**** BEGIN LOGGING AT Sun Sep 02 02:59:57 2007 Sep 02 03:53:07 pupnik, did you add smw to the garage? Sep 02 04:32:21 bye Sep 02 05:23:42 Does anyone know how to switch from the fingerboard to the keyboard in VNC? Sep 02 05:43:24 err Sep 02 05:43:31 depending on what you actually mean Sep 02 05:43:59 the rumor implies you'd need to stick a maemo gtk immodule in the host system Sep 02 07:30:28 straind`: oh, from finger to virtual? shouldn't it be sufficient to minimize the finger keyboard and single tap an input field? Sep 02 07:30:54 It doesn't seem to work in VNC. Sep 02 07:31:54 you're vncing to your device? Sep 02 07:32:42 from Sep 02 07:33:18 from i.e. you're running vnc client on your device? Sep 02 07:34:03 Yeah. VNC viewer connecting to VNC server on my desktop. Sep 02 07:34:32 so, getting those keyboards to work requires the application be able to explain widgets well enough to a gtk im Sep 02 07:34:55 I get the virtual for password entry, but once connected, it's finger. Sep 02 07:35:07 im? Sep 02 07:35:16 the simplest test of this is trying to run vnc on your desktop to a second workstation and trying to use japanese input methods [im] on your client desktop Sep 02 07:35:36 if you can't make that work, then expecting this to work is unreasonable Sep 02 07:36:34 I need sleep. That looks like english, but I don't understand it. :) Sep 02 07:40:26 some languages are so complicated you can't really have a keyboard w/ just ~100 keys to input the characters Sep 02 07:40:43 so they generally require software and some sort of window for dealing w/ input Sep 02 07:41:02 this begat Input Method Editors [IMEs] Sep 02 07:41:24 It's weird. I can get the finger up with the select button, and close it with the same. When I try to get the keyboard up by tapping the icon of the paper and pencil, nothing happens until I tap it again and get the fingerboard again. Sep 02 07:42:19 some devices are so frustratingly small, that they too needed some sort of window for dealing w/ input, this resulted in using the same system of IMEs on small devices Sep 02 07:43:04 but typically in order to support an IME you need to either use widgets predesigned for the IME (e.g. Gtk widgets will sorta automatically work w/ gtk-im-modules) Sep 02 07:43:12 or you need to write your own glue to connect to them Sep 02 07:43:26 opera, mozilla, and vnc would fall into that latter catgeory Sep 02 07:43:32 s/geo/ego/ Sep 02 07:43:33 timelyx meant: opera, mozilla, and vnc would fall into that latter category Sep 02 07:43:51 since they all draw their own "widgets" to the extent that they even understand what a widget is Sep 02 07:44:20 in the case of vnc, it would basically have to pretend to be an IM in order to ask the apps it's rendering enough questions to make a local IM happy Sep 02 07:46:06 Ok. So it's VNC on the 770s fault for not requesting the right thing, but it could be fixed by someone who understands it more than I do right now. :) Sep 02 07:46:22 that's my presumption, yes Sep 02 07:46:31 and the way to test it is to use japanese w/ desktop boxes Sep 02 07:46:44 if japanese "works correctly", then the problem could be maemo's fault instead Sep 02 07:46:50 could. IMEs are pains Sep 02 07:47:08 i think they make general unicode handling seem like a piece of cake Sep 02 07:47:14 Ok. I'm hoping I won't have to deal with that too soon. :) Sep 02 07:47:51 While I've got you, how hard is it to modify a virtual keyboard? Sep 02 07:48:36 please be more specific Sep 02 07:48:53 I use the US & UK keyboards. Sep 02 07:48:55 there are keyboard layouts, i think that someone reverse engineered enough to get hebrew working Sep 02 07:49:12 there's actually a public doc explaining how to write your own layout Sep 02 07:49:25 and there's a package w/ the program to "compile" layouts Sep 02 07:49:35 Ok. I'll look for it when my eyes aren't closed. :) Sep 02 08:11:27 woohoo Sep 02 08:11:31 * timelyx goes off to kill 3 products Sep 02 08:13:07 * timelyx frowns Sep 02 08:13:13 we have an Other for bugs about maemopad? Sep 02 08:18:29 timelyx: couldn't you maybe sneak in an inconspicuous product "3rd party" with skype and flash as components? Sep 02 08:18:44 lle2: skype is in communications Sep 02 08:18:55 because it was already in the bug db before i started Sep 02 08:19:03 so it moved to communications when i moved things Sep 02 08:19:15 before i was told not to have it Sep 02 08:19:16 really? I couldn't see it at the time Sep 02 08:19:22 it's there Sep 02 08:19:29 i put ~5 bugs there recently Sep 02 08:19:31 alright, no worries then Sep 02 08:19:35 flash is in browser because i own browser Sep 02 08:19:41 :) Sep 02 08:19:44 and i need a place to say "not my fault" Sep 02 08:19:54 yup Sep 02 08:20:07 the only thing is that browser has nokians, so we can touch bugs Sep 02 08:20:22 it's unclear how places that don't have nokians will deal with bugs given standard contract procedure Sep 02 08:20:36 but at least flash has promised to look periodically Sep 02 08:20:41 maybe jaaksi will handle them Sep 02 08:20:48 who? Sep 02 08:21:05 ok. Sep 02 08:21:08 dr. ari j. our benevolent leader ;) Sep 02 08:21:19 does he have a bugs.maemo.org account? Sep 02 08:21:26 no idea, but he should Sep 02 08:21:33 it'd be amusing to make him watch misdirected:nokia ;) Sep 02 08:21:40 :D Sep 02 08:22:22 is a bug about mameopad (sdk docs) something that should go in development platform, documentation, or website:documentation?. the bug includes a fix for a compile error Sep 02 08:22:40 * timelyx needs to see about killing website:documentation Sep 02 08:23:19 he has an account Sep 02 08:24:20 i'm not quite sure i feel like sending an email saying "dear benevolent leader, i've impersonated you to make you watch misdirected nokia related bugs" Sep 02 08:24:54 pre-texted, that's more mgmt language Sep 02 08:25:00 what i could do, i suppose, is write a blog entry with suggestions for what various people should watch and detailed steps Sep 02 08:25:08 pre-texted? Sep 02 08:25:20 you've not followed the hp saga? Sep 02 08:25:44 i followed it, but i didn't expect someone to mention it Sep 02 08:26:03 so what would that mean? Sep 02 08:26:09 pre-texting is supposedly when somebody (il)legitimately pretends to be someone else to gain something Sep 02 08:26:18 actually Sep 02 08:26:22 there's a simple way to do that Sep 02 08:26:51 i could send him an html message w/ an image which would subscribe him to the mail as long as he's already logged into bugzilla :) Sep 02 08:27:06 I guess it was first used to describe the process of acquiring somebody else's phone records from the phone companies Sep 02 08:27:15 * timelyx nods Sep 02 08:27:36 in this case, one of the nokia qas sent me an email @ my nokia address asking me if i had impersonated them and why Sep 02 08:27:48 i said, yes. but i was fairly shocked that they recognized it was me Sep 02 08:28:03 since afaik my real name and timeless aren't very connected in nokia circles Sep 02 08:28:22 heh Sep 02 08:28:36 there's a great quote in mozilla.org circles Sep 02 08:29:09 is a bug about mameopad Sep 02 08:29:11 eep Sep 02 08:29:13 http://quotes.burntelectrons.org/1388 Sep 02 08:30:52 ooh, i got a kudo, wow Sep 02 08:30:57 <`0660> so what's with this new nick? are you trying to be anonymous? :) Sep 02 08:30:57 http://www.ohloh.net/accounts/6653/positions/3393 Sep 02 08:31:07 _Monkey me? Sep 02 08:31:07 <_Monkey> hmmm... timelyx is your mac, it has scrollback Sep 02 08:31:18 _Monkey timele61i? Sep 02 08:31:18 <_Monkey> somebody said timele61i was well, your phone or timeless using a nokia e61i (blueberry) Sep 02 08:31:40 no, i just have a number of accounts w/ various properties Sep 02 08:31:52 timeless is my identity, the others are variations based on connectivity Sep 02 08:32:43 <`0660> i didn't know