Aug 24 11:31 <chen> Lets get started,
Aug 24 11:32 *jpunit (~email@example.com) has joined #evolution-meet
Aug 24 11:32 <chen> I would like to start this meeting by mentioning a important phenomenon happening in India, there are very peaceful protests happening at the moment,
Aug 24 11:32 <chen> by a person named Anna hazzare who is sitting on his 9th day fast against corruption drinking just water
Aug 24 11:33 <chen> I was there the whole last week and was amazed to see the rising consciouness..
Aug 24 11:33 *adborden1 (~firstname.lastname@example.org) has joined #evolution-meet
Aug 24 11:33 <chen> Getting started with the project updates,
Aug 24 11:33 *mauricelynch (~email@example.com) has joined #evolution-meet
Aug 24 11:34 <chen> we had the 3.1.5 release last week and would be having another release coming monday which 3.1.90
Aug 24 11:34 <chen> the freezes which are alive at the moment are,
Aug 24 11:34 <chen> API, ABI, Feature freezes,
Aug 24 11:34 <chen> and we are on String, UI announcement period
Aug 24 11:34 <chen> Some freezes dates to keep in mind,
Aug 24 11:34 *vibs has quit (Connection reset by peer)
Aug 24 11:34 <chen> Aug 29 UI freeze
Aug 24 11:34 <chen> Sep 05 String freeze
Aug 24 11:34 <chen> Sep 19 Hard code freeze
Aug 24 11:35 <chen> We had to finalize on the release date for 3.0.3
Aug 24 11:35 <chen> i would like to propose the date aug 31 for the same,
Aug 24 11:35 *vibs (~firstname.lastname@example.org) has joined #evolution-meet
Aug 24 11:35 <chen> i would like fix the following bug for the same,
Aug 24 11:36 *mcrha is fine with the date for 3.0.3
Aug 24 11:36 <chen> mbarnes, ?
Aug 24 11:36 <chen> andre_, i hope ur fine with it?
Aug 24 11:37 <andre_> oh sure :)
Aug 24 11:37 <mbarnes> chen: worksforme. I shall update the release assignments on the wiki
Aug 24 11:37 <chen> mbarnes, cool, thanks!
Aug 24 11:38 <chen> mbarnes anounced another crucial update for evolution landing soon during this release - CamelSettings,
Aug 24 11:38 <chen> mbarnes, i hope it wouldn break any freezes?
Aug 24 11:39 <chen> lakhil, any updates on the idea you had proposed on the collecting the usage statistics iirc ?
Aug 24 11:40 <mbarnes> chen: well, no one but us really consumes Camel and the platform API freeze doesn't really apply
Aug 24 11:40 <chen> mbarnes, cool
Aug 24 11:40 <lakhil> nope, i haven't got any response from banshee developer, will ping him again
Aug 24 11:40 <chen> lakhil, thanks
Aug 24 11:41 <chen> Evolution ews is also supported now for 2.28 with a branch gnome-2-28
Aug 24 11:42 <chen> it was specially done for sled, but if there are any others interested, they can consume the same
Aug 24 11:42 <mbarnes> hmm...
Aug 24 11:42 <chen> am yet to start to port it to Evolution 3.2 though
Aug 24 11:42 <chen> i made an effort to form a single compatibility patch for all versions,
Aug 24 11:43 <chen> but due the camel-object -> gobject changes which happened in 2.30 wasn able to,
Aug 24 11:43 <chen> as the code started to look ugly
Aug 24 11:43 <mbarnes> eek, that sounds incredibly painful
Aug 24 11:43 <chen> yeah, lots of typing to do :)
Aug 24 11:44 <mbarnes> I don't know how cooly does it...
Aug 24 11:44 <chen> will try to get a single clean compatibility patch for other versions
Aug 24 11:44 <chen> Anything to be discussed before we move on to individual updates?
Aug 24 11:45 <chen> updates from my side,
Aug 24 11:45 <mcrha> line wrapping discussion?
Aug 24 11:45 <chen> yup,
Aug 24 11:45 <chen> mcrha, had a chance to go through http://stackoverflow.com/questions/2151836/c-coding-style-line-length-wrapping-lines ?
