=== asac_ is now known as asac === Kopfgeldjaeger is now known as Kopfi|offline === ember_ is now known as ember [01:46] night. === superm1|away is now known as superm1 [02:09] <__iron> hi [02:10] <__iron> could it be that linux ram-management have any failure ? [02:17] __iron: Of course. But I don't think this is the place to raise whatever question it is that you'll eventually raise. [03:07] http://www.joshuazeidner.com/2008/02/ted-gould-svg-inkscape-and-web.html [03:32] A few things about installing from debootstrap. 1 - root is enabled and has no password by default. 2 - no default users, no prompt to create one 3 - in kubuntu install, there seems no sound mixer backend [03:40] * genii puts on a pot of coffee before he leaves [04:43] gee I need to get more involved in the fixing bugs or sending patches. :/ [04:47] wolfe: If you're bored, and want work to do, #ubuntu-motu is a good place to ask for tasks. [04:47] ^_^ thanks [04:48] I can't think of anything ubuntu needs program/gui wise :P since I know the the people working on said lacking programs already. [04:49] wolfe: Well, personally, I'd like to see better drivers for Brain-Computer interfaces, and I don't know that anyone is working on a GUI to help tune those. Mind you, the drivers need a lot of work first. [04:58] persia: i have the actual IO working, but the calibration tools are somewhat lackluster [04:59] "OH GOD! DON'T REFORMAT! I just wanted pancakes!" [05:01] See, it's all about having good drivers, a good set of defaults, and some interface to help the user to avoid that sort of confusion. [05:01] :) [05:02] * mneptok tootles off for home [05:16] persia: send me a brain-computer interface and I'll work on drivers :P [05:17] wolfe: That works. Please email me your address and a quick summary showing that you can write input drivers. I'll send it in about three weeks. [05:17] heh, no exp sorry :P [05:18] In that case, buy your own :p [05:18] how much are the related devices? [05:18] The good ones are a couple thousand euro, but you can also get kits for about 300 euro. [05:18] Mostly serial interface, but hacking USB for a kit oughtn't be that bad. [05:19] wouldn't a person want the highest quality interface device when writing a driver? [05:19] There's some talk about consumer devices coming out soon for a couple hundred euros, but some of them are skin galvanic detectors, so really process muscle changes rather than brainwaves. [05:20] Probably. If anyone ever takes me up on the offer of writing a driver for one that works cleanly with X, I'd send them one of the ones approved for medical use in Germany. [05:22] I'm guessin they all only work with Windows? :) [05:22] Most of them only work with specialised software right now. There's a couple generalised drivers for Windows (old versions) for mouse pointing and MIDI, and some simple linux drivers for MIDI. [05:23] On the other hand, nobody seems to really believe that the devices exist, so most of the drivers are either made for specific sales niches (e.g. OCS's new skin galvinometer for improved game play in Vista) or research purposes. [05:25] wouldn't the drivers need to be a type of processing engine? Do the devices send a converted data stream of raw data? [05:27] Most of the devices seem to send some sort of waveform signal, so one has to do DSP, etc. to translate this into something useful. [05:27] As different devices have a different number of sensors, they provide multiple streams of data. [05:27] so its more than just knowing input device drivers... :) [05:28] There are a couple standards for the data delivery, but nothing yet adopted industry-wide. [05:29] Well, yes, but I figure anyone who actually wants the project is likely to ask useful questions and look up the devices a bit :) Also, there are plenty of good tools for pulling digital signals from analog waveforms already available in Ubuntu. [05:30] Also, it's a learning process: one probably doesn't want to trigger a mouse movement up based on the user's thought "up", but rather on some arbitrary