you could do irc with Lyx :P Sep 02 08:33:12 <`0660> i wasn't even able to do my thesis with it Sep 02 08:33:16 <`0660> i had to use latex Sep 02 08:36:14 <`0660> so what's your nick when using n800? :) Sep 02 08:39:22 typically timeless here Sep 02 08:39:40 from my n800 i tend to use ssh/screen/epic Sep 02 08:39:56 on moznet it's timeless/timely depending on the channel Sep 02 08:40:04 (two screen sessions) Sep 02 08:40:08 <`0660> pretty confusing :) Sep 02 08:40:30 i don't use nick coloring, always address timeless, i'll respond if i'm alert Sep 02 08:40:39 or when i read scrollback Sep 02 08:40:56 i'll respond to the others too, since i'm reading everything Sep 02 08:41:21 do you also have a life? :) Sep 02 08:41:51 i have a summer vacation starting next week, does that count? Sep 02 08:42:37 depends on what you do there Sep 02 08:42:46 if you'll spend it hanging on irc reading backlogs, no Sep 02 08:42:50 travel/pray Sep 02 08:42:52 <`0660> i think 5 irc nicks should count as one life :) Sep 02 08:43:19 i expect to spend less than half my days w/ access to anything electronic Sep 02 08:43:27 that includes things like lights Sep 02 08:43:35 where to you travel to? Sep 02 08:43:36 we'll see how close i get to that Sep 02 08:44:02 vienna, bologna, milano, munich, frankfurt, bern, strausborg, paris Sep 02 08:44:12 only the first and last are guaranteed Sep 02 08:44:35 i'm currently considering swapping bologna, milano for milano, tonio, or milano, bologna Sep 02 08:44:37 going along some pilgrims way? Sep 02 08:44:51 i might drop bern or trade it for zurich Sep 02 08:45:09 not specifically, it's sorta a "hit classic central europe" Sep 02 08:45:55 i live halfway between zurich and strasbourg Sep 02 08:46:01 i have plans to hit the Mediterranean and British Isles in subsequent years (along w/ a couple of others) Sep 02 08:46:26 oh yeah, Nordic countries is one of the other trips :) Sep 02 08:46:41 my next trip is to tokyo Sep 02 08:46:48 nice Sep 02 08:46:54 checking whether the sites from mega tokyo actually exist Sep 02 08:46:59 i have an invite but no allotment Sep 02 08:47:05 heh Sep 02 08:48:02 * timelyx tries to figure out what to do w/ a bug asking for bluetooth printing Sep 02 08:48:16 connectivity:bluetooth, desktop:user interface, misdirected:nokia Sep 02 08:51:51 Hmm.. where have the sardine repositories gone? Sep 02 08:52:02 ? Sep 02 08:55:02 http://repository.maemo.org/sardine/ says "Sardine has moved to the Maemo Garage. If your browser doesn't automatically redirect there click here" Sep 02 08:55:35 And the getting started document from that page redirects to wiki document, that gives the http://repository.maemo.org/sardine/ -repository Sep 02 08:56:39 http://sardine.garage.maemo.org/ Sep 02 08:57:02 * timelyx sighs Sep 02 08:57:06 finns are strange Sep 02 08:57:09 That's the page that the old repository redirects to Sep 02 08:57:17 only a finn would think to call sardine's fresh or tasty Sep 02 08:57:32 hahahaha Sep 02 08:58:21 now, that's not to say i have anything personally against sardines, i just won't eat them :) Sep 02 08:58:34 they're ok imo Sep 02 08:59:01 Unfortunately we are poisoning the oceans - mercury levels etc Sep 02 08:59:05 but offtopic.. Sep 02 08:59:47 http://timeless.justdave.net/mxr-test/garage/source/sardine/ Sep 02 08:59:51 sardine is walkable Sep 02 09:01:25 * timelyx loves seeing ip addresses in files: http://timeless.justdave.net/mxr-test/garage/source/sardine/trunk/sardine-experimental.sources.list Sep 02 09:01:54 timelyx: They are internals Sep 02 09:03:45 theril: that's fairly stupid for a public project, no? Sep 02 09:04:44 timelyx: Yep Sep 02 09:05:17 but what makes you say they're internal? Sep 02 09:05:37 Because 192.x are intranet IPs Sep 02 09:06:19 timelyx: why do you think sardine was coined by a finn? Sep 02 09:06:23 no Sep 02 09:06:24 http://www.geocities.com/technofundo/tech/misc/abcofip.html Sep 02 09:06:34 There are 3 IP network addresses reserved for private networks. Sep 02 09:06:35 * 10.0.0.0 to 10.255.255.255 Sep 02 09:06:35 * 172.16.0.0 to 172.31.255.255 Sep 02 09:06:35 * 192.168.0.0 to 192.168.255.255 Sep 02 09:06:46 timelyx: isn't that one of carlos' projects? Sep 02 09:07:21 lle2: maybe, perhaps i'm wrong. it just seems like a finnish choice. perhaps i should apologize :| Sep 02 09:07:35 Then I'm glad to be wrong :) Sep 02 09:08:10 But the IP doesn't answer anyway Sep 02 09:09:09 Hmm.. lets try chinook then Sep 02 09:09:28 Does anybody have chinook's repository's URL at hand? Sep 02 09:09:51 timelyx: 192. addresses can still be internal to nokia Sep 02 09:10:17 lle2: to the extent that they're a specifically allocated netblock, sure Sep 02 09:10:29 but not for the standard "oh, it's 192.168.*" Sep 02 09:10:45 yes Sep 02 09:10:46 http://timeless.justdave.net/mxr-test/garage/source/sardine/trunk/sardine-experimental_closed.sources.list Sep 02 09:10:49 now that's a funny list Sep 02 09:11:51 Argh, developing outside Nokia is pain Sep 02 09:12:09 i hope you don't think developing inside nokia is cake Sep 02 09:12:35 Well I didn't have to hunt down stuff like this Sep 02 09:12:43 * timelyx chuckles Sep 02 09:12:46 sure you do Sep 02 09:12:50 or at least, i do Sep 02 09:13:00 instead of 3 inconsistent wikis Sep 02 09:13:04 you have 50 you don't know about Sep 02 09:13:31 and hundreds of people you don't know well enough to know if they know what you need to know Sep 02 09:13:40 I know, but I'm now talking about getting the sources Sep 02 09:13:51 sure Sep 02 09:13:53 same problem Sep 02 09:14:05 getting sources involves figuring out who has sources and how to retrieve them Sep 02 09:14:19 to /some/ extent, my computer at work solves that problem Sep 02 09:14:26 but only when i found out, which can be months late Sep 02 09:14:36 Wasn't atleast for me, I just put the integration repository to sources.list and pulled all the fresh stuff Sep 02 09:15:05 and since it isn't cron'd yet, my server will be basically 4 weeks stale for a heavy period Sep 02 09:15:31 * timelyx shrugs Sep 02 09:15:44 working off integration repositories is /sometimes/ useful, and usually frustrating Sep 02 09:16:03 they keep changing their structure Sep 02 09:16:51 http://timeless.justdave.net/mxr-test/garage/source/sardine/trunk/Y.txt Sep 02 09:16:57 can someone please explain that file? Sep 02 09:16:59 I had no problems Sep 02 09:17:24 A workaround for lack of /bin/yes? Sep 02 09:18:36 * timelyx considers enrolling in sardine to fix a typo in http://timeless.justdave.net/mxr-test/garage/source/sardine/trunk/README.rootstraps#4 Sep 02 09:18:55 lle2: think anyone would mind? Sep 02 09:20:42 timelyx: mind what? Sep 02 09:21:06 if i fixed the spelling of "built" by forcibly adding myself to svn access and committing the fix Sep 02 09:21:18 lle2: BTW, is it possible in Sb2 to have the .sbconfig file somewhere else than in the "root" directory? Sep 02 09:21:25 that's counter-revolutionary thinking Sep 02 09:21:45 I'm thinking of using a mount of N800 root as Sb2 root fs Sep 02 09:21:50 theril: sure it would be possible, what would be a good place? Sep 02 09:22:13 what's a "root" directory in this case? Sep 02 09:22:20 the root of the scratchbox? Sep 02 09:22:25 the root of the buildroot Sep 02 09:22:36 as in $HOME/buildroot/.sb2config Sep 02 09:22:51 lle2: Well perhaps home directory or someplace writable by 'user' (instead of root) Sep 02 09:23:07 does sb2config let me have more than one scratchbox? Sep 02 09:23:15 timelyx: yes Sep 02 09:23:24 I really don't want to have sb2 running with root access on my device ;) Sep 02 09:23:31 timelyx: as many as you like Sep 02 09:23:38 * timelyx is confused Sep 02 09:23:41 theril: what am i missing? Sep 02 09:23:44 where does root come in? Sep 02 09:24:14 timelyx: It would use a remotedly mounted FS of