Aug 24 11:45 <chen> my observation from that is,
Aug 24 11:45 <mcrha> yup, I didn't find any special argument for shorter lines there, they also said there that it's mostly per-person preference, thus the best to make a consensus between us and update the coding style page, as chen suggested the last time. Only do not take me wrong, I'm also wrapping some long lines, that's only I do not wrap on 80 letters anymore, even my laptop LCD is showing large enough screens to have wrapping on 80 letters just waste of space.
Aug 24 11:45 *chen reads the long sentence :)
Aug 24 11:46 <chen> mcrha, i felt 120 max should be reasonable ?
Aug 24 11:46 *vibs has quit (Read error: 131 (Connection reset by peer))
Aug 24 11:46 <mcrha> count in that also 8 letters per tab, which limits line length quite quickly
Aug 24 11:47 <mcrha> hmm, mine console has 174 at the moment
Aug 24 11:47 *vibs (~email@example.com) has joined #evolution-meet
Aug 24 11:47 <chen> mcrha, same as mine :)
Aug 24 11:47 *vibs has quit (Remote closed the connection)
Aug 24 11:48 <chen> mcrha, but shouldn we consider viewing multiple files in the same page ?
Aug 24 11:48 *vibs (~firstname.lastname@example.org) has joined #evolution-meet
Aug 24 11:48 *mbarnes still strongly prefers 80 chars
Aug 24 11:48 <mcrha> it's not like "full stop on max line length", is it? because sometimes it's really better readable when it's long line
Aug 24 11:48 <mbarnes> short, simple lines are easier to read
Aug 24 11:48 <chen> mbarnes, i agree
Aug 24 11:48 <mcrha> mbarnes, do you remember the change I gave a link to, where was shown just the opposite?
Aug 24 11:49 <mcrha> short lines, actually harder to read?
Aug 24 11:49 <chen> mcrha, i actually tried writing a patch with longer lines and it really looked bad
Aug 24 11:49 *mcrha votes for 140 soft-max
Aug 24 11:49 <chen> i had to commit a incremental patch later to make it look good :)
Aug 24 11:50 <mbarnes> so you're suggesting this is readable?
Aug 24 11:50 <mcrha> yeah, it depends, both cases on both sides, I agree.
Aug 24 11:50 <chen> hehe
Aug 24 11:50 <rgs> haha - thats insane
Aug 24 11:51 <chen> mbarnes, i guess the function declarations look nice with one argument per line
Aug 24 11:51 <mcrha> reason for that is that it doesn't bring in any extra information being on multiple lines, because half of the page would be just arguments of the function ;)
Aug 24 11:51 <mbarnes> chen: yes, and my cleanup script can now automate that so we're consistent everywhere
Aug 24 11:51 <chen> mcrha, but it looks good :)
Aug 24 11:52 <chen> mbarnes, i just wonder how devhelp would show it ?
Aug 24 11:52 <chen> oh sorry
Aug 24 11:52 *vibs1 (~email@example.com) has joined #evolution-meet
Aug 24 11:52 <chen> bad question
Aug 24 11:52 *vibs has quit (Read error: 131 (Connection reset by peer))
Aug 24 11:52 <mbarnes> chen: not relevant to devhelp, it does its own formatting regardless of ours
Aug 24 11:52 <chen> mbarnes, yup
Aug 24 11:53 <chen> mcrha, maybe for function declaration and definition which are exposed
Aug 24 11:53 <chen> we could have one line per argument
Aug 24 11:53 <chen> ?
Aug 24 11:53 <mcrha> chen, mbarnes, that example you gave is that way for easier eye-catching, as that was just a monkey work to write those stubs. At least for me.
Aug 24 11:53 <chen> for the rest we could have a longer limit?
Aug 24 11:54 <mcrha> can be. I'll format ema myself, once I get to it ;)
Aug 24 11:54 <chen> good, seems like we are getting a concensus :)
Aug 24 11:55 <chen> mcrha, we could have 120 as a mid-ground and leave it to people for the rest, but make sure code looks nice :)
Aug 24 11:55 <chen> for others too :)
Aug 24 11:56 <mcrha> ok, so mbarnes will run his script and that's it, right? will mbarnes increase line length for function calls?