signal, and to have a training GUI that helps users learn how to move the mouse, type, set focus to windows, etc. [05:31] As there's already lots of code that does the basework, it's getting the existing drivers (of various sorts) mangled into the linux input event framework and putting together a training GUI that I see as the main problems remaining to be solved. [05:32] As an example, if one looks at the MIDI driver, that controls 128 independent 7-bit axes, along with 128 possible event triggers, each with a corresponding optional 7-bit value (ignoring tricky stuff one can do with MIDI). [05:33] This is more than one needs to handle a few mouse axes (W,X,Y,Z,Rw,Rx,Ry,Rz are all that the kernel defines), and some keys. [05:34] *grin* you know it would be quite odd if someone were able to think a tune with an instrument selection and have the music pop on the screen. [05:34] Supposedly, that works now, if one configures denemo and lilypond correctly for MIDI input, but I suspect it's a bit buggy. [05:35] sorry, you said MIDI, I've been dinking around with making a video :P I hate not having a real 83+ keyboard [05:36] persia: if someone wrote a driver, apple would take it and package it all up for use with Garageband and their other software, heh. [05:36] >:/ [05:36] and I bet they wouldn't contribute anything back [05:36] Well, that hasn't happened yet. [05:38] persia: apple uses lots of open source software in osx though. When the main developer said "no" to Apple's request to relicense the libntfs library, that story seemed to make it large. Some people even said nasty things about the developer because he said no >:/ [05:38] since the libntfs filesystem library seems to be GPL [05:40] Apple uses the BSD license [05:40] Big Scary Duopoly [05:40] * wolfe likes the BSD license, but I like when people well... give back? :) [05:40] wolfe: I don't really follow apple development, but I know that nobody ships BCI drivers with the default OS. [05:41] I mean, sure... Apple contributes greatly to GCC. I dont see too much of apple giving app code out unlss its samples. [05:41] or WebKit [05:41] persia: brain comptuer interface? [05:41] wolfe: Yes. [05:42] [of course, with WebKit the original software license kind of forced their hand] [05:44] Nifty. Google just told me about a 16-lead EEG system with USB for only 1200 USD (http://www.contecmed.com/cart/product_show.asp?id=47). Perhaps I have the pricing wrong (because I was only looking at prebuilt headsets). [05:45] persia: isn't there anything well.. headset like around that price? [05:45] wolfe: I've not seen it actually shipping yet, but I don't follow as closely as in the past. [05:45] I suppose people would still have to use electrode glue on a headset as well to get the best contact possible.. [05:46] This is probably why skin galvinometers are becoming more popular. [05:47] persia: you don't want to shave your head bald to use an input device so you don't have to wash your hair out every time you use the computer? :) [05:48] wolfe: I suppose. I'm happy with only a couple electrodes, and I'd shave my temples and skull base for that. [05:49] * slangasek reads the last few lines of scrollback, and starts squinting upwards to check whether we're talking about tasps [05:50] No. It's brain -> computer. tasps are more about computer -> brain. [05:54] I didn't realize there was so much consumer software out there for BI devices [05:54] BCI [05:54] BioExplorer looks interesting [05:54] persia: do you know much on this device? http://www.transparentcorp.com/products/eeg/pendant.php [05:54] Yeah, there's a fair bit, but not yet something where I can plug in a list of supported devices, and use my computer without my hands. [05:55] pendant device for 600USD.. [05:55] without software or electrodes [05:55] No, I hadn't hear about that. That's well into reasonable pricing. Thanks for the link! It's only 2-electrode, but one can still do a lot with that. === thegodfather is now known as fabbioen === fabbioen is now known as fabbione [05:58] persia: seems so :) the BioEra package, the open source one, has a little game you can drive around in a car. [05:58] * persia celebrates another convert to the school of thought that DWIM oughtn't really be that hard to implement [05:59] *grin* [06:02] persia: what I'd love to see is an actual device like that pendant which can just be worn on the head but in a not so obvious fashion. I can think of a thousand different tasks I'd love to do if there was software making system administration really easy. [06:02] :( wishful thinking [06:03] wolfe: Sure, and then you think 'rm -rf /' and it executes it [06:03] wolfe: The trick is to tape the electrodes to your HMD, so nobody can see that it's not the audio feed. [06:04] good morning [06:04] (and there are "covert" HMDs that fit in mirrorshades: Oakley even sells a pair) [06:04] StevenK: I've plenty of self control, especially to not open a new email to swear at an employee and send.. [06:04] *googles* [06:05] fabbione: good morning - how are you doing? [06:06] do you think you could take a look at bug 259579 when you have the time? [06:06] Launchpad bug 259579 in redhat-cluster "cman init script points at an incorrect doc path" [Low,Confirmed] https://launchpad.net/bugs/259579 [06:11] dholbach: hey.. barely awake :) [06:11] * dholbach hugs fabbione [06:11] * dholbach gets more coffee himself [06:12] * fabbione hugs dholbach [06:15] dholbach: how often are the (sponsoring) and (patches) refreshed on Harvest? [06:15] LaserJock: you need to ask bdmurray - the harvest scripts run every hour and check if there's anything new [06:16] hmm [06:16] I was actually using it today but nothing has left the list :( [06:17] dholbach: See, this is why every set of opportunities ought have a homepage: so feed admins can be more clearly identified. [06:17] persia: I never disagreed with having it - right now I just don't the time :-/ [06:18] but it shouldn't be too hard to do [06:20] dholbach: Understood :) [06:36] halöle === superm1 is now known as superm1|away [07:18] who is archive admin today? === Hobbsee` is now known as Hobbsee [07:24] fabbione: seb according to the wiki: https://wiki.ubuntu.com/ArchiveAdministration [07:26] jpds: thanks === MacSlow_ is now known as MacSlow [07:45] ?C [07:45] ?C [07:45] ugh stuck keys === tkamppeter_ is now known as tkamppeter [09:35] ArneGoetje: can you please disable the cron job that updates langpacks? ... tomorrow there will be a change that will break mozilla translations - which Ill have to fix first. [09:39] asac: What does NM mean by "Automatic (VPN) addresses only"? === DreamThief is now known as LameSucker [09:41] asac: it is disabled. the current ones are still in -proposed === WhiskeySour is now known as DreamThief === LameSucker is now known as WhiskeySour [09:48] dholbach: Hey.. Could you test another kvm thing for me? [10:01] soren: sure [10:09] Riddell: is it a known problem that the menus of QT4 applications are not visible? See this screenshot: http://albertomilone.com/menus.jpg [10:09] Riddell: I'm using GNOME [10:12] tseliot: nope [10:12] not seen that [10:12] tseliot: only time I've seen that problem was in pyKDE applications for 4.0 [10:13] but designer isn't pyKDE [10:14] Riddell: I can reproduce the problem only with QT4 apps (such as eric) but not with KDE4 apps (such as Kompare) [10:16] tseliot: presumably if you run it with -style=plastique it doesn't have a problem? [10:16] Riddell: no, it's only the Oxygen style that causes the problem [10:16] Riddell: QtCurve or Cleanlooks work well === ara_ is now known as ara [10:35] asac: ooh, much win, I have a working openvpn connection :D === Tonio__ is now known as Tonio_ [10:39] I finally got VPN plugins working with NM0.7 tonight - it took a few reboots... is that intentional/known? [10:41] sounds like windows, reboot till it works :) [10:47] geser: That is why I brought it up. It was horrifyingly like Windows. [10:50] seb128, my evo shows me 5 unread mails in Trash since yesterday (there are definately none) and one in Junk since this morning ... [10:51] do