| 00:49:36 | mibus | y'all shoulda been here 12 hours ago, it wasn't even slightly quiet :) |
| 00:52:26 | mibus | anyone installed appmgr 0.07 yet?? |
| 00:52:29 | mibus | i have :) |
| 00:54:47 | mibus | Zzzzzz.... :P |
| 01:21:00 | prpplague | anyone alive? |
| 01:21:21 | prpplague | is there a reason why i can't locate source for the gpe apps at gpe.handhelds.org? |
| 01:23:51 | prpplague | nm i'll grab it via cvs |
| 01:32:09 | prpplague | anyone got any suggestion why the gpe-contact app won't compile on x86 |
| 01:34:03 | prpplague | seems to be looking for 'parser.h" |
| 01:34:45 | flypiper | prolly the same reason I can't get mmenu to compile for x86 notebook.. something missing in png.h |
| 01:36:23 | prpplague | hmm, so how are you guys testing these apps? all on arm? |
| 01:37:48 | flypiper | I do |
| 01:39:22 | prpplague | that seems like a inefficient way of doing things |
| 01:39:42 | prpplague | besides what if someone wanted to run these apps on a webpad thats x86 based? |
| 01:40:59 | flypiper | excactly.. im working on matchbox on my notebook.. |
| 01:41:25 | flypiper | it's prolly a config someplace im overlooking |
| 01:41:54 | prpplague | flypiper: hmm, mmenu? seems like i had the same problem |
| 01:42:25 | flypiper | LOL... NO.. LOL |
| 01:42:41 | flypiper | always gotta work on it :-) |
| 01:47:42 | prpplague | flypiper: well looks like the .configure script is incorrectly call a xml2-config instead of xml-config |
| 01:48:10 | flypiper | your good..kewl |
| 01:49:41 | prpplague | flypiper: i don't think so, i'm just another joe smoe wannabie |
| 01:50:09 | flypiper | LOL..OK.. just trying to boost morale |
| 01:55:16 | prpplague | hmm, compilied and installed, all i get is "aborted" |
| 01:55:53 | flypiper | congrats it compiled |
| 01:58:56 | prpplague | this sucks |
| 02:08:16 | prpplague | there appears to be a # of include missing from the headers of libgpewidget |
| 08:22:19 | pb_ | morning |
| 08:22:49 | PaxAnima | morning |
| 09:03:04 | mallum | hey pb_ |
| 09:04:53 | pb_ | yo mallum |
| 09:05:09 | pb_ | been working on GTK 2 |
| 09:08:44 | mallum | cool, good progress ? |
| 09:08:57 | mallum | I had a cool idea last night for xkbd |
| 09:09:05 | pb_ | well, I have gpe-edit working with the new libraries. |
| 09:09:19 | mallum | pb_: excellent ! |
| 09:09:24 | pb_ | looking at gpe-todo and gpe-calendar now. |
| 09:09:50 | mallum | pb_: I believe dillo will just about compile against gtk2 aswell |
| 09:10:14 | pb_ | unfortunately both of those rely on the libgpewidget date selector widget, which is a bit of a nightmare to port - it's the first widget I ever wrote, and it's like some kind of black museum of coding horrors. |
| 09:10:31 | pb_ | so, I need to redo that first of all. |
| 09:10:35 | pb_ | what was your cool idea? |
| 09:10:50 | mallum | pb_: I worked out last night, it would not be too hard to write a perl/python script to parse xkeycap .h keyboard definetion files to produces *alot* of various layout for xkbd |
| 09:11:19 | pb_ | oh right, that would be rather nice. |
| 09:11:58 | mallum | pb_: you could have little mini sun keyboards and all sorts |
| 09:12:06 | pb_ | excellent! |
| 09:12:24 | mallum | pb_: now I have to pester a non-c coder to do it ;) |
| 09:12:36 | pb_ | try dc_ |
| 09:12:45 | pb_ | he seems to be some kind of python guru. |
| 09:13:10 | mallum | pb_: yeah my python is a little rusty - I could do it in Perl, but python would be better |