the N800 Sep 02 09:24:52 And if I want to write the config to root of the N800, it would have to be mounted as root Sep 02 09:24:53 oh Sep 02 09:25:05 so you're using sb2 for sbrsh transparency Sep 02 09:25:07 And then sb2 magic would probably fuck up the system totally :P Sep 02 09:25:10 or trying to Sep 02 09:25:14 timelyx: Not exactly Sep 02 09:25:20 * pupnik ponders Sep 02 09:25:26 theril: that's interesting way of setting it up Sep 02 09:25:28 * timelyx doesn't exactly get it Sep 02 09:25:38 might be useful Sep 02 09:25:40 In fact I'm thinking of using raemo to mount and run stuff with sb2 Sep 02 09:25:51 Raemo? Sep 02 09:26:05 <`0660> Raimo Sep 02 09:26:05 timelyx: http://raemo.garage.maemo.org/ Sep 02 09:26:16 theril: but raemo doesn't do that, as far as I know Sep 02 09:26:34 lle2: Do what? Sep 02 09:26:45 BTW, I'm the one that wrote raemo ;) Sep 02 09:26:49 theril: mount entire root filesystems anywhere Sep 02 09:26:55 theril: right :D Sep 02 09:27:00 lle2: Sure it can Sep 02 09:27:22 theril: are you planning to sshfs the entire root from the device to the pc? that would work Sep 02 09:27:27 theril: raemo didn't exist as a public svn creature when i pushed garage to timeless.justdave.net and practically speaking, i don't look at garage directly if i can avoid it :) Sep 02 09:28:06 lle2: Exactly Sep 02 09:28:17 theril: .sb2config is only used by the sb2 script to read in and set some environment variables Sep 02 09:28:30 And I was about to run sb2 on it, but I didn't really want to mess around as root Sep 02 09:28:54 lle2: So I can use ~/.sb2config to override it? Sep 02 09:29:16 theril: basically yes, you just need to modify sb2 script a bit Sep 02 09:29:21 OK Sep 02 09:29:40 theril: it currently looks for .sb2rc file to find the buildroot from which to dig up .sb2config Sep 02 09:30:12 And it goes upwards the directory structure until it finds one? Sep 02 09:30:26 theril: you might change it so that if the buildroot doesn't have a .sb2config it just uses whatever was in .sb2rc Sep 02 09:30:30 theril: yes Sep 02 09:30:34 And the cache stuff could be configured in the same way Sep 02 09:30:49 lle2: did i ever resend that patch for sb2 Sep 02 09:30:57 timelyx: no Sep 02 09:31:07 * timelyx grumbles Sep 02 09:31:15 Hmm.. perhaps even more elegant would be to put the .sb2config one directory below the root and in that .sb2config there would be path to the root Sep 02 09:31:35 So the magic still works but it doesn't clutter the rootfs Sep 02 09:31:39 theril: that wouldn't gain you anything Sep 02 09:31:43 config files should not be visible in the environment system Sep 02 09:31:46 theril: I want to have multiple roots Sep 02 09:32:01 ideally an app running in a scratchbox should not be able to figure out it's in scratchbox :) Sep 02 09:32:45 theril: just do it so that the .sb2config parameters can be defined in .sb2rc as well Sep 02 09:33:04 lle2: Yes, but you'd have like sb2/root1 sb2/root2 and sb2/.sb2config.root1 (or something like that) and when sb2 is run from inside sb2/root1 it would find the .sb2config.root1 from sb2/ Sep 02 09:33:24 theril: I don't run sb2 from inside the buildroot Sep 02 09:33:24 And also it would have sb2/.sb2config.root2 etc Sep 02 09:33:26 lle2: got sb2 sources handy? Sep 02 09:33:33 lle2: I do ;) Sep 02 09:33:45 timelyx: http://freedesktop.org/Software/sbox2 Sep 02 09:33:52 lle2: i mean to make changes Sep 02 09:34:08 i don't know where my secureid is, so i figure i'll tell you the patch here Sep 02 09:34:13 timelyx: uh, no, not now, I'm selling my PCs :) Sep 02 09:34:14 and you fix it while i wait Sep 02 09:34:21 !?@ Sep 02 09:34:46 timelyx: getting rid of everything, even the mac mini Sep 02 09:35:00 bah, i'll write the general patch here, and you can pull it from the irc logs later :) Sep 02 09:35:02 <`0660> are you turning into a monk? Sep 02 09:35:08 preload/sb_exec.c Sep 02 09:35:18 do_exec(...) Sep 02 09:35:18 `0660: something along those lines Sep 02 09:35:26 timelyx: I think I fixed that already Sep 02 09:35:43 lle2: you fixed the strdup? Sep 02 09:35:46 <`0660> cool :) Sep 02 09:35:49 errr basename() Sep 02 09:35:53 Morning, all Sep 02 09:35:53 timelyx: yeah, maybe Sep 02 09:36:11 <`0660> maybe you have a burnout? Sep 02 09:36:45 `0660: no, maybe I had too many f'ing pcs that made my apartment look like a hosting facility Sep 02 09:36:57 <`0660> :) Sep 02 09:37:05 <`0660> how many are you going to keep? :) Sep 02 09:37:22 `0660: just two tiny laptops Sep 02 09:38:21 lle2: Are you going to sell the stuff? Sep 02 09:38:54 theril: sold already, unless you're interested in a slightly whining 17" lcd Sep 02 09:39:30 theril: it's not too bad, makes this annoying little sound when there's no signal Sep 02 09:40:42 lle2: Hmm.. I could be Sep 02 09:40:52 I'm planning to get rid of my TV Sep 02 09:41:16 theril: if you come and pick it up, you can get it for the low price of one good finnish beer Sep 02 09:41:45 theril: one dead pixel Sep 02 09:41:58 OK. Which beer is the good one? Sep 02 09:42:11 theril: koff! Sep 02 09:42:15 OK, if it has more than that, I'll take the beer back Sep 02 09:43:04 theril: sure, I'll return it if it comes to that Sep 02 09:45:50 Anyone know of any full-screen, pretty, child-friendly URL/app launchers for XP? Something a bit like Apple's Front Row and/or their dock. Sep 02 09:47:26 just buy an iphone w/o a sim card :) Sep 02 09:48:20 Hmm.. has anybody got composite working on N800? Sep 02 09:49:10 I read some slides that said it's working on sardine and thought of backporting its xserver and now when I got the sources, they are the same as in bora Sep 02 09:49:20 And config has composite disabled Sep 02 09:49:28 And building with the flag fails Sep 02 09:49:36 useful Sep 02 09:49:47 timelyx: this is for an old ThinkPad being used by Jaffa Jr ;-) Sep 02 09:51:04 http://www.slideshare.net/jobi/xcomposite-on-internet-tables Sep 02 09:51:22 http://it.slashdot.org/comments.pl?sid=286201&cid=20437865 Sep 02 09:51:36 "Already in sardine, will come with chinook". Yeah right Sep 02 09:53:59 * timelyx shrugs Sep 02 09:55:12 And the damned composite compiling fails on wrong arguments given to libtool (two libraries on row without spaces, like ../foo.a../bar.la) Sep 02 09:55:35 And of course it has some autogenerated autogeneration stuff so no human being can debug them Sep 02 09:59:38 that's a feature Sep 02 10:01:08 www.maemo.com ?? Sep 02 10:03:25 Why the hell people even use autotools? Sep 02 10:03:55 because build environments differ? Sep 02 10:04:02 to make life difficult :). especially to ensure that if people upgrade autotools, builds will fail Sep 02 10:04:41 i've been confused about that stuff since the early 90s Sep 02 10:05:13 run-around in circles waving hands in the air and hollering and then running into the wall confused Sep 02 10:05:25 Its a lot easier to hand write the magic in Makefile than fight with autohell Sep 02 10:06:01 well it's been working fairly well for me Sep 02 10:06:09 lately Sep 02 10:06:52 Hack on cludge on hack Sep 02 10:13:40 ok. all products in bugs.maemo.org are things i want to exist Sep 02 10:13:49 that only leaves a couple of components that should die Sep 02 10:14:47 helpwanted, there are 11 "applications:core", 8 "applications:hardware" and 12 "applications:User interface and graphics and style" Sep 02 10:16:40 Has there been any developments with the memory card corruption bug? Sep 02 10:16:56 I don't dare to put my card in the device until its fixed :( Sep 02 10:17:05 no Sep 02 10:18:00 I heard that the cause is now known though Sep 02 10:18:06 really? Sep 02 10:18:38 yes Sep 02 10:18:55 If you have access to internal bugzilla, there's probably more info Sep 02 10:19:09 i'd have to find my password Sep 02 10:19:18 which is more trouble than it's worth Sep 02 10:19:28 But AFAIK