you have anything open about that ? [10:51] known issue [10:51] oki [10:51] no bug upstream knows about it so no need to create launchpad noise [10:51] good, thanks [10:55] asac: hmm, so I am connected and stuff, but nm-openvpn-service is now eating a whole core :/ [10:56] it seems to be tightly looping on poll()s on some sockets [10:57] and even after disconnecting, it's still doing it :/ [10:57] oh, now it stopped [11:00] You know, I was just thinking that it would have been really neat if the first Ubuntu release were named adjective animal [11:00] * YokoZar is editing the wikipedia article [11:02] YokoZar: warty warthog wasn't adjective animal? [11:03] Ng: No, I mean "adjective animal" as a release name ;) [11:03] ah [11:03] That way we'd have started alphabetical. And been cute from the getgo [11:04] wgrant: not sure. where do you read that? [11:04] kirkland: well, I'm almost ready to upload, aside from a slight bug that causes /etc/pam.d/common-* to be filled with endless duplicated lines [11:04] Ng: cool. (not the core eating thign) [11:04] kirkland: I'll have to fix that in the morning :/ [11:04] asac: at least it works, I'll check out the CPU usage thing later [11:05] Ng: which process is eating the cpu` [11:05] ? [11:05] asac: nm-openvpn-service [11:06] strace showed it doing poll() and read() and nothing else [11:06] I was about to write "using an adjective and an animal alliteration" only to discover that that, too, is a long-ass alliteration [11:06] Ng: so either looping thing or hard-crypto? [11:06] asac: IPv4 Settings when creating a VPN. [11:07] asac: I'd be surprised if it takes a whole core of a core2duo to do the encryption and LZO. Also, doesn't "read(4, "", 1024)" imply that it didn't read anything? [11:08] wgrant: "addresses only" means that you will only get the addresses from VPN ;) [11:08] wgrant: e.g. you set dns + search domains manually [11:08] wgrant: not sure about routes [11:08] asac: Ahh. I think that could be made more clear. [11:09] (DNS servers are still addresses...) [11:09] wgrant: well. that text was just changed to that ;) [11:09] because users complained that nobody knows what a DNS is ;) [11:12] I think it should be radio buttons, as it just affects those two lower widgets. Somewhat like how Windows XP does it. [11:24] asac: that config dialog certainly isn't easy to use, but it's perfect for power users [11:24] the routes bit works fine, I can tell it to ignore automatic routes and just set one for the /24 of the VPN network with a higher metric than the default dev route [11:25] hi === dholbach_ is now known as dholbach [11:25] I am using Network Manager to connect to openvpn [11:25] every time i need to connect to openvpns server i need to add [11:25] sudo ip route add 10.0.0.0/8 via 10.10.50.12 dev tap0 [11:26] kaushal: which version of NM? [11:26] Ng: nm-applet 0.6.6 [11:28] kaushal: hmm, mine was ok when I set the "Only use VPN for these addresses" option and gave 10.8.0.0/24 in the settings for the VPN [11:28] Ng: where can i find openvpn client configs using Network Manager [11:29] kaushal: (btw this is offtopic for #ubuntu-devel). [11:30] Ng: I am aware of it [11:30] if you can help me where can i look out for help [11:40] Riddell, can yu acces wiki.kubuntu.org atm ? [11:41] (i dont get response from wiki.edubuntu.org and kubuntu seems to be the same here) [11:45] seb128: ping? [11:45] hi fabbione [11:45] hey dude [11:46] seb128: i saw you are the archive admin of the day ;) [11:46] seb128: do you have 2 minutes? [11:46] yes [11:46] awesome [11:46] seb128: main src package openais has been splitted by upstream in 2 source packages. [11:47] i have the split ready.. just need to know if: [11:47] - new source need an official SRU (code is the same as it was before the split and it's already in main) [11:47] - do you prefer a normal upload or copy from ppa? [11:47] - there are 3 new binaries, but i am sure that's not a problem at all [11:47] fabbione: s/SRU/MIR? [11:48] whops [11:48] yes MIR sorry [11:48] * fabbione is not exactly awake