Aug 24 11:56 <chen> mcrha, oh increase the line length, i wonder if its possible through script?
Aug 24 11:56 <chen> mbarnes, ?
Aug 24 11:57 <mcrha> chen, just the limit, not joining back already wrapped lines
Aug 24 11:57 <mbarnes> the script can enforce one-param-per-line in declarations. it doesn't touch function calls.
Aug 24 11:58 <mcrha> mbarnes, do you line-wrap them manually?
Aug 24 11:58 *Corsac (~firstname.lastname@example.org) has joined #evolution-meet
Aug 24 11:58 <Corsac> hey
Aug 24 11:58 <mbarnes> usually, yes
Aug 24 11:58 <chen> hey Corsac
Aug 24 11:58 <mcrha> mbarnes, oh
Aug 24 11:58 *dave_largo (~email@example.com) has joined #evolution-meet
Aug 24 11:58 <chen> hey dave_largo
Aug 24 11:59 <dave_largo> Hey guys,just back from lunch
Aug 24 11:59 <mcrha> anyway, let's move on, when talking about mbarnes' script, though I wanted to speak about this anyway: white space and coding style cleanups
Aug 24 11:59 <chen> cool
Aug 24 11:59 <chen> anything else before we move to individual updates?
Aug 24 11:59 <mcrha> I would like to suggest to do that only once per cycle, the best around .91/.92, to have less work with modifying patches for stable release when done on development version
Aug 24 12:00 <mbarnes> that's fair
Aug 24 12:00 <chen> yup
Aug 24 12:00 <chen> updates from my side,
Aug 24 12:00 <mcrha> personally, I thought it'll be a one time change, though such changes are more and more ;)
Aug 24 12:00 <chen> :)
Aug 24 12:01 <chen> mostly fixing bugs and porting ews to 2.28 branch,
Aug 24 12:01 <chen> and now fixing some bugs specific to 2.28
Aug 24 12:01 *mcrha is getting hands off for 2.32 and chen is divin into 2.28. Interesting
Aug 24 12:01 <chen> andre_, am yet to test ur patch, will get it done by tomorrow, was held up in meetings today..
Aug 24 12:02 <andre_> np
Aug 24 12:02 <chen> hehe /me envies mcrha :D
Aug 24 12:02 <chen> adborden1, anything to share?
Aug 24 12:02 <dave_largo> 2.28 is of interest to the corporate customers :)
Aug 24 12:02 <adborden1> nope, just listening :)
Aug 24 12:02 <chen> adborden1, ok :)
Aug 24 12:02 <chen> andre_, anything to share ?
Aug 24 12:03 <andre_> I assume I'm de-facto now Evo docs maintainer and would like to introduce myself to the Evo team but I didn't prepare anything. Maybe next time. :-P
Aug 24 12:03 <andre_> There's still enough TODOs in the files (some need developer input) and I welcome any bug reports against Evo's "User documentation" component in bugzilla.gnome.org
Aug 24 12:03 <andre_> And lakhil has already pointed out about 20 mistakes he found - thanks a bunch!
Aug 24 12:03 <lakhil> welcome andre_ , hard work was done by you :)
Aug 24 12:03 <chen> lakhil, +1
Aug 24 12:03 <chen> andre_, thanks :)
Aug 24 12:04 <andre_> it was interesting. :-P
Aug 24 12:04 <chen> Corsac, anything to share?
Aug 24 12:04 <chen> andre_, i should now inform this update to radhika to get back at you :D
Aug 24 12:04 <chen> dave_largo, anything to share?
Aug 24 12:04 <andre_> chen, :-D we had a quick talk when I was in Bangalore
Aug 24 12:05 <dave_largo> The only thing that comes to mind is that I'd like to explore some ideas for a stronger UI roadmap.
Aug 24 12:05 <chen> dvratil, updates?