the corruption happens when the device goes to power saving mode Sep 02 10:19:30 lle2: why isn't useful information in the public bugzilla? Sep 02 10:19:38 timelyx: no idea Sep 02 10:20:11 If I was in charge, all bugs would be in the public bugzilla Sep 02 10:20:45 theril: this bug is in the public bugzilla Sep 02 10:20:48 wrong problem Sep 02 10:20:58 the trick is getting all people to actually *use* the public bugzilla Sep 02 10:21:13 and not, e.g. email, cubical hopping, or another bug tracker Sep 02 10:21:15 Hmm.. my login to internal bugzilla is still working :) Sep 02 10:21:24 i can fix that for you :) Sep 02 10:22:17 Why did I say anything :P Sep 02 10:22:32 i'd have to figure out who you are, and i'm way too lazy to do that Sep 02 10:22:55 Maybe I could talk about extending my NDA to keep the login Sep 02 10:22:57 just disable access for everybody Sep 02 10:23:34 theril: you couldn't talk to me about it :) Sep 02 10:23:44 but it sounds like a good idea Sep 02 10:24:20 do any of you know anything about gtk/hildon? Sep 02 10:24:25 do windows have "names"? Sep 02 10:24:40 I thought bugzilla accounts couldn't be deleted if they have bugs on them Sep 02 10:24:43 I heard that there are some community developers with access there Sep 02 10:24:47 so just file a few Sep 02 10:24:52 lle2: disabled != delete Sep 02 10:25:00 hmph, details, details Sep 02 10:25:01 disabled = bugzilla refuses to let you login Sep 02 10:25:03 timelyx: What kind of names? They do have window IDs Sep 02 10:25:32 theril: is that a consistent human readable string that's likely to be unique per app Sep 02 10:25:44 but consistent per app across app instances? Sep 02 10:25:51 "bob" Sep 02 10:26:09 sounds good Sep 02 10:26:19 what happens if i want two "bob" windows? Sep 02 10:26:36 "bob2" Sep 02 10:26:41 that's what happened to me Sep 02 10:26:53 bob2 won't work very well Sep 02 10:28:29 timelyx: Hmm.. I don't think so. But you can set XProperty to the window Sep 02 10:29:22 There is a function to set application's name in glib, but it may be just related to debugging output etc Sep 02 10:30:02 application names are useless Sep 02 10:30:13 applications i care about have multiple window classes Sep 02 10:30:21 and yes, i know about the xproperties Sep 02 10:30:37 gecko tries (and probably fails miserably, it typically does...) to use them Sep 02 10:30:42 Why would you need the name? Sep 02 10:30:51 persisting window state Sep 02 10:31:05 because some annoying free thinker thinks that maemo sucks Sep 02 10:31:15 (maemo does suck, but solving the problems he cares about is non-trivial) Sep 02 10:31:31 maemo can't suck, it's not a physical entity Sep 02 10:31:49 maemo as deployed in each product i can buy and or flash sucks Sep 02 10:32:00 is that acceptable? Sep 02 10:32:24 maemo as manifested in the form of ITOS sucks Sep 02 10:32:29 although even that's backwards Sep 02 10:32:29 timelyx: So that every time window opens, it opens to same place or something like that? Sep 02 10:32:46 theril: typically it has the same toolbar state Sep 02 10:32:58 or the same column order/sort, or widths Sep 02 10:33:32 I think that's more of the programs concern Sep 02 10:33:43 program's Sep 02 10:33:44 well, err Sep 02 10:33:53 he claims and in a way he's right that it's a platform bug Sep 02 10:34:14 the fact is, if programmers have to do things, they won't, or at least they won't consistently Sep 02 10:34:24 I think that GTK+ API is too low level to even enable that Sep 02 10:35:01 surely there must be a way to give a window a name? Sep 02 10:35:28 I'm not aware of system that would be needed for this Sep 02 10:35:44 lle2: i asked first =b Sep 02 10:35:53 timelyx: :D Sep 02 10:36:14 That it would detect that a window is instance of some widget configuration Sep 02 10:36:15 but couldn't gtk just use X to do the dirty work? Sep 02 10:36:34 whaat? Sep 02 10:37:28 What is the common denominator with two instances of "same" window? Sep 02 10:37:38 the name? Sep 02 10:37:58 or "class" or whatever they choose to call it Sep 02 10:38:20 Yes, but GTK doesn't force (or even allow) the program to specify that Sep 02 10:38:53 uhhuh Sep 02 10:38:55 So every program would have to be hacked to explicitly set the name using XProperty or something like that Sep 02 10:39:16 I'm so glad I don't have to use gtk to do my work Sep 02 10:39:26 i'd love to get rid of gtk Sep 02 10:39:32 Is not that bad IMHO :) Sep 02 10:39:34 it's such a stupid system upon which to build a platform Sep 02 10:39:40 theril: no, it's much worse Sep 02 10:40:02 maybe we shouldn't get rid of it, just build up something more sane on the side Sep 02 10:40:11 lle2: gecko Sep 02 10:40:11 that integrates well with it Sep 02 10:40:11 The abstraction level is a bit wrong though, but it makes a fine low level widget set Sep 02 10:40:18 xul based applications Sep 02 10:40:30 gecko is a pile of hack Sep 02 10:40:31 they scale, they flex, they persist Sep 02 10:40:59 three things the maemo platform doesn't do Sep 02 10:41:22 maemo scales, in man months Sep 02 10:41:23 oh, and generally, gecko doesn't kill itself when it runs out of memory Sep 02 10:41:38 They could be easily incorporated in the API Sep 02 10:41:41 lle2: that's something like 200 man months Sep 02 10:41:54 Just force programmer to set window name when creating a HildonWindow Sep 02 10:42:35 And properly coded GTK apps do scale Sep 02 10:42:36 yeah, more maemo specific crap will solve the problem Sep 02 10:42:45 ++lle2 Sep 02 10:42:59 using Gazpacho (?) might solve the problem Sep 02 10:43:19 since it's basically getting rid of C initializers and using an xml dialect (xul done w/ gtk instead) Sep 02 10:43:51 I'm not too big a fan of that kind of stuff Sep 02 10:44:10 You'll soon end up with having a gecko (which IMO isn't very good thing) Sep 02 10:44:26 But a API like that over GTK would be good Sep 02 10:44:29 theril: gazpacho is basically yet another gecko Sep 02 10:44:34 which is silly Sep 02 10:44:42 * timelyx wishes maemo would give up and just move to gecko Sep 02 10:44:54 we'd have a notes app that wouldn't suck Sep 02 10:45:01 we'd have a chance at a mail app that didn't suck Sep 02 10:45:03 AFAIK gazpacho is yet another glade Sep 02 10:45:15 we could make the ui designers actually write the UI instead of specifying it Sep 02 10:45:34 yeah, i originally wrote glade but replaced it w/ gazpacho Sep 02 10:45:45 And we'd have apps that take minutes to load Sep 02 10:45:48 i can't remember or understand the difference Sep 02 10:46:02 theril: we have skype :) Sep 02 10:46:06 And 30Mb of javascript and XML loaded in the memory at all times Sep 02 10:46:25 but actually, it becomes less of a problem if it's already loaded Sep 02 10:47:10 Yeah, but the 30MB could be used to something more sane Sep 02 10:47:27 like the current email client? :) Sep 02 10:47:28 theoretically sure Sep 02 10:47:34 but do you really trust nokia to manage that? Sep 02 10:47:49 actually, at this point, i think you're likely to get the current email client :) Sep 02 10:48:35 Well maybe they'd make good use of a few of those 30 MBs :) Sep 02 10:48:43 pigs might fly Sep 02 10:48:58 would you like a meta crawler with that? Sep 02 10:49:02 timelyx: Have you done any gecko programming? Sep 02 10:49:09 theril: only around 7 years Sep 02 10:49:20 that's what i do for a living Sep 02 10:49:25 that's why i'm working for nokia Sep 02 10:49:27 And you like the mess that's its architecture? Sep 02 10:49:40 in comparison to nokia/osso/maemo? absolutely Sep 02 10:50:08 i'm glad you asked ;-) Sep 02 10:50:20 I think the comparsion is more against GTK+ etc Sep 02 10:50:40 glib/gdk/gtk/gnome/hildon/dbus Sep 02 10:50:49 how many cards did i miss in my house of cards? Sep 02 10:50:54 * timelyx blows and watches it collapse Sep 02 10:51:19 but even