| 09:13:23 | mallum | pb_: python scripts always end up more maintainable |
| 09:13:57 | mallum | pb_: I need to post to fam list requesting any one post me any broken / missing keys with current xkbd layouts |
| 09:14:35 | pb_ | yah, good plan. |
| 09:14:48 | mallum | pb_: My other job at the moment is doing some work for a design collective |
| 09:14:58 | mallum | pb_: more php work :( |
| 09:15:11 | pb_ | :-/ |
| 09:15:20 | mallum | pb_: I need to suggest to them if they would like to maybe do icons etc |
| 09:15:29 | mallum | pb_: as there stuff looks pretty nice and clean |
| 09:16:08 | pb_ | ah, that would be cool. |
| 09:16:16 | pb_ | what kind of things do they work on normallu? |
| 09:17:50 | mallum | pb_: I'll find a link to there site |
| 09:18:21 | mallum | http://www.londondesigncollective.com/ |
| 09:20:34 | mallum | pb_: is 0.6 still on for release end of coming week ? |
| 09:21:18 | pb_ | not sure, we'd need to check with Jamey. |
| 09:21:26 | pb_ | I haven't heard anything to the contrary though. |
| 09:22:12 | pb_ | Does the mmenu in unstable include a .dock file? It still seems to be missing in 0.6. |
| 09:41:57 | mallum | pb_: 0.8 does have one |
| 09:42:04 | mallum | pb_: 0.8 of xkbd will too |
| 09:42:18 | mallum | pb_: as will xstroke and xmonobut |
| 09:42:40 | pb_ | okay, cool. |
| 09:42:43 | pb_ | any eta on those? |
| 10:17:34 | mallum | pb_: probaboly by about wednesday |
| 10:17:47 | mallum | pb_: depending on how much a can get done covertly |
| 10:17:57 | mallum | pb_: did you get a chance to test xkbd BTW ? |
| 10:20:01 | mallum | cool, looks like there is now a netbsd matchbox package |
| 10:20:04 | mallum | ftp://ftp.netbsd.org/pub/NetBSD/packages/pkgsrc/wm/matchbox/README.html |
| 10:20:44 | BBrox | mallum: want Python consulting :-) ? |
| 10:20:51 | pb_ | mallum: not yet, my ipaq is languishing at the office and I couldn't be bothered to go across town to pick it up. I'll have a go with xkbd tomorrow. |
| 10:21:19 | mallum | pb_: its fine to test on your desktop, just start it up and check you dont get the 'mode_switch' error |
| 10:21:43 | pb_ | mallum: Oh right. I'll try it in a couple of minutes then. |
| 10:22:01 | mallum | BBrox: Im 90% sure its possible to write a script which converts xkeycaps keyboards into xkbd keyboards |
| 10:22:23 | mallum | BBrox: you'll need to grab the xkeycaps source and look in the kbds directory |
| 10:22:29 | BBrox | mallum: xkeycaps.. The tool that displays a nice graphical keyboard on your screen ? |
| 10:22:32 | mallum | BBrox: yep |
| 10:22:43 | mallum | BBrox: the keyboards are defined in the source |
| 10:22:56 | BBrox | mallum: let me download the source :-) |
| 10:24:19 | pb_ | mallum: yay, gpe-todo (mostly) works with GTK 2 as well. |
| 10:24:29 | mallum | pb_: :)) |
| 10:25:03 | pb_ | mallum: looks like I should stop using a GtkDrawingArea for the main display, and switch to some kind of list of widgets. |
| 10:28:16 | BBrox | mallum: what do you want to convert ? Everything ? Geometry ? And why :-) ? |
| 10:29:39 | mallum | BBrox: yeah, dont think we need quite everything - I dont know how well the geometry will translate - Have a look at the existing xkbd layouts |
| 10:30:35 | BBrox | mallum: sources of xkbd somewhere ? |
| 10:31:13 | mallum | BBrox: HH.org cvs apps/xkbd |
| 10:33:12 | BBrox | mallum: the config are the data/*.in files ? |
| 10:34:44 | mallum | BBrox: yeah the kbdconfig.in files |
| 10:37:40 | BBrox | Will see what I can do :-) |