today [11:48] no need of a new MIR [11:48] ok [11:48] normal upload please [11:48] ok [11:48] and the new binaries are no issue ;-) [11:48] indeed ;) [11:48] * fabbione executes seb128's orders [11:48] * seb128 hugs fabbione [11:48] seb128: done with syncs for now? [11:49] Riddell: yes, it's all yours [11:49] Riddell: thanks for not doing those yesterday :-p [11:49] had to do over one hundred of those ;-) [11:50] seb128: I left you an empty new queue! [11:51] ok, that's a fair deal then ;-) [11:53] seb128: all done [11:53] Subject: New: corosync 0.91-0ubuntu2 (source) [11:53] this is the new source that will spit the 3 new packages [11:53] the other uploads (openais and redhat-cluster) will just go dep-wait till corosync is sorted [11:54] fabbione: cool [11:54] will do the newing [11:54] seb128: thanks a lot for your help [11:54] next major change is going to be the redhat-cluster SONAME [11:54] but not today [11:54] * fabbione needs to do that upstream first [12:09] seb128: hey there. Thanks for processing bug #259288, but you didn't think the latest version (and the version I requested) [12:09] Launchpad bug 259288 in wpasupplicant "Please sync wpasupplicant 0.6.4-1 (main) from Debian unstable (main)." [Wishlist,Fix released] https://launchpad.net/bugs/259288 [12:13] where can i get help regarding Network Manager [12:14] what's the correct procedure for adding a package which requires itself to compile to the archive? [12:14] ScottK: you reverted the dget change I sponsored saying that it's only needed to work around a PPA-related bug. I disagree - surely it's useful if you happen to have a URL to a .dsc on launchpadlibrarian.net, regardless of whether it's in a PPA [12:15] ScottK: and even if it's a PPA bug it seems useful to work around it for the moment [12:16] * cjwatson takes it to the bug [12:20] infinity: in particular, directhex wants to add vbnc to ubuntu (VB .net compiler, written in... VB.net)... maybe you have some hints? [12:24] meanwhile, i still need to get write access to the appropriate section of the alioth svn repo the packaging work is stored in [12:31] directhex, is this the Visual Basic mono compiler? [12:31] NCommander, aye [12:31] If memory serves, isn't the only way to build that from scratch is use a binary of the VBC, or use Microsoft's visual basic? [12:31] NCommander, aye. [12:32] (I would cite for the record that are packages in Ubuntu that require themselves to exist to be bootstrapped) [12:32] NCommander, orig.tar.gz ships with a vbnc.exe for bootstrapping [12:32] gnat, gdc, mono itself used to [12:33] Well, if its vbnc.exe, then isn't it by definition a .NET app? [12:33] yep [12:34] NCommander, i just want a definitive answer to the question "should the source package use bootstrap binaries, or should it depend on itself" - and if the latter, what's the procedure? [12:34] The later is what's used in Debian [12:34] It involves someone manually building the package for that architecture and uploading it though [12:34] *did it for GNAT on m68k) [12:34] multiverse allows for binaries [12:35] Its probably best if you package the binary of the vb compiler, with the depends on mono, and then have the mono vb package depend on that, or itself [12:35] That way it can also bootstrap itself automatically on new architectures [12:35] well it's a managed app, so it's arch:all [12:36] Right :-P [12:36] Well, you could put the binaries in multiverse, upload the compiler to universe, then request removal of the original one if the archive admins can't add a binary directory to the repo. [12:36] actually it's best to have the final packaging build-depend on itself (as soon as its bootstrapped) [12:36] That's the Debian policy [12:36] AFAIK [12:37] and in regards to bootstrapping, it's best to ask a buildd-admin for help (that's why I referred to you here) [12:37] it's *really* not hard to do either. i just need direction [12:37] i'll build both now. a self-bootstrapping and a self-depending package [12:37] Your dedication to mono is amazing ;-) [12:38] i like the