ignoring all those layers Sep 02 10:51:22 IMHO putting that monster on a portable device is plain insane Sep 02 10:51:48 the uis designed by/for mozilla.org are better than the junk shipped by nokia Sep 02 10:51:50 So you compare the whole gecko to the separate libraries Sep 02 10:52:05 i'll compare it however Sep 02 10:52:14 i can compare nss(works) with libcst (POS) Sep 02 10:52:36 i can compare PSM(works) w/ certman (has to hack around libcst) and still won't always work Sep 02 10:52:41 How about something like glib vs nsoldcludge, dbus vs the interface spec hacks etc Sep 02 10:53:16 And I don't think gecko even has functionality that could be compared with dbus Sep 02 10:53:20 theril: nspr v. glib? i don't mind, nspr works, is stable, and can be fairly safely used whereever Sep 02 10:53:34 Same for glib Sep 02 10:53:55 glib and nspr are fairly comprable Sep 02 10:54:06 yes Sep 02 10:54:18 dbus is basically a library guaranteed to enable your app to crash when the system runs out of memory Sep 02 10:54:25 i'm happy not to be forced to have such a library Sep 02 10:54:26 And doesn't gecko use GDK? Sep 02 10:54:34 it doesn't have to use gdk Sep 02 10:54:42 at some point i'm hoping for cairo-xlib Sep 02 10:54:51 But it uses Sep 02 10:54:57 and on windows and osx and qnx and beos it certainly does not use gdk Sep 02 10:55:13 it only uses it on a handful of backwards platforms :) Sep 02 10:55:18 I think that apps should be free to crash when run out of memory Sep 02 10:55:24 sure Sep 02 10:55:30 but they shouldn't be killed by libraries Sep 02 10:55:32 there's a difference Sep 02 10:55:50 Working around memory ending situations is IMO too complex for application level anyway Sep 02 10:55:50 if my app wants to crash or not crash when it runs out of memory, that should be its cohice Sep 02 10:56:06 it shouldn't be in a position where a fundamental system library will crash/kill the app Sep 02 10:56:18 you think the system does better? Sep 02 10:56:21 you're nuts. Sep 02 10:56:39 I don't usually even bother to check mallocs Sep 02 10:56:54 the os is the thing that gets apps into the problem in the first place, and the only thing it can do to deal with it is go around shooting things Sep 02 10:57:16 the os is *not* in a good, or even particularly useful place to make anything close to sane decisions Sep 02 10:57:43 in 2006, the os decided that the only thing it could do when it ran out of memory was shoot alarmd Sep 02 10:57:53 I have yet to see program that manages handling out of memory really better than just shutting down Sep 02 10:57:59 after it shot alarmd, it remembered it was supposed to run alarmd whenever alarmd died Sep 02 10:58:03 so, it went off and started it again Sep 02 10:58:07 and then it shot it again Sep 02 10:58:10 this repeats Sep 02 10:58:48 they eventually did fix that one ... Sep 02 10:59:06 but integrating sanity into the os is clearly rocket science Sep 02 10:59:06 I think the memory use of apps should be compartmentalized, and localise the killing to the box that is actually overflowing instead of going out and killing some random process Sep 02 10:59:29 The D-Bus implementation admittedly isn't very great Sep 02 10:59:38 lle2: what do you do when the user is annoying and tries to build make -j 100 on a box? Sep 02 10:59:47 no one specific process uses a lot of memory Sep 02 10:59:58 now sure, you could claim a job pool is Sep 02 11:00:04 i'd claim it's the X server :) Sep 02 11:00:15 I vote for gecko Sep 02 11:00:15 or init :) Sep 02 11:00:21 :P Sep 02 11:00:31 timelyx: those processes would sit in the "misc" box which gets only a limited chunk to be split among them Sep 02 11:00:39 lle2: but the main problem w/ shooting the biggest consumer Sep 02 11:00:48 is that on average you shoot the app the user was actually *using* Sep 02 11:00:49 timelyx: and X resource allocations should be taken into account Sep 02 11:00:55 namely, the web browser or media player Sep 02 11:01:14 timelyx: I'm not shooting the biggest user, as long as it stays within its boundaries, it will live Sep 02 11:01:42 lle2: best way to set the boundaries is by actually returning NULL for alloc failures Sep 02 11:01:45 and letting the app deal Sep 02 11:01:52 everything else is just bad hacks Sep 02 11:01:55 that don't work Sep 02 11:02:02 timelyx: yes, that would be returned when the boundary is exceeded Sep 02 11:02:07 Programs should tell the OS how much memory they need and after the process tries to allocate more, it's shut down Sep 02 11:02:18 theril: how much does a web browser need? Sep 02 11:02:25 answer: as much as the user demands the browser need Sep 02 11:02:28 timelyx: A lot less than it uses now Sep 02 11:02:31 theril: it needs to be more dynamic than that Sep 02 11:02:48 lle2: Of course, but something in those lines Sep 02 11:02:52 theril: user decides to open 100 very heavy web pages Sep 02 11:03:02 user expects them all to be alive and fully functional Sep 02 11:03:07 timelyx: the user is then told that's not possible Sep 02 11:03:07 (user is crazy, but whatever) Sep 02 11:03:14 lle2: i'd love to Sep 02 11:03:23 lle2: theril wants the app to die instead Sep 02 11:03:34 timelyx: yup, murderous bastard Sep 02 11:03:43 yup Sep 02 11:03:51 or at least bloodthirsy Sep 02 11:03:55 s/y/yt Sep 02 11:03:59 arg Sep 02 11:04:01 or at least bloodthirsty Sep 02 11:04:12 I'd rather let my application die than to try to write a desperate attempt in rescuing it Sep 02 11:04:20 theril: fwiw, i'd rather not let users exceed 4 windows Sep 02 11:04:30 but i atm have no budget for making such a ue Sep 02 11:04:34 s/e$/i/ Sep 02 11:04:46 set a tax on windows :P Sep 02 11:04:49 theril: the browser must be able to deal with that, it's just bad form to crash if you go to a massive slashdot comment page or something Sep 02 11:05:10 Let them have all the windows they want, but write the DOM to HD and let it rerender it when needed Sep 02 11:05:37 lle2: Yes, but I'm talking from developers' point of view Sep 02 11:05:38 theril: there's no space for that on the flash, besides that would suck anyway Sep 02 11:06:06 So the better solution is just to spam the memory until it ends? Sep 02 11:06:06 theril: for kicks, consider: http://timeless.justdave.net/stress/1024-768.html Sep 02 11:06:15 theril: it's a start Sep 02 11:06:27 ideally gecko will recognize that it can drop an image that's not currently visible Sep 02 11:06:34 and replace it w/ one that it needs to show *now* Sep 02 11:06:40 theril: anyway, try loading that page Sep 02 11:06:49 butt be warned, i killed a Samsung Q1 w/ that last year Sep 02 11:06:51 timelyx: I loaded it in lynx :P Sep 02 11:07:01 and i've killed various unix boxes in the past Sep 02 11:07:15 the Q1 bluescreened, unix boxes panic Sep 02 11:07:25 or the xserver gets killed, which is equally amusing Sep 02 11:07:55 theril: anyway, how do you "render" that page? Sep 02 11:08:11 specifically, user loads the page and jumps to the bottom Sep 02 11:08:21 how do you know where the "bottom" is? Sep 02 11:08:31 hrm, wrong url Sep 02 11:08:41 True, I got the point Sep 02 11:08:47 sorry, http://timeless.justdave.net/stress/stress.html Sep 02 11:08:50 that's the right url Sep 02 11:09:03 But then we'll get to the subject of how much HTML sucks Sep 02 11:09:08 * timelyx shrugs Sep 02 11:09:14 i have to live with it Sep 02 11:09:16 browser is special of the apps that we have on the device Sep 02 11:09:19 i'd rather you not kill me over it Sep 02 11:09:29 lle2: mail and media player are not too far from it Sep 02 11:09:33 maybe email is a second one Sep 02 11:09:34 rss reader too Sep 02 11:09:39 kill that Sep 02 11:09:41 depending on how stupid you make them Sep 02 11:09:42 Well, I'm sort of pro-euthanasia Sep 02 11:09:44 lle2: gladly :) Sep 02 11:10:08 anyway, yeah, the browser is