Aug 24 12:05 <chen> andre_, :)
Aug 24 12:05 <dvratil> yeah, regarding the webkit port
Aug 24 12:05 <dave_largo> Whereby we kick around ideas and do mockups
Aug 24 12:05 <chen> dave_largo, looks like ur back to form :)
Aug 24 12:06 <dave_largo> I was always in form, but we haven't developed anything new for a while now :)
Aug 24 12:06 <dvratil> embeding GtkWidget into WebKit is totally broken in webkit and nobody know how to fix it and I was told that the API for this will be dropped anyways sooner or later
Aug 24 12:06 <lakhil> chen, you dared to wake up sleeping tiger ;-)
Aug 24 12:06 <chen> hehe
Aug 24 12:06 <mbarnes> dvratil: well shit. that's kinda important...
Aug 24 12:06 *vibs (~firstname.lastname@example.org) has joined #evolution-meet
Aug 24 12:06 <dvratil> yop, I'm rewriting the EMFormat & co. so that it split into multiple webviews
Aug 24 12:06 <dave_largo> We had someone log on the other day mentioning that highlighting was not supported in email, which is a part of the webkit thing
Aug 24 12:07 <mbarnes> dvratil: urg. srag strongly recommended against that after trying it with Anjal...
Aug 24 12:07 <mbarnes> but I don't know what else to suggest if embedded widgets aren't supported
Aug 24 12:07 <dave_largo> Outlook is sending email now with sections highlighted, an attribute similar to bold or italic
Aug 24 12:09 <dvratil> mbarnes, but the split-webview thing works already a little bit, I think I'm on the right track
Aug 24 12:09 <dave_largo> chen, lahkil, let's have a short Largo meeting after we are done.
Aug 24 12:09 <chen> dave_largo, ok
Aug 24 12:10 <dvratil> I'll clean up the code and fix some ugly crashes, push the recent code to git and send a mail with full description of all the evil things I've done to EMFormat on Friday
Aug 24 12:10 <mbarnes> dvratil: printing may be an issue with split-webview
Aug 24 12:10 <chen> would it be good to discuss with srag as well maybe to check if there are any other issues he faced?
Aug 24 12:11 <dvratil> for printing I'd create a separate single-webview generator, we don't need to print the EAttachmentBar and similar widgets, do we?
Aug 24 12:11 <Corsac> chen: Debian recently switch from 2.32 to 3.0.2 so we expect various bug reports which we'll try to forward
Aug 24 12:11 <Corsac> nothing new other than that :)
Aug 24 12:12 <lakhil> Corsac, you will get 3.0.3 with some fixes soon :)
Aug 24 12:12 <mbarnes> dvratil: I'll wait for your mail before commenting further. don't want to derail the meeting :)
Aug 24 12:12 <chen> Corsac, i just mentioned during the start of the meet,
Aug 24 12:12 <dvratil> mbarnes, sure :)
Aug 24 12:12 *vibs1 has quit (Ping timeout: 600 seconds)
Aug 24 12:12 <chen> that we would be making the 3.0.3 release on aug 31
Aug 24 12:12 <chen> if u have missed
Aug 24 12:13 <chen> Corsac, plz let me know if u have any critical issues that needs to be fixed for 3.0.3..
Aug 24 12:13 <chen> and we can add it to the same
Aug 24 12:13 <chen> dwmw2, updates?
Aug 24 12:13 <chen> jpunit, updates?
Aug 24 12:14 <jpunit> chen: worked on downstream mostly on submission of sp2 and some bugs..
Aug 24 12:14 <chen> kov, anything to share?
Aug 24 12:14 <chen> jpunit, thanks
Aug 24 12:14 <chen> lakhil, updates?
Aug 24 12:14 <lakhil> ews testing in master and downstream, random testing on master, evolution user doc review and bug triaging
Aug 24 12:14 <chen> lakhil, thanks!
Aug 24 12:15 <chen> mauricelynch, anything to share?
Aug 24 12:15 <chen> mbarnes, updates?