| 10:37:49 | pb_ | mallum: xkbd seems to work fine on my desktop |
| 10:38:04 | BBrox | (problem is that most keyboard in xkeycaps sport keypads and there is no easy way to remove them) |
| 10:38:24 | BBrox | Well, I cannot build it on my box... I have problem with autoconf :-/ |
| 10:38:40 | BBrox | ./ltconfig: ./ltconfig: No such file or directory |
| 10:40:51 | mallum | pb_: cool, and are you running xkbd extension ? |
| 10:41:05 | mallum | BBrox: did you run autoconfigure.sh first ? |
| 10:41:05 | pb_ | mallum: yeah, I think so. |
| 10:41:29 | mallum | pb_: thats good, Its broken on my laptop and it still works ok |
| 10:41:34 | BBrox | mallum: I think I had a too old version of libtoolize |
| 10:44:38 | mallum | BBrox: ok, grab a tarball for hh.org/~mallum/downloadables |
| 10:44:47 | mallum | BBrox: the format has not changed in CVS |
| 10:46:07 | BBrox | mallum: it's working now. Just had to go around some FreeType include problems :-) |
| 10:52:47 | mallum | BBrox: cool, CVS will now compile without xft gubbins |
| 11:52:40 | pb_ | mallum: I'm gonna build some GTK 2.0 packages for familiar. |
| 11:53:12 | mallum | pb_: nice |
| 11:53:24 | pb_ | be interesting to see how big they actually turn out. |
| 11:53:28 | mallum | pb_: how are you finding it speedwise BTW ? |
| 11:53:54 | pb_ | Fine, but I've been doing everything on my athlon machine so far. |
| 11:54:03 | mallum | cool |
| 11:54:22 | mallum | I got a 1.2 duron yesterday for 35quid, apparently it got a 200 fsb |
| 11:54:28 | pb_ | excellent. |
| 11:54:29 | mallum | havn't tried it yet |
| 11:55:09 | mallum | I got a new dv-->vga converter thingy for my graphics card but the second head is still blurred :( |
| 11:55:29 | mallum | ok, I gotta go to B&Q now - bbl |
| 11:56:36 | colin_ | Hi all. |
| 12:01:58 | colin_ | In a Makefile, how do I get the output of a shell command into a variable? |
| 12:02:17 | spung | pb_: i've been watching CVS messages... are you starting to port to gtk2? |
| 12:09:02 | colin_ | Argh, found it. Was running in the wrong directory... |
| 12:09:30 | colin_ | Will be back later at night, bye. |
| 12:36:29 | pb_ | spung: yes |
| 12:41:25 | spung | pb_: when are you planning to use just gtk2? is it still slow on arm like mallum said? |
| 12:42:20 | pb_ | spung: I don't know yet. Haven't done any testing on arm so far, though I'm building some gtk2 packages for familiar at the moment. |
| 12:42:42 | pb_ | spung: if it's still slow on arm we will just have to find ways to speed it up, I guess. |
| 12:43:13 | spung | pb_: nice, let me know when you make some progress |
| 12:43:37 | spung | meanwhile i'll look at your code and see what #ifdefs i need, i'll do the same |
| 12:43:51 | pb_ | ok, cool. |
| 12:44:35 | pb_ | libgpewidget and gpe-edit are GTK2 capable now. I have some patches for gpe-todo to make it kind of work, but it's a bit hacky. |
| 12:44:53 | pb_ | Simple stuff like gpe-login ought to "just work", though I haven't tried it yet. |
| 14:00:32 | PaxAnima | flypiper: hi |
| 14:01:16 | flypiper | PaxAnima: morningg |
| 14:01:27 | PaxAnima | flypiper: intimate was too big for the CF card :/ |
| 14:01:51 | PaxAnima | but I've found a way (I think), but usbnet makes installing very difficult |
| 14:02:06 | flypiper | the base image was too big? |
| 14:03:00 | PaxAnima | yes |
| 14:03:06 | PaxAnima | but usbnet is _bad_ |
| 14:03:33 | flypiper | wonder why.. mine is very stable? |
| 14:03:49 | PaxAnima | what kernel version do you use? |