special Sep 02 11:10:15 most of the other apps have little excuse Sep 02 11:10:39 consider images which seems to get sick of the themes graphics directories :) Sep 02 11:10:52 it sorta goes off and hangs itself Sep 02 11:11:17 note: it is possible for browser to actually deal w/ that page properly Sep 02 11:11:33 and in fact, it doesn't do /such/ a bad job today in gecko, although it can do a better one Sep 02 11:11:49 actually, all things considered, that page is tame, albeit slightly overweight Sep 02 11:11:50 timelyx: Couldn't that be handled by just downloading size of the images and rendering empty boxes in place of them? Sep 02 11:11:56 theril: no Sep 02 11:12:03 you need to know if the image is valid Sep 02 11:12:08 which requires decoding it start to finish Sep 02 11:12:14 because if it's corrupt you're required to replace it Sep 02 11:12:26 By some spec? Sep 02 11:12:31 oh, and what happens if the image is dynamically changing between corrupt and not corrupt Sep 02 11:12:34 theril: html spec, yes Sep 02 11:12:54 Well, I'd say fuck you to that Sep 02 11:12:56 now, you could go off and download each one in turn, and determine its size and then discard it Sep 02 11:13:03 and then pray that the image doesn't change Sep 02 11:13:10 which i'm fine w/ doing Sep 02 11:13:29 but if you're on GPRS and paying by the byte, you /might/ not appreciate that :) Sep 02 11:13:48 actually, at 100+mb for the entire directory, you will never appreciate that page if you pay by the byte :) Sep 02 11:14:20 If I were to write a browser (that nobody would use :P), it would throw a parse error on corrupt data just like any other sane parser does Sep 02 11:14:43 no one would use it, it'd render maybe 5 pages Sep 02 11:14:50 It would also ease lots of the pain in creating the pages Sep 02 11:14:56 and i'm not sure which sort of corrupt data you're talking about Sep 02 11:14:57 Yes, but that's how it should work Sep 02 11:15:03 timelyx: Any Sep 02 11:15:04 corrupt images throw dom events Sep 02 11:15:11 which web pages can react to :) Sep 02 11:15:17 by e.g. changing the image location ... Sep 02 11:15:39 see... users are annoying, they expect web pages and web browsers to self correct Sep 02 11:15:43 Why the hell should corrupt images be supopported? Sep 02 11:15:45 offering them something instead of an error Sep 02 11:15:57 what do you mean "supported"? Sep 02 11:16:07 if an image is corrupt, you get a dom event and the image is replaced Sep 02 11:16:12 That browser should do anything else than stop rendering when one is encountered Sep 02 11:16:25 ?? Sep 02 11:16:32 the browser does lots of other things Sep 02 11:16:50 but the web app might want to pick an alternate image if the one it got from some server is broken Sep 02 11:16:59 the browser isn't a program, it's a platform Sep 02 11:17:10 Would you like GCC to start quessing stuff (which may or may not be wrong) for you when you have incorrect syntax? Sep 02 11:17:19 the web app is the program, and it's the browser's job to offer the web app full flexibility Sep 02 11:17:28 theril: you're confused Sep 02 11:17:34 web browsers aren't compilers Sep 02 11:17:44 they're translators Sep 02 11:17:53 What's the difference? Sep 02 11:17:54 <_Monkey> the difference is Archos TI chip is more powerfull DSP Sep 02 11:18:13 and yes, when i speak through an interpreter, i want feedback so i can try to pick another explanation if the translator can't figure out what i'm saying Sep 02 11:18:28 that'd be an interesting feature, auto-search for image based on page text content when you see a broken image Sep 02 11:18:38 it could get psychedelic Sep 02 11:18:41 _Monkey forget the difference Sep 02 11:18:41 <_Monkey> timelyx: I forgot difference Sep 02 11:18:46 _Monkey the difference is Sep 02 11:18:47 <_Monkey> OK, timelyx. Sep 02 11:18:55 theril: compilers are used by programmers, browsers are supposed to eat random crap from real people Sep 02 11:20:02 users have higher expectations than programmers Sep 02 11:20:09 oh, and they have money Sep 02 11:20:15 that helps too Sep 02 11:20:15 and they vote with their wallets Sep 02 11:20:44 programmers either know not to use gcc or don't have a choice Sep 02 11:20:51 :) Sep 02 11:23:42 So everybody has to suffer so that any moron can spit crap from Frontpage that should be rendered fine? Sep 02 11:24:02 define suffer Sep 02 11:24:04 but sure Sep 02 11:24:09 I don't think that people who don't know HTML write HTML anymore Sep 02 11:24:20 you clearly don't browse the web Sep 02 11:24:27 I think the official definition of suffering is to get any non-trivial website working with all browsers Sep 02 11:24:31 or you don't use anything interesting to check for web quality Sep 02 11:24:59 because it's more important that i can get access to poorly written and unmaintained content than that you have some pristine value Sep 02 11:25:03 And people would learn to write HTML if the parser/browser would just spit a parse error Sep 02 11:25:16 * timelyx chuckles Sep 02 11:25:20 chicken, meet egg Sep 02 11:25:36 and no, they won't Sep 02 11:25:41 validator.w3.org exists Sep 02 11:25:46 extensions exist to integrate it Sep 02 11:26:00 most decent web design toolsets include some sort of validator Sep 02 11:26:09 Yeah, lets drag some 1985 weight just to not to have to deal with chicken-egg Sep 02 11:26:34 Yes, but because the browser renders badly formatted stuff, the validator is essentially useless Sep 02 11:26:54 http://validator.w3.org/check?uri=http%3A%2F%2Fmaemo.org&charset=%28detect+automatically%29&doctype=Inline&group=0 Sep 02 11:27:14 And also as long as same HTML is rendered differently by browsers (especially with the current extent) Sep 02 11:27:19 http://validator.w3.org/check?uri=http%3A%2F%2Fmaemo.org%2F&charset=%28detect+automatically%29&doctype=XHTML+1.0+Transitional&group=0 Sep 02 11:27:55 timelyx: That is useless Sep 02 11:28:24 theril: what about it? Sep 02 11:28:27 As long as browsers accept to render that, that has no importance at all Sep 02 11:28:31 timelyx: Validating the code Sep 02 11:28:35 it's basic web 101, include ALT tags for all images Sep 02 11:28:43 maemo.org can't even get that right on their main apge Sep 02 11:28:47 s/apg/pag Sep 02 11:29:01 And that basic web 101 is idiocy Sep 02 11:29:10 * timelyx shrugs Sep 02 11:29:14 you were the fool using lynx Sep 02 11:29:19 or was it links? Sep 02 11:29:35 I use both from time to time Sep 02 11:30:05 But for me it doesn't matter if it says [IMG] or "Image of cat" Sep 02 11:30:07 don't go complaining about me complaining about lack of alt tags when you use multiple browsers who benefit from it Sep 02 11:30:23 that's because you don't care about content Sep 02 11:30:36 which makes you a user i don't need to support Sep 02 11:30:40 users care about content Sep 02 11:30:45 that's why they browse the web Sep 02 11:30:46 Yes, but I've done years of web development and demanding alt is just idiotic Sep 02 11:31:01 Ask how many cares about the freaking alt-text Sep 02 11:31:06 how often have you been sued under sec508? Sep 02 11:31:33 sp3000: did i get the number right? Sep 02 11:31:41 Never Sep 02 11:31:58 probaby Sep 02 11:31:59 l Sep 02 11:32:19 theril: try being blind for a while Sep 02 11:32:19 theril: the only one suffering is the poor programmer writing the browser Sep 02 11:32:27 it should open your eyes when you're done Sep 02 11:32:39 theril: nobody eles cares how bad things are, as long as it works Sep 02 11:33:13 timelyx: I do understand the accessibility issue, but forcing the alt tag doesn't solve it the least Sep 02 11:33:23 "Poor programmer" writing the browser? Spare me (apologies to timelyx) Sep 02 11:33:26 it's a start Sep 02 11:33:35 timelyx++ Sep 02 11:33:45 Well, forcing the alt would be ok, if CSS wouldn't be such crappy Sep 02 11:34:00 * timelyx doesn't follow Sep 02 11:34:01 So that you'd have to use the img-tag only for "content" images Sep 02 11:34:22 CSS is fine - it's idiotic web designers using images for text that cause problems in the CSS/no images combo Sep 02 11:34:34 Jaffa: CSS is far from fine Sep 02 11:34:39 theril: you mean if people could use css to make a page that required a 10000x9000 cnavas size? Sep 02 11:34:44 s/cna/can/ Sep 02 11:34:44 timelyx meant: theril: you mean if people could use css to make a page that required a 10000x9000 canvas size? Sep 02 11:34:47 oh wait, they can Sep 02 11:34:56 i'm not sure how that helps matters Sep 02 11:35:02 theril: CSS is fine in the situation you're describing. I'm well aware of all the limitations of CSS, having been fighting with IE6 yesterday Sep 02 11:35:16 in fact, that makes life difficult for embedders (a hat i get to wear this year) Sep 02 11:35:17 timelyx: So that you could really put content in the HTML and formatting in CSS Sep 02 11:35:36 theril: if people actually used css for formatting Sep 02 11:35:41 i'd be in much worse shape Sep 02 11:35:52 i'd have to figure out what to do w/ pages such as the one i described Sep 02 11:38:13 timelyx: I do as much as I can Sep 02 11:38:59 Well, the CSS spec sucks for both HTML and browser developers Sep 02 11:39:31 hildon is worse :) Sep 02 11:39:37 Floating mode has to be abused to get block elements side by side etc Sep 02 11:40:26 Hildon is about 10000x better than HTML/CSS specwise at least Sep 02 11:40:49 very hard to compare Sep 02 11:40:55 I can write a simple webpage Sep 02 11:41:07 I failed to write a simple hildon app when I tried Sep 02 11:41:24 I didn't follow any instructions for either feat Sep 02 11:43:27 I find doing it using plain xlib is easier, but that's just me Sep 02 11:43:44 hildon doesn't have a spec Sep 02 11:43:48 and it changes yearly Sep 02 11:43:59 plus you need to recompile every 6 months Sep 02 11:44:02 none of these are better Sep 02 11:44:07 xlib is better Sep 02 11:44:13 You are insane Sep 02 11:44:36 i care about having a product that doesn't require a rewrite every six months Sep 02 11:44:42 just because someone got clever and decided to break apis Sep 02 11:55:18 someone here have a normal gtk app w/ a scrollbar handy? Sep 02 11:55:34 theril: when are you going to pick up the monitor? Sep 02 11:59:59 lle2: Where can I pick it up from? Sep 02 12:01:19 http://science.slashdot.org/comments.pl?sid=286139&cid=20439513 Sep 02 12:23:55 lle2: comparing a webpage to an application is not really interesting. it's like comparing a leaflet to tv-series Sep 02 12:24:20 tko: that's why I said it's hard to compare Sep 02 12:24:25 or whatever bad analogy you'd want to use Sep 02 12:24:29 right Sep 02 12:25:46 timelyx: so how's the mozilla xlib/motif backend going? (speaking of stable APIs...) Sep 02 12:32:48 Does anybody know where I could get an adapter for N800 which would split the 3-pole audio plug to separate input and output plugs? Sep 02 12:33:23 I couldn't even find a handsfree with wires (and don't want to destroy the one that came with the device) Sep 02 12:34:32 Or even just a 3-pole jack Sep 02 12:51:07 <`0660> theril, i take you mean left and right channel? Sep 02 12:52:18 `0660: he means mic and headset Sep 02 12:53:39 <`0660> hmm, is it possible? Sep 02 12:57:47 oh man Sep 02 12:57:51 * timelyx curses Sep 02 12:57:55 osso sucks Sep 02 13:01:00 what did it do this time? Sep 02 13:01:37 the contacts status bar item is a sym link to the contacts task navigator item Sep 02 13:02:05 so in order to disable the task navigator item and get the status bar item Sep 02 13:02:21 it isn't sufficient to delete or rename the task navigator item Sep 02 13:02:29 you must actually move it to statusbar Sep 02 13:02:48 and then of course, it doesn't actually work Sep 02 13:02:56 let's not forget that detail Sep 02 13:03:28 * timelyx gives up Sep 02 13:09:45 ok, i'm down to two components Sep 02 13:09:50 people really don't want to help? Sep 02 13:10:40 what? sorry i want to help Sep 02 13:11:00 bugs.maemo.org, query applications:hardware or applications:core Sep 02 13:11:04 move the bugs to anywhere else Sep 02 13:11:08 ok Sep 02 13:11:10 preferably where they best belong Sep 02 13:11:24 for hardware, misdirected:nokia is often correct Sep 02 13:11:35 although sometimes systemsoftware:(xserver/dsme/...) is better Sep 02 13:11:51 for core, desktop/ or systemsoftware/ are probably best Sep 02 13:12:00 i have a long un posted comment for 1262 Sep 02 13:12:04 O_o Sep 02 13:12:32 https://bugs.maemo.org/buglist.cgi?product=Applications&component=Core&component=Hardware Sep 02 13:13:10 ok i'll schedule a block here - 19:00-21:00 CET Sep 02 13:13:22 sp3000: hrm.. Sep 02 13:15:35 sp3000: timely dcc? Sep 02 13:15:56 I usually give up on dcc Sep 02 13:16:01 try it Sep 02 13:16:38 * sp3000 looks for an irssi manual Sep 02 13:16:46 /dcc chat victim Sep 02 13:16:54 it doesn't get through the proxy to cz Sep 02 13:17:07 ? Sep 02 13:17:12 you're proxying chatzilla? Sep 02 13:17:40 cz talks to irssi-proxy Sep 02 13:17:59 ah Sep 02 13:18:30 ok, lemme know if that looks ok Sep 02 13:28:01 hola Sep 02 13:28:02 <_Monkey> hey, Blacksito Sep 02 13:28:13 hi _Monkey Sep 02 13:48:14 theril: hello, so are you coming? Sep 02 13:48:32 lle2: How are you home today? Sep 02 13:48:45 theril: right now, and I'm not going anywhere Sep 02 13:48:46 I'll have to finish one Linux installation first Sep 02 13:48:50 x) Sep 02 13:49:17 lle2: OK, I'll try to come before 19:00 Sep 02 13:50:03 theril: cool, you got my address? Sep 02 13:50:04 Oh, do you have a box for it or should I bring something myself (I'm coming by bus) Sep 02 13:50:20 no box, sorry Sep 02 13:50:49 lle2: I *had* it, but I don't have logging on and irssi got shut down when I pulled the power plug :P Sep 02 13:50:54 So could you resend pleasE? Sep 02 13:50:59 will do Sep 02 13:51:09 I'll come up with somekind of box Sep 02 13:52:01 I think you'd manage even without one, it's in one piece Sep 02 13:54:28 It just may start raining by look of the clouds Sep 02 13:55:34 that'll just add to the street credibility Sep 02 13:56:26 Yeah, and a broken monitor :P Sep 02 14:03:38 ok. all gone Sep 02 14:21:44 hi Sep 02 14:29:38 made by a genuinely retarded guy Sep 02 14:29:39 www.alternativetentacles.com/octopodes/483/u0HBOB9KbXNa1lR233g/Wesley_Willis-Rock_n_Roll_McDo.mp3 Sep 02 15:12:24 Hi, I hope you don't mind if I ask more device question than actual maemo one... Sep 02 15:13:22 I just got a N800 and the screen seems to get scratches very easily, so I'm considering a hard case and a screen protector. Recommendations? Sep 02 15:15:21 Lynoure: did you remove the film that it ships with? Sep 02 15:15:36 I've /seriously/ punished some of those displays and I can't get them to scratch Sep 02 15:16:31 lle2: this was used... I see some layer on the screen, but a part of it (about 2mm) seems to go under the bevel at the top... so I thought it was the touch sensor layer and decided not to try remove it with force Sep 02 15:17:10 hmm, in the lower left corner there should be a "flap" that you can get it off from Sep 02 15:17:21 maybe the original owner cut that off Sep 02 15:17:24 no flap, whatsoever Sep 02 15:18:13 I wish I could show this, but no way I can get that good macro shot Sep 02 15:19:12 it's really easy to get smudges that look like scratches Sep 02 15:19:57 But, when you guys look at the very edge of the screen, can you see a border about 1-2 mm from the edge? Sep 02 15:20:11 the display coverings are made out of a chemical with a long name that i forget Sep 02 15:20:19 that said I do have a few scratches but the smudges usually win Sep 02 15:20:31 I really do not want to put force in removing something that should stay there... Sep 02 15:21:08 So I need