Aug 24 12:15 <mbarnes> chen mentioned the new settings API for Camel
Aug 24 12:15 <mbarnes> that was a strategic move to remove some roadblocks in my account-mgmt branch
Aug 24 12:15 <mbarnes> which I'm now dusting off and trying to get running again after several months of neglect
Aug 24 12:15 <mbarnes> I'm at the point where I'm trying to merge mail accounts into the new ESource API
Aug 24 12:15 <mbarnes> so that's my only real focus now, as I'm trying to finish it in time for 3.3.1 or .2
Aug 24 12:15 <chen> mbarnes, thats awesome :)
Aug 24 12:16 <chen> mbarnes, unified account management on the way :)
Aug 24 12:16 <chen> mcrha, updates?
Aug 24 12:16 <mcrha> only simple work-through more-or-less recent bugs, catching things before 3.1.90 release.
Aug 24 12:16 <Corsac> chen: I myself had issues with the sidebar size (not critical, already reported) and few times evo wouldn't close or switch to offline mode
Aug 24 12:16 <mauricelynch> Sorry in another mtg with my managers, and trying to debug something with David
Aug 24 12:16 <chen> mcrha, thanks1
Aug 24 12:16 <Corsac> maybe related to network changes while suspended
Aug 24 12:17 <chen> Corsac, ok, let me know the bug id for the second issue.. i will try to make a fix
Aug 24 12:17 <chen> mauricelynch, np
Aug 24 12:17 <chen> rgs, anything to share?
Aug 24 12:17 <rgs> Folks 0.6 was finally released - with a shiny e-d-s backend which is being used from Gnome Contacts, on the way to this release e-d-s' introspection support received some attention, so i hope to see new and wonderful projects using e-d-s' introspected binding
Aug 24 12:17 <mauricelynch> We are close to a EWS Beta release, just as an update
Aug 24 12:17 <chen> mauricelynch, nice :)
Aug 24 12:18 <chen> rgs, nice :)
Aug 24 12:18 <Corsac> ok, in fact I just closed two evo on two different boxes, and it freezes at quitting
Aug 24 12:18 <chen> rodrigo, anything to share?
Aug 24 12:18 <Corsac> so there's definitely something fishy :)
Aug 24 12:18 <mcrha> Corsac, imapx in use?
Aug 24 12:18 <Corsac> yes
Aug 24 12:19 <chen> vibs, updates?
Aug 24 12:19 <Corsac> ha, first one finally closed
Aug 24 12:19 <Corsac> and second too
Aug 24 12:19 <Corsac> so it's just really slow
Aug 24 12:19 <chen> mcrha, is it available in stable branch as well ?
Aug 24 12:19 <chen> s/it/the patch
Aug 24 12:20 <mcrha> Corsac, should get to your nearest hardware store and give them some money for better machine ;)
Aug 24 12:20 *vibs was mostly on downstream
Aug 24 12:20 <chen> vibs, thanks
Aug 24 12:20 <mcrha> chen, nope, 3.1.90+ says the last comment there
Aug 24 12:20 <vibs> and was on leave
Aug 24 12:21 <chen> mcrha, can you commit to stable as well?
Aug 24 12:21 <mcrha> by the way, when I was chasing these pointers, I extended e_pointer_tracker routines from e-data-server-util.c, it can show backtraces where the pointer was added to the tracker, which is kinda useful
Aug 24 12:21 *vibs has quit (Read error: 104 (Connection reset by peer))
Aug 24 12:21 <andre_> dvratil: I know I asked this before, but hearing about your issues I assume that WebKit is now targetted for evo 3.4 instead of 3.2 as we'd expect too many rough edges?
Aug 24 12:21 <mcrha> chen, I'm not sure, it didn't get enough testing, neither proper review
Aug 24 12:22 <mbarnes> andre_: no chance in hell we'd merge that for 3.2 at this late date
Aug 24 12:22 <dvratil> andre_, definitely not making it to 3.2 :)
Aug 24 12:22 *mauricelynch has quit (Remote closed the connection)
Aug 24 12:22 *mauricelynch (~email@example.com) has joined #evolution-meet
Aug 24 12:22 <andre_> thanks. just need it for my planning :-P
Aug 24 12:22 *vibs (~firstname.lastname@example.org) has joined #evolution-meet
Aug 24 12:22 <chen> mcrha, ok, will have a look at it and port it to gnome-3.0 branch
Aug 24 12:22 <mcrha> good chen
Aug 24 12:22 <dvratil> andre_, after three weeks of work I managed to display single plaintext email two days ago :)
Aug 24 12:22 <andre_> dvratil, heh
Aug 24 12:23 <chen> andre_, sure!