technology, and i dislike the blanket "use suse" response handed out in official channels whenever debbuntu isn't behaving 100% [12:38] oersonally VB makes me vomit, but some people need it badly [12:38] siretart: fixed now [12:40] My solution is to convert VB.NET apps to C# [12:40] THere really isn't that much difference between the two [12:40] NCommander, nice solution! [12:40] NCommander, i could run monodis on vbnc.exe, then depend on ilasm... [12:40] directhex, there are automated tools that require it [12:41] er, do that [12:41] wait [12:41] what O_O? [12:41] Your going to convert the compiler to C# .... and then have it build itself?! [12:41] *twich* === Kopfi|offline is now known as Kopfgeldjaeger [12:42] NCommander, nah, i like the depend-on-self solution [12:42] then what was the nice solution? [12:42] NCommander, but someone somewhere needs vbnc in $PATH for the package to build [12:42] ... I feel like I'm missing something in context [12:42] NCommander, the nice solution was some prpopsed evil to work around shipping "binaries" [12:43] Well, you could covert the compiler to bootstrap itself, and then recompile it [12:44] directhex: I suggest that you start with the depend on self solution, get the package reviewed and we'll fiddle with bootstrapping issues once its in the archive, ok? [12:44] (once the source is in the archive even) [12:44] how can I help? [12:45] NCommander: review the package for example ;) [12:45] How do I do that O_o?, I'm no MOTU [12:46] NCommander: anyone can comment on revu (unless you broke that :P) [12:46] probably :-) [12:47] directhex: definitely don't do bootstrap binaries - talk with infinity to get a manual bootstrap done [12:50] seb128: thanks! [12:50] siretart: np, sorry for syncing the wrong version before [12:50] directhex, I was referring to if I must compile a VB.NET app on linux before the advent of the VB.NET compiler [12:51] ick. messy [12:59] NCommander: actually, rewiewing is a good training to become a MOTU [13:00] I can't comment on REVU without a package getting marked needs-work :-/ [13:00] no? [13:02] I'm a REVU admin, which also marks me an MOTU in revu's database, regardless of the truth [13:07] hmmm not really [13:07] the REVU db doesn't know about MOTU and non MOTU iirc [13:08] it knows about uploaders, reviewers and REVU admins [13:08] you're a MOTU if you belong to the MOTU group on LP [13:09] does anyone has ever seen a mann file (with 2 n)... ? [13:26] huats: tcl used to do that [13:26] huats: nowadays it tends to be man3/foo.3tcl.gz instead [13:26] huats: the list of valid man sections, technically, goes 1-9 l n o [13:26] huats: but we only use 1-9 nowadays [13:27] cjwatson: ok [13:27] cjwatson: because I am packaging a tcl lib [13:27] and I have a mann dir with a foo.n [13:28] so it is the man page :) [13:28] /usr/share/doc/man-db/man-db-manual.txt.gz documents it as "new [obsolete]" [13:28] ok [13:30] cjwatson: so do you think I might not include it ? [13:38] dpkg-deb: building package `mono-vbnc' in `../mono-vbnc_1.9+dfsg-0ubuntu1_all.deb'. [13:39] huats: it should go in /usr/share/man/man3/foo.3tcl.gz instead [13:39] cjwatson: I just replied in the bug. I'm sorry for doing something that felt abrupt. Details in the bug. [13:40] ok, thanks [13:46] ion_, hi [13:47] Hi tkamppeter [13:51] soren: FYI, that tasksel change of mine does work as far as it goes, but there's some complicated cdebconf bug that makes the display completely wrong [13:53] Ion_, I have seen your http://heh.fi/tmp/cups/01-adobe-ps. This is for the Debian SVN? [13:53] soren: (specifically, I get the text of debian-installer/dummy rather than of tasksel/first, although the text for tasksel/first in templates.dat is correct afterwards [13:56] tkamppeter: Yes, but it’s not fully working yet, because i haven’t managed to produce a .types file that recognizes the DRMed PS files as other than application/postscript. I will make a version of the patch for Ubuntu CUPS as soon as the Debian svn patch is ready. [13:57] tkamppeter: priority(200) didn’t affect it, and