| 14:04:39 | flypiper | i have 2.4.18 on the ipaq and 2.4.9-31 on notebook desktop |
| 14:04:56 | PaxAnima | what hh-version? |
| 14:05:23 | flypiper | the latest I think hh15 |
| 14:06:24 | flypiper | I have used it before all day long and not lost connection or dropped connection |
| 14:06:27 | PaxAnima | same here (I have 2.4.18) |
| 14:06:43 | PaxAnima | 2.4.18 on the desktop |
| 14:06:48 | flypiper | I don't use dhcp tho.. just static to notebook |
| 14:07:06 | flypiper | and usually I mv dhcpcd to dhcpcd-1 so it won't try to grab an ip |
| 14:07:38 | PaxAnima | I use static ips |
| 14:08:03 | flypiper | hmmm you have 3800 |
| 14:09:04 | PaxAnima | yes |
| 14:10:22 | flypiper | do you keep it in any kinda sleeve while in the cradle? I think mine doesn't stay on the connector like my 3700 does |
| 14:10:55 | flypiper | very easy to move the 3800 to drop the connection. I keep it on a single sleeve most of the time |
| 14:20:27 | PaxAnima | flypiper: use a pcmcia sleeve |
| 14:20:42 | PaxAnima | could be it |
| 14:28:28 | flypiper | dunno then |
| 14:29:07 | PaxAnima | haven't tried it yet... working on my pim at the moment.. |
| 14:30:26 | flypiper | sure..working on getting xfree86-4.2 on a box |
| 16:20:58 | mallum | hey |
| 16:23:07 | mallum | pb_: just killed my desktop putting that processor in |
| 16:27:36 | prpplague | mallum: howdy |
| 16:27:46 | prpplague | mallum: fun |
| 16:28:30 | prpplague | mallum: you seen this - http://www.tuxia.org/viper/index.html |
| 16:28:53 | pb_ | mallum: :-( |
| 16:29:02 | pb_ | mallum: what happened? |
| 16:32:02 | spung | prpplague: hmmm, does viper run on ipaq? |
| 16:32:50 | prpplague | spung: ya thats what i was designed for |
| 16:34:30 | pb_ | prpplague: hmm, none of the links seem to work. |
| 16:34:45 | spung | prpplague: could its look be improved? |
| 16:37:28 | spung | prpplague: is nano x better than tiny-x ? |
| 16:39:33 | mallum | pb_: I put the new cpu in, came up with some wierd error, saying I had to run softmenu |
| 16:39:41 | mallum | pb_: ( the bios program ) |
| 16:40:13 | mallum | pb_: but the machine just seemed to lock, then after a few more power cycles it just did nothing ( fans come on ) but no vga |
| 16:40:31 | mallum | pb_: put the old cpu back in and still nothing :( |
| 16:40:35 | pb_ | gutted! |
| 16:40:47 | mallum | prpplague: that vipers been around for a while ... |
| 16:40:58 | pb_ | sounds very bizarre, I can't think of any obvious way a cpu could damage the rest of the board. |
| 16:41:10 | pb_ | have you tried clearing the CMOS RAM, in case it's in a weird state? |
| 16:41:19 | mallum | pb_: yeah tried that |
| 16:41:39 | mallum | pb_: YYou just switch the jumber, and then switch it back right ? |
| 16:42:03 | pb_ | Yeah. You might have to leave it switched for a little while (say 30 seconds to a minute). |
| 16:42:21 | mallum | pb_: while the machine is off right ? |
| 16:42:24 | pb_ | Right. |
| 16:42:43 | mallum | ok lets leave it off for a while |
| 16:43:39 | prpplague | the links are dead due to the fact that tuxia is out of biz |
| 16:43:50 | pb_ | oh right. |
| 16:43:51 | prpplague | but i have gotten a copy of all the latest source |
| 16:43:58 | pb_ | is it free software? |
| 16:44:15 | mallum | prpplague: didn't they change into a different co ? |
| 16:44:31 | mallum | prpplague: I was always waiting for there xlib only mozilla |
| 16:44:32 | prpplague | spung: as far as nano being better than x its tit-for-tat in my book |