Aug 24 12:23 <vibs> hi chen, my connection is getting disconnected often
Aug 24 12:23 <vibs> last month I tried hands on EWS addressbook and downstream bugs
Aug 24 12:23 <chen> vibs, cool
Aug 24 12:24 <mbarnes> chen: have you heard anything from srag about how this email-factory branch of his is progressing?
Aug 24 12:24 *mcrha noticed rumors of abandoning the work on it
Aug 24 12:24 <chen> mbarnes, i had a discussion with him couple of weeks back and he said,
Aug 24 12:24 <chen> he wanted to push it for 3.4 release
Aug 24 12:24 <Corsac> mcrha: hmhm, it takes ages to quit even just after starting
Aug 24 12:25 <chen> he said he wanted the api changes to slow down and he could merge the changes for 3.4
Aug 24 12:25 <mcrha> Corsac, if there is anything in the status bar, then it'll "never" quit
Aug 24 12:25 <chen> mcrha, hmm, let me check with him tomorrow morning
Aug 24 12:25 <chen> to be sure
Aug 24 12:25 <mbarnes> chen: hmm. my branch and his branch aren't gonna get along well. we'll need to coordinate
Aug 24 12:26 <mcrha> chen, might be I mixed up with someone else, do not listen this insane guy writing insanely long lines ;)
Aug 24 12:26 <Corsac> nop, here the status bar is empty
Aug 24 12:26 <chen> mbarnes, i can understand :)
Aug 24 12:26 <Corsac> all in all I find imapx operations slower in 3.0
Aug 24 12:26 <chen> mbarnes, he had the concern as well, but would it have a clash even with 3.3 proceeding ?
Aug 24 12:27 <mcrha> ok Corsac, let's talk after the meeting
Aug 24 12:28 <chen> mbarnes, maybe you can discuss with him tomorrow on it and keep track of things..
Aug 24 12:28 <chen> anything else before we wrap up the meeting ?
Aug 24 12:28 *mcrha has nothing more
Aug 24 12:29 <mbarnes> chen: he's working against the 2.32 branch, so he's got some adaptations to make regardless
Aug 24 12:29 <chen> mbarnes, its better he gets the code merged once we branch for 3.3
Aug 24 12:30 *dwmw2 returns; sorry
Aug 24 12:30 <mbarnes> agreed
Aug 24 12:30 <chen> lets check with him and i hope he has the same plan
Aug 24 12:30 <dwmw2> update: I've mostly been ignoring EWS for the last few weeks and concentrating on the ActiveSync back end
Aug 24 12:30 <dwmw2> which is working well for mail; I've been using it in anger on my company mail for the last couple of months in fact.
Aug 24 12:30 <chen> dwmw2, thanks!
Aug 24 12:30 <chen> dwmw2, anger ;)
Aug 24 12:30 <dwmw2> I haven't done calendar/addressbook back ends; we're doing that through SyncEvolution atm although it would be nice to get an evo backend for htem.
Aug 24 12:31 <dwmw2> I *really* want to fix the config ui stuff for camel providers. Having to do an evolution ui plugin just to configure the basic account settings is a PITA
Aug 24 12:31 <dwmw2> and if we fix that, all the groupware stuff could actually move into e-d-s where it belongs. Because it wouldn't need to link against evo itself.
Aug 24 12:31 <chen> dwmw2, agree
Aug 24 12:32 <dwmw2> I reckon we can get away with putting the ActiveSync code into eds too; the patents don't bother us.
Aug 24 12:32 <dwmw2> the allegedly patented stuff is in the separate ActiveSyncd dÃ¦mon, which the camel provider talks to by dbus
Aug 24 12:33 <dwmw2> (and anyone clainming to have a patent on that should be jailed for fraud anyway, but that's a different issue)
Aug 24 12:33 <chen> dwmw2, i wonder what evolution-ews backend and activesync backends are targetted for, the question is
Aug 24 12:33 <chen> were does active sync find the use ?