i didn’t get around to figuring out why yet. [13:59] cjwatson: Um.. Ok :) Sounds like fun. [14:01] ion_, for me every PS file was detected as DRMed (without your patch). [14:01] tkamppeter: Huh. Strange. [14:24] ion_, it seems that the CUPS auto-typing is somewhat buggy and we must find a workaround to distinguish between DRMed and non DRMed PS. === superm1|away is now known as superm1 === Kopfgeldjaeger is now known as Kopfi|offline [14:29] tkamppeter: Yeah [14:30] tkamppeter: One alternative would be just processing *all* postscript files with the filter. ;-) The gawk script is fast, it wouldn’t be much of a performance hit. [14:30] tkamppeter: But yeah, it would be best to fix the detection. [14:35] ion_, so we should pass all through the filter for now and report a bug against CUPS about the broken auto-detection- [14:37] tkamppeter: Well, on my box, nothing passes through it. [14:38] tkamppeter: To make everything pass through it deterministically, we probably should change the /etc/cups/mime.types application/postscript entry to say application/x-noredistillps. [14:39] ion_, more or less like that [14:41] yes exactly that, rename application/postscript to application/x-noredistillps and create a new application/postscript without autodetection rule. [14:41] Add a .convs rule converting application/x-noredistillps to application/postscript with your script [14:42] This will stay this way until the CUPS bug is fixed. [14:43] tkamppeter: I’ll update my patch, a moment. [14:48] is it just for me that firefox 3.0.1 in hardy (amd64) is hanging so much it's completely unusable? [14:48] only with nspluginwrapper, IME [14:49] no such thing installed here. [14:49] nspluginwrapper should exactly achieve the opposite [14:49] (make ff not hang on flash sites) [14:50] ogra, generally it does. usually it makes any instances of flash just disappear. like magic! [14:51] it's like flashblock, only more reliable [14:52] Mithrandir: It was for the brief period I had either flash or java in the browser, but when I got rid fo them things started working just fine [14:52] maswan: I have neither, so I have no idea why it seems to hang every now and then. For something like 30s. [14:55] lifeless: http://benjamin.smedbergs.us/blog/2008-04-17/reviewing-merges-in-mercurial/ ... interesting POV imo .... what can bzr do to help reviewing the actual merge-changes done? [15:02] tkamppeter: Updated the patch, building CUPS now to test. [15:20] init-top in initrd, is that called before /sbin/init or just after? [15:21] Ok, I've had enough of evo eating my inbox now... [15:22] Who's responsible for evo so I can get this fixed, ASAP? [15:22] haha, [15:22] you're talking about evolution? [15:22] yep [15:22] i was faithful to that @#$%^ for 6yrs [15:22] I've been using it for 3, and it never screwed me over till now [15:22] i'm a much happier person after moving to thunderbird [15:23] +sunbird [15:23] I'm think it's a 64-bit issue, else more people would have noticed [15:23] ion_, there must also be a line only containing [15:23] application)/postscript [15:23] application/postscript [15:23] BenC, it works fine for me on x86 here [15:23] i only had minor issues this time [15:24] BenC, seb128 maintains it for us [15:24] in /etc/cups/mime.types, to define application/postscript [15:24] ogra: x86-32? I'm on -64 :) [15:24] seb128: ping [15:24] hi [15:24] are you up to date ? there were some fixes recently [15:24] what evolution-data-server version do you use? [15:25] seb128: 2.23.90.1-0ubuntu1 [15:25] and what is the issue? [15:25] using local boxes, imap, other? [15:25] seb128: I get weird things like Junk folder and Trash folder highlighted like that have new mail in them...when I empty it all, it shows -X in the status [15:25] seb128: local mboxes [15:26] seb128: in my Inbox, it keeps zeroing out my mail (not the stuff flagged as Junk or Trash) [15:26] seb128: confirmed with mutt [15:26] urg [15:26] 2.23.90 had issue that should be fixed in 2.23.90.1 [15:27] did you restart evolution since the update? [15:27] seb128: I