| 16:44:48 | prpplague | spung: i'll just compile them as fltk apps instead of flnx |
| 16:44:53 | mallum | pb_: how much did you pay for your new motherboard cpu etc ? |
| 16:45:14 | prpplague | mallum: don't know about the company name, i've been looking for information all morning |
| 16:45:21 | spung | mallum: that'd be nice |
| 16:45:44 | spung | prpplague: tit-for-tat? sorry.. |
| 16:46:04 | pb_ | mallum: 51ukp for the motherboard, 63.18 for the cpu, 54.00 for the PC2700 sdram, and I think about another 45 for the PC2100 sdram. |
| 16:46:06 | pb_ | all plus VAT. |
| 16:46:15 | pb_ | came to about 250 quid. |
| 16:46:51 | pb_ | which reminds me, _still_ no sign of a new credit card. I think the bank must have forgotten about me. |
| 16:47:52 | prpplague | spung: sorry american slang, the good things on one offset the bad on another and vise versa |
| 16:48:27 | spung | prpplague: OIC |
| 16:54:41 | mallum | pb_: hmm cmos reset made no difference :( |
| 16:54:51 | mallum | pb_: I think I fixed the video card too :( |
| 16:57:50 | pb_ | mallum: hell. |
| 16:58:14 | pb_ | mallum: What wattage is your PSU rated at? I wonder if you toasted that. |
| 16:59:01 | mallum | pb_: how could I have done that ? |
| 16:59:21 | pb_ | mallum: new cpu requiring more power. |
| 16:59:49 | mallum | pb_: what and it toasted the motherboard ? |
| 17:00:56 | pb_ | mallum: well, anything's possible, but blowing up the PSU won't necessarily damage anything else. |
| 17:01:39 | prpplague | mallum: you know of any mirrors for viper stuff? |
| 17:01:52 | mallum | prpplague: no ... sorry |
| 17:02:00 | prpplague | mallum: ok thanks |
| 17:02:23 | prpplague | mallum: btw, i was able to do ifdefs on the pop menus on mdock |
| 17:02:35 | prpplague | mallum: i'm testing to make sure i didn't break anything |
| 17:02:37 | mallum | prpplague: cool, do you want to send me a patch ? |
| 17:03:13 | prpplague | mallum: ya probably in a day or two |
| 17:03:31 | mallum | prpplague: cool |
| 17:10:48 | mallum | pb_: *sigh* I guess I will have to try and buy a new motherboard tommorow, but I really dont want to have to shell out for new RAM etc |
| 17:11:28 | pb_ | mallum: Yeah. :-( |
| 17:14:12 | pb_ | hmm, has anybody had any thoughts on syncing recently? Now that we (more or less) have the new database sorted out, syncing is the main thing that we still need in order to stop the opie guys laughing at us. |
| 17:15:28 | spung | pb_: ohh.. news from dave? |
| 17:15:37 | pb_ | spung: still nothing. |
| 17:15:37 | spung | pb_: what did opie guys decide to do? |
| 17:15:46 | pb_ | spung: they seem to be using kitchensync. |
| 17:16:07 | spung | pb_: but you say we have the new database sorted out |
| 17:16:21 | pb_ | spung: well, I said more or less :-) |
| 17:16:28 | pb_ | spung: gpe-todo and gpe-contacts are both using the "new style" now. |
| 17:16:40 | pb_ | spung: gpe-calendar still needs to be converted. |
| 17:17:00 | spung | right |
| 17:17:43 | pb_ | I might try to fix up gpe-calendar next week. |
| 17:20:03 | pb_ | oh, also, gpe-contacts and gpe-todo are still using separate databases; they probably ought to be cut over to using a combined one in ~/.gpe/pim or something. |
| 17:20:55 | pb_ | spung: oh, did you mean "what did the opie guys decide to do about the database"? |
| 17:23:49 | spung | pb_: yes |
| 17:24:15 | pb_ | spung: sorry, I thought you were talking about syncing. They still haven't decided for sure. I sent Max some more information about ours this morning. |