Aug 24 12:33 <mbarnes> dwmw2: I'm convinced the whole CamelProviderConf crap needs to die. we need a better framework.
Aug 24 12:33 <dwmw2> ews has more features, is generally better and faster. But you have to be on the VPN
Aug 24 12:34 <dwmw2> ActiveSync is available anywhere
Aug 24 12:34 <dwmw2> mbarnes: makes sense.
Aug 24 12:34 <dwmw2> I am still seeing Evolution build up to 5GiB or so
Aug 24 12:34 <dwmw2> I found some gobject leaks that valgrind wasn't complaining about
Aug 24 12:34 <dwmw2> the biggest in EWS, which I fixed. Others in the composer I think.
Aug 24 12:35 <dwmw2> but I can't actually find the main offenders
Aug 24 12:35 *dave_largo hates leaks on multi-user servers!
Aug 24 12:35 <dwmw2> doesn't seem to be objects, doesn't seem to be shown by valgrind even when I run with --show-reachable.
Aug 24 12:35 <dwmw2> perhaps we mmap() some space that isn't accounted for by what I've been using?
Aug 24 12:35 <chen> maybe memprof would help ?
Aug 24 12:35 <dwmw2> I will keep poking at it
Aug 24 12:36 <chen> cool
Aug 24 12:36 <dwmw2> 16478 dwmw2 20 0 4177m 1.8g 19m S 0.0 47.9 6:25.78 evolution
Aug 24 12:36 <dwmw2> starts up at 2Â½G, grows until I kill it.
Aug 24 12:36 <chen> dwmw2, bad
Aug 24 12:36 <dwmw2> yes
Aug 24 12:37 <dwmw2> I want something like the kernel's /proc/slabtop where I can see how many objects of various sizes are allocated
Aug 24 12:37 <dwmw2> will memprof do that? Or do I need to hack the gslice allocator?
Aug 24 12:37 <dwmw2> valgrind should have done it, I suppose, but didn't seem to.
Aug 24 12:37 <chen> dwmw2, it would show which objects are holding all the space
Aug 24 12:38 <chen> dwmw2, i used it long time back to figure out a memory buildup in libical
Aug 24 12:38 <mcrha> dwmw2, do you have suspicion on a concrete object?
Aug 24 12:38 <dwmw2> 374 new GtkButtons every time I open a compose window and close it :)
Aug 24 12:39 <dwmw2> but I don't think that accounts for the huge sizes I'm seeing.
Aug 24 12:39 <dwmw2> it grows even when I'm not in front of it.
Aug 24 12:39 <mcrha> dwmw2, hmm, for those from eds/evo I used e_pointer_tracker from e-data-server-utils.h to track where they were allocated (it can show backtraces too)
Aug 24 12:41 <dwmw2> cool, thanks. I'll play
Aug 24 12:42 <dwmw2> see what I can find
Aug 24 12:42 <chen> cool
Aug 24 12:42 <dwmw2> would like to use talloc-like allocations
Aug 24 12:42 <dwmw2> belong to an object
Aug 24 12:42 <mcrha> dwmw2, ./configure --enable-backtraces
Aug 24 12:42 <dwmw2> go away when obj destroyed
Aug 24 12:42 *mcrha hates talloc
Aug 24 12:42 <dwmw2> ok, thanks
Aug 24 12:43 <dwmw2> I'm dubious about it, but if it helps reduce ;leaksa... :)
Aug 24 12:43 <mcrha> talloc is too transparent, one can access already freed memory even not knowing that
Aug 24 12:44 *lakhil has quit (Leaving)
Aug 24 12:44 *mcrha gotta go
Aug 24 12:44 <mcrha> chen, is meeting over?
Aug 24 12:44 <mcrha> ;)
Aug 24 12:44 <chen> Yup :)
Aug 24 12:44 <chen> lets then meet next month same time :)
Aug 24 12:44 <mcrha> good, see you another day. bye bye
Aug 24 12:44 <chen> have fun :) thanks everyone :)