googlers speak about new products and were in a way promoting them. Besides that we(the students) participated in events, cultural performances and presentations.
[00:42] <kshitij8> We also had a fun night. The trip was good but the schedule was pretty tight. :)
[00:43] <kshitij8> so wasnt a highly productive week overall.
[00:46] <cboltz> well, I wouldn't call your commits small ;-)
[00:48] <kshitij8> i tried to do a bit of extra work to compensate for the summit. :)
[00:50] <cboltz> I usually don't manage to write any code while on a conference, so your "small" things like testcases are better than what I expected ;-)
[00:52] <cboltz> so for this week you said you have to fix several things to pass the testcases
[00:52] <cboltz> does this fill the whole week or do you have more things planned?
[00:52] <kshitij8> I actually ended up finally seeing the bug in aregex in regexparser  code on-flight. :D
[00:54] <kshitij8> well i dont think it'll take the entire week. I'll probbaly move to another tool after logprof maybe midweek or by weekend.
[00:55] <cboltz> good candidates may be genprof (quite similar to logprof, with the exception that it starts with a new profile) 
[00:55] <cboltz> and mergeprof (at least the two-way merge should be quite easy)
[00:56] <kshitij8> i'll have to read a some code for threeway merge.
[00:56] <sarnold> aa-enforce, aa-complain, aa-disable, aa-audit, might be nice, smaller tools to build o the work of the module..
[00:56] <kshitij8> maybe that'll make for a good weekend read.
[00:57] <cboltz> sarnold: yes, good idea
[00:57] <kshitij8> i dont think i'll start mergeprof right now.. sarnold's idea looks good! :)
[00:57] <cboltz> basically those small tools mean set/remove a flag and write the modified profile
[00:58] <cboltz> which reminds me of the "write" part ;-)
[00:58] <kshitij8> ohh I forgot. I have a bunch of interesting fixes
[00:59] <kshitij8> comments one specially. and the vim file. mean i have some work this week.
[01:00] <cboltz> sounds like my hint to have a look at create-apparmor.vim.py was useful ;-)
[01:00] <kshitij8> i'll get back on the review soon. :)
[01:01] <kshitij8> s/mean/means/
[01:02] <cboltz> for writing the profiles - how do you like John's idea to
[01:02] <kshitij8> I'll encode new strings to avoid user data in testcases in a commit or two.
[01:02] <cboltz> a) by default, preserve profile order, comments etc.
[01:03] <cboltz> b) have a commandline switch to "cleanup" the profile
[01:03] <cboltz> ?
[01:03] <kshitij8> yes, thats on of the interesting things I have to work this week with.
[01:03] <kshitij8> switch to cleanup the profile means?
[01:03] <kshitij8> remove all comments, reorder entries?
[01:04] <cboltz> correct
[01:04] <cboltz> basically as if there was no existing profile
[01:04] <kshitij8> that'd be fairly easy i guess.
[01:04] <sarnold> reorder and collapse, anyway; maybe keep the comments near the nearest line, or shove them all up top if nearest-line is too ugly still?
[01:05] <kshitij8> we've previously dealt with maintaining comments and order thing. wonder if same trick will still work here
[01:05] <kshitij8> cboltz: what would you guess?
[01:06] <cboltz> basically it will work
[01:06] <cboltz> but it's a bit more difficult because you don't have one big array with the profile
[01:06] <cboltz> you have one for file rules, one for cababilities, one for network etc.
[01:07] <kshitij8> sounds fun.
[01:07] <cboltz> nevertheless, you'll probably be able to reuse some code from writing config files
[01:08] <kshitij8> maybe i'll  try to port out some common code to a module.
[01:08] <kshitij8> :)
[01:10] <cboltz> as much as I like to avoid code duplication - in this case it could be hard to find more than 10 lines of common code ;-)
[01:10] <cboltz> you'll have lots of similar code, but not exactly the same
[01:13] <kshitij8> I too expect little common code seeing the size of config writer.
[01:14] <kshitij8> but having some general code could be useful in some other tool. maybe mergeprof?
[01:15] <cboltz> mergeprof will of course use the same code to write the merged profile
[01:15] <cboltz> but that still has nothing to do with writing config files ;-)
[01:18] <cboltz> besides that, anything else we need to discuss?
[01:19] <kshitij8> not that I can think of.
[01:20] * kshitij8 wonders when john will be back.
[01:20] <cboltz> sarnold: ^^^ ?
[01:20] <sarnold> probably tomorrow'
[01:20] <kshitij8> btw does his mail filters have something to filter in mails that have @John mention or something?
[01:21] * cboltz guesses it's done by /dev/eye
[01:21] <sarnold> kshitij8: I don't think so
[01:21] <sarnold> @name is very rare in our field..
[01:22] <kshitij8> ohh.. just had a thought. 
[01:22] <cboltz> hmm, maybe I should add more @John in my reviews to force him to read them? ;-)
[01:23] <sarnold> heh, mutt colors the ### in such a way that it _discourages_ reading them
[01:23] <kshitij8> cboltz: maybe next week you can bring in your pending items list for me to have a look. :)
[01:23] <sarnold> .. since it presumes you're three levels of quoting deep at that point, and don't need to read it. interesting content is unquoted, you know? :)
[01:23] <cboltz> sarnold: would a single # be better?
[01:24] <sarnold> cboltz: I'd personally prefer if you prepended a "> " to every line of code, and then just write without any prefixes. but maybe I'm alone in my cranky ways. :)
[01:25] <cboltz> I'd like to keep the diff style (and also the diff coloring in $EDITOR)
[01:25] <cboltz> which means prepending > won't really work
[01:25] * kshitij8 laughs at mutt's tagline: "All mail clients suck. This one just sucks less."
[01:27] * cboltz finally got kmail un-sucking - for some unknown reason, it had created two outbox folders end then entered an endless event loop
[01:30] <kshitij8> ohh the resource-hungry bug resolved?
[01:30] <cboltz> yes, finally
[01:30] <kshitij8> on its own?
[01:30] <kshitij8> or did you have to spell out some charms. :P
[01:30] <cboltz> after a debugging session on IRC with the main kmail developer ;-)
[01:31] <kshitij8> ohh..one less bug in kmail :)
[01:31] <cboltz> fixing it (as in "delete the second outbox folder") was easy after we found the problem
[01:31] <cboltz> well, strictly speaking, the bug is still there
[01:32] <cboltz> I "just" removed the duplicate outbox that triggered it
[01:32] <cboltz> (and proposed to add a check for duplicate folders ;-)
[01:33] <kshitij8> ohh.. I thought they patched it.
[01:34] <kshitij8> anyways, i cant think of anything for the meeting atm. you cboltz?
[01:34] <cboltz> I also think we can close the meeting ;-)