| 17:33:03 | mallum | pb_: kitchensync ? |
| 17:33:14 | pb_ | mallum: it's the KDE sync architecture, apparently. |
| 17:33:42 | mallum | spits |
| 17:33:59 | spung | lol |
| 17:34:29 | mallum | Im writing a game next |
| 17:34:34 | pb_ | cool! |
| 17:35:32 | spung | mallum: what kind of game? |
| 17:35:57 | mallum | spung: The project is in skunk mode, I cant tell you anything ;P |
| 17:37:01 | spung | lol |
| 17:41:20 | pb_ | hurrah, gtk build is finished. |
| 17:42:03 | mallum | woohoo |
| 17:42:23 | pb_ | so, file sizes are: |
| 17:42:27 | pb_ | -rwxr-xr-x 1 pb phil 1687608 Aug 4 19:47 libgtk-x11-2.0.so.0.0.6 |
| 17:42:32 | pb_ | -rwxr-xr-x 1 pb phil 50692 Aug 4 19:47 libgdk_pixbuf-2.0.so.0.0.6 |
| 17:42:35 | pb_ | -rwxr-xr-x 1 pb phil 302424 Aug 4 19:47 libgdk-x11-2.0.so.0.0.6 |
| 17:42:45 | pb_ | -rwxr-xr-x 1 pb phil 347144 Aug 4 14:47 libglib-2.0.so.0.0.6 |
| 17:42:46 | BBrox | Does it contain an integerized version of the gdk-pixbuf ? |
| 17:42:49 | pb_ | -rwxr-xr-x 1 pb phil 10236 Aug 4 14:47 libgmodule-2.0.so.0.0.6 |
| 17:42:53 | pb_ | -rwxr-xr-x 1 pb phil 154976 Aug 4 14:47 libgobject-2.0.so.0.0.6 |
| 17:42:54 | mallum | BBrox: yes |
| 17:42:58 | pb_ | -rwxr-xr-x 1 pb phil 14792 Aug 4 14:47 libgthread-2.0.so.0.0.6 |
| 17:43:01 | BBrox | Cool :-) |
| 17:43:27 | mallum | pb_: there are some other loader libs too I believe |
| 17:43:27 | BBrox | And are some never used widgets removed ? |
| 17:43:32 | BBrox | libpango and libatk |
| 17:43:50 | pb_ | mallum: yes, there are. I plan to only ship the png, xpm and (maybe) jpeg loaders by default. |
| 17:44:07 | pb_ | mallum: there are also a variety of input method modules. I don't plan to ship any of those in the main package. |
| 17:44:29 | BBrox | My opinion would be to remove all the 'obsolete' routines. If people want to use Gtk2, they need to use the new Lists and Text widgets :-) |
| 17:45:57 | pb_ | BBrox: yeah, but that would make porting much harder. Virtually every GPE program uses at least one obsolete widget, and quite a few of them need to be built with GTK_ENABLE_BROKEN at the moment. |
| 17:46:24 | pb_ | BBrox: haven't compiled pango and atk yet, so I don't have accurate file sizes for those ones. |
| 17:47:42 | pb_ | BBrox: looks like atk is about 100KB. Pango I'm not sure about, there are four libraries and I don't know if all of them are required. |
| 17:47:47 | pb_ | -rw-r--r-- 1 root root 214608 Apr 15 02:32 /usr/lib/libpango-1.0.so.0.0.1 |
| 17:47:47 | pb_ | -rw-r--r-- 1 root root 192996 Apr 15 02:32 /usr/lib/libpangoft2-1.0.so.0.0.1 |
| 17:47:47 | pb_ | -rw-r--r-- 1 root root 49416 Apr 15 02:32 /usr/lib/libpangox-1.0.so.0.0.1 |
| 17:47:47 | pb_ | -rw-r--r-- 1 root root 131080 Apr 15 02:32 /usr/lib/libpangoxft-1.0.so.0.0.1 |
| 17:49:51 | mallum | pb_: did you configure without X input support ? |
| 17:50:35 | mallum | pb_: ideally we want jg to package up the xft2 stuff ( which I think he is working on ) and just use that for pango |
| 17:51:14 | mallum | pb_: for now I think you wont need it, but probably best to include the other 3 |
| 17:53:13 | BBrox | pb_: I thought that Gtk2 won't work without pango ? |
| 17:53:20 | pb_ | mallum: I've no idea about X input support. I didn't put any explicit option, so I guess I got the default. What should I have done? |
| 17:53:27 | pb_ | BBrox: that's right, pango is required. |
| 17:53:33 | BBrox | Well, in my opinion, XInput is useless on the iPAQ |