removed everything from ~/.evolution/mail/local/ (backed it up) and started fresh, and it keeps happening [15:28] seb128: the .1 update was installed yesterday, and I've done "evolution --force-shutdown" about 10 times since then [15:29] (on the phone) [15:32] BenC, its a kernel issue ... you simply forgot to modprobe evolution :P [15:34] * BenC modprobe chuckle [15:40] tkamppeter: A moment still, the conversion still failed with one of my test PS files. [15:43] re [15:44] seb128: For instance, on a fresh restart, after deleting all indexes and cmeta, Trash shows 858 deleted, but when I ^A, it shows only 429 selected [15:44] BenC: the junk and trash count issues is a known bug and is being worked upstream, it's mostly cosmetic though [15:44] right, trash count is broken [15:45] seb128: Ok, then corrupting my Inbox is the main issue...sometimes it shows dupes of every email (and double count) and sometimes it zeroes out all of my messages (headers remain) [15:46] seb128: usually this happens in combination with emptying Junk and/or Trash, which is why I associated the count bug with it [15:46] tkamppeter: Huh. This fails: /usr/lib/cups/filter/psadobedrmtops seb128: I've tried pruning my Inbox by hand (with mutt), and it still will end up corrupted somehow [15:47] tkamppeter: This works: /usr/lib/cups/filter/psadobedrmtops foo.ps; /usr/bin/ps2pdf13 -dAutoRotatePages=/None -dAutoFilterColorImages=false -dColorImageFilter=/FlateEncode -dNOPLATFONTS -dWRITESYSTEMDICT foo.ps /dev/null [15:47] BenC: disk/partition full? (or did evo copy some code of kmail there *g*)? [15:47] disk has 80Gigs free [15:48] seb128: not sure if this is pertinent, but on console I keep seeing this: (evolution:25371): camel-WARNING **: Error during searching SELECT uid FROM 'Templates' WHERE ((deleted = 1) and ((deleted != 1) and (junk = 1))): no such table: Templates [15:48] seb128: same message for Outbox [15:48] tkamppeter: Should i just make the pstopdf filter collect the document to a tempfile before processing if it comes from stdin? [15:51] BenC: well, the corruption issue is exactly why evolution-data-server 2.23.90.1 has been rolled and other users confirmed it fixes the issue for them, did you upgrade all the libe* too? [15:52] seb128: I did a full update just before msg'ing you [15:52] ion_, are you intending to put your new DRM filter into pstopdf? This is also a possibility, as the change is only needed for converting the PS to PDF. [15:52] tkamppeter: Hmm, good idea. I’ll do that. [15:54] BenC: can you make sure there is evolution* process running, start a new evolution and tell me what you do to get a corruption [15:54] seb128: I made a mistake....1 update just got installed this morning...let me make sure I can reproduce the corruption still [15:54] BenC: when you do get a corruption are mails displayed as "?" [15:54] seb128: yes, mails were display as "?" [15:54] that's definitively what 2.23.90.1 should fix [15:55] seb128: right now I can't empty my trash :) [15:55] right, trash is known to be broken [15:55] they are having a hard time to get the summary disk issues fixed and focussed on the important ones first [15:56] trash and counts should be fixed for next week [16:36] tkamppeter: This seems to work. http://heh.fi/patches/cups/01-adobe-ps [16:36] ion_, with DRM handling in pstopdf no extra mime type needs to be defined. [16:38] tkamppeter: If that looks okay to you, i can make the Ubuntu patch. === Kopfi|offline is now known as Kopfgeldjaeger [16:38] kopfgeldjaeger: Thanks a lot for the information! [16:39] ion_, is OK, go ahead with the Ubuntu patch. [16:39] if you're talkin 'bout the "* Kopfi|offline is now know as Kopfgeldjaeger" - I use a bnc, so it's not more information than "* Kopfgeldjaeger has joined #ubuntu-devel" === superm1 is now known as superm1|away [16:44] seb128, i'm playing with evo in a mobile setup atm ... do you think it would be possible to get a (gconf switchable) button added for the "goto next unread mail" function ? its kind of tricky to use evo without keyboard [16:46]