to first be sure it's something removable :) Sep 02 15:21:16 sure I can see the edge of the whole lcd panel, but the covering shouldn't go underneath the case Sep 02 15:22:12 lle2: this edge is inside the lit area... Sep 02 15:22:21 ever so slightly Sep 02 15:23:48 and I can feel my fingernail running along it, if I try that. The scratches are definately scratches, or at least microfiber cloth would not remove them Sep 02 15:24:05 hmm, maybe try it with something not too pointy but slim enough to fit there and try to lift it Sep 02 15:25:05 no way could I get a fingernail to the edge of the lcd, but that's mostly 'cause I ain't got any Sep 02 15:26:30 btw, do the testing at the bottom middle of the screen, least used area ;) Sep 02 15:27:23 lle2: you could feel the same edge with an edge of a paper or business card, if it there Sep 02 15:27:35 s/it/it is Sep 02 15:28:33 I only hit the metal edge of the lcd, under which the glass + ts films go Sep 02 15:29:39 if you're not sure about it, just leave it. easiest way to break the lcd is to press with a sharp object near the edge Sep 02 15:30:28 I need to do something... now it gets a scratch every time I draw a line with a stylus, no matter how light touch I use Sep 02 15:31:08 for sure those are not real scratches, I've hit one with a stylus so hard the stylus broke in pieces and there was no trace of it Sep 02 15:31:09 That's why I asked recommendations for screen protectors Sep 02 15:31:50 lle2: how do you define real scratches? I just rubber them with microfiber for 20 seconds, and they still seem real enough Sep 02 15:32:10 s/rubber/brubbed Sep 02 15:32:42 (typos, sorry) Sep 02 15:32:52 real scratch would be something you can feel when drawing a line across it with a stylus Sep 02 15:33:55 to me, real scratch is anything that I can see and feel on the screen, and that is not dirt. I wish they were Sep 02 15:35:26 I've had the n800 going around without any cover at all, for about 6months, and there are no scratches I could see when the backlight is on and it's showing something other than solid black Sep 02 15:37:58 I guess it must be then that the protective film is still on and it gets scratches easily... I'm waiting a verification from the ex-owner before pulling it off with a tape Sep 02 15:39:47 I've seen screens that went through 30 000 - 60 000 chinese characters to the same place by a robot and I have to look *very* carefully to see any sign of that load Sep 02 15:40:33 sure it affects the electrical properties, but it's very difficult to see it visually Sep 02 15:40:45 That's nice to hear Sep 02 15:40:45 cool Sep 02 15:41:43 Then, once I get that thing off thing, I'll only need a case, and no protector Sep 02 15:42:02 well as long as the protector is on, only the protector is getting scratched Sep 02 15:43:31 pupnik: yes, sounds so. I bet no pulling on the screen has been tested, though, so I hope this film is not too stickily on. Sep 02 15:47:39 http://www.internettablettalk.com/forums/showthread.php?t=4975 sounds like my story (minus the fact I never saw the tab) Sep 02 15:51:02 especially this bit sums up why I was wondering about whether it was the touchfilm: " I've looked closely at the factory 'screen protector' - there's a strange tiny tab at the top that disappears under the plastic rim (top mid-right) - never removed it as I feared the screen protector may actually be the touchscreen film itself!" Sep 02 15:52:25 we really should mark it with "remove before use" Sep 02 15:52:52 in big bold opaque letters diagonally across the whole screen Sep 02 15:53:55 Got it out now with the tape trick Sep 02 15:54:21 :) Sep 02 15:55:08 I bet that "protector" wears off pretty badly as it's only designed to protect the lcd during the manufacturing process Sep 02 15:56:47 same thing as some people leave that film in place for their mobile phones, can hardly see anything through after couple of months Sep 02 15:56:56 it seems many people managed to pull off the removal tab, but not the film Sep 02 15:57:11 heh Sep 02 15:58:06 I'll have to post the trick to that thread. A lot safer than trying to press in a fingernail Sep 02 16:14:13 Thanks, all, again :) Sep 02 17:06:57 jumpula: hey, do you know if chinook will be moving the debian-etch devkit? Sep 02 17:11:55 gah, i keep getting the "new version" message when starting pidgin. but its not available on the repos :( Sep 02 17:13:42 <`0660> try not start it :) Sep 02 17:14:17 :P Sep 02 18:29:44 lle2: The display is dead silent Sep 02 18:30:05 ? Sep 02 18:30:13 no way Sep 02 18:30:15 lle2: The display I "bought" from you Sep 02 18:30:32 Did you use DVI or VGA? Sep 02 18:30:35 you've connected the power cable? Sep 02 18:30:36 both Sep 02 18:30:42 Yes Sep 02 18:30:51 checked all the connections? Sep 02 18:31:30 does it light up the orange led in the power button? Sep 02 18:31:30 No, I mean that there's no extra sounds :P Sep 02 18:31:40 The picture is ok :) Sep 02 18:31:45 oh Sep 02 18:31:47 man Sep 02 18:33:40 I think I muted it from the menu Sep 02 18:34:05 theril: check the menu thingy for something related to audio Sep 02 18:34:51 So the sound comes from the speakers? Sep 02 18:34:57 yes Sep 02 18:35:19 or so I think Sep 02 18:35:25 maybe it has some psychic interface Sep 02 18:35:30 OK. I'm not planning to use it anyway Sep 02 18:35:38 it's better to keep it muted Sep 02 18:35:51 Yeah, maybe it has something personal against you Sep 02 18:35:51 I remember that it made hideous noise Sep 02 19:12:33 lle2: Do you know what the USB-port in the display is for? Sep 02 19:16:10 theril: there is a usb port? Sep 02 19:17:09 lle2: Yep, with a weird kind of connector Sep 02 19:17:17 Like one on some printers Sep 02 19:18:01 no idea Sep 02 22:01:54 ergh Sep 02 22:02:04 does maemo mapper have a "back" button somewhere? Sep 02 22:02:52 * sp3000 jumped with an address search from neuwied to ...uh... Sep 02 22:02:55 * sp3000 zooms out Sep 02 22:03:22 oh, portugal :D Sep 02 22:09:58 * sp3000 is priming maps for the german tour, can't exactly download maps on the fly can I Sep 02 22:10:14 well, or I guess it's either that or food Sep 02 22:18:19 hrm Sep 02 22:18:34 how do I add a poi in the middle of the view? Sep 02 22:18:42 or, place the, um, caret there Sep 02 22:19:08 * sp3000 finds it Sep 02 22:19:15 I think you just have to be accurate with the stylus Sep 02 22:19:32 I found the maemo mapper UI a bit weird Sep 02 22:19:40 it seems to be hidden only in the context menu Sep 02 22:19:46 Yes Sep 02 22:20:29 which is weird if I landed on the view with a text search, now I'm supposed to find the center of the screen manually :) Sep 02 22:20:39 IMHO it should pan the map like eg browser window does and have a caret/cursor Sep 02 22:21:07 And a toolbar would be nice Sep 02 22:23:11 Ah well, it has 54 active feature request tickets :) Sep 02 22:56:13 greeting maemo-ians. I am joining your ranks. Just unboxed my N800 and charging the battery Sep 02 23:00:56 check the version of firmware and upgrade Sep 02 23:02:11 I just downloaded the latest firmware.. but haven't checked to see what is installed yet. waiting for the battery to charge. it was almost completely dead Sep 02 23:24:01 also weird Sep 02 23:24:14 can't see a way to go to a poi Sep 02 23:24:24 except the nearest Sep 02 23:26:32 had 2007-3 updated to 2007-4 Sep 03 00:00:04 is there a problem with using the N800 with a WPA - Pre-Shared Key WLAN? Sep 03 00:08:25 works fine with open WAPs, but can't get it to work with my WPA Sep 03 00:08:44 I have it working with wpa psk Sep 03 00:11:18 maybe a limit in the length of the psk? I have 10 devices working on my WAP, but can't seem to get my n800 up Sep 03 02:33:21 who's in charge of planet maemo? Sep 03 02:34:09 My feed isn't getting updated. **** ENDING LOGGING AT Mon Sep 03 02:59:56 2007