| 17:54:10 | mallum | pb_: during the configure you need to --disable it - I dont think it will even work on the ipaq with it enabled .... /me hides |
| 17:54:25 | pb_ | mallum: ah. |
| 17:54:29 | pb_ | mallum: okay, I will rerun the build. |
| 17:54:38 | BBrox | And well, libpangoft2 and libpangoxft are a bit the same |
| 17:54:48 | pb_ | ? |
| 17:55:04 | BBrox | One uses Render + FreeType2, the other uses Xft |
| 17:55:25 | BBrox | Both are to display True Type fonts but in two different ways (AFAICT) |
| 17:55:32 | pb_ | mallum: is it XIM I want to turn of, or is that something else? |
| 17:56:32 | BBrox | XIM == XInput |
| 17:56:46 | BBrox | (ie tablets and such things) |
| 17:56:46 | pb_ | okay. |
| 17:57:01 | pb_ | so, my command line is: |
| 17:57:06 | pb_ | CXX=g++-3.1 CC=gcc-3.1 CFLAGS="-mcpu=strongarm110 -Os -fomit-frame-pointer -frename-registers" ./configure --disable-debug --prefix=/usr --sysconfdir=/etc --enable-integer-pixops --disable-xim |
| 17:57:07 | mallum | BBrox: xft is a combo of render and freetype |
| 17:57:11 | pb_ | any other advances before I launch it? |
| 17:57:39 | mallum | pb_: you integer patch got integrated ? |
| 17:57:48 | BBrox | pb_: you do not need an march=.... when you put the mcpu=strongarm110 thingy ? |
| 17:58:04 | BBrox | mallum: yup, but at the X side |
| 17:58:15 | pb_ | mallum: no, I haven't sent it to the gtk gurus yet. |
| 17:58:23 | pb_ | BBrox: -mcpu=strongarm110 implies -march=armv4 |
| 17:58:27 | BBrox | 'k |
| 17:59:00 | pb_ | actually, I don't know why I did that, "-march=armv4" would have been less typing, but still. |
| 17:59:36 | BBrox | pb_: is there any way to ask GCC to *NOT* unroll multiplications (ie to use 'mul' instead of adds and shifts) ? |
| 18:00:16 | pb_ | BBrox: hmm, I'm not sure. I guess -Os ought to suppress that transformation really, but I don't know if it does at the moment. |
| 18:00:51 | BBrox | pb_: because it's often faster (for small multiplicands) to do a 'mul' than the (most often) three lines of ASM for the shifts and adds |
| 18:02:19 | pb_ | BBrox: mmm, right. |
| 18:02:40 | pb_ | BBrox: I did a small test: int f(int a) { return a * 0x150; } and gcc does seem to do that with shifts and adds even at -Os. |
| 18:03:01 | BBrox | This is why my ASM version of the UV->RGB routine is much faster than the C code.... I should try a C version with inline ASM to force usage of 'mul'. |
| 18:03:31 | BBrox | And on StrongARM, 'mul' can be one cycle if the second register is 'short' enough (do not remember the number of bits though) |
| 18:04:12 | pb_ | BBrox: can you send me a test case where mul would be faster? |
| 18:04:38 | BBrox | pb_: hunting for my Intel Application Note on SA-110 ASM performances :-) |
| 18:06:20 | pb_ | BBrox: okay, cool. It may just be that the compiler needs its ideas of relative instruction costs tweaking a bit. |
| 18:08:08 | mallum | BBrox: have you done anywork lately of BBplay ? |
| 18:09:38 | BBrox | pb_: well, if Rs is using less than 11 bits, the multiplication will be in one cycle (from what I understood in the application note). Maybe it could be two (due to some pipeline problems for the next instruction). One should do tests to check :-) |
| 18:09:49 | BBrox | mallum: nope |
| 18:11:07 | pb_ | BBrox: okay, do you want to look into that? |
| 18:11:36 | pb_ | BBrox: if you can come up with a testcase where the compiler uses shifts and mul would definitely be faster, I'll have a look at fixing gcc. |
| 18:12:01 | BBrox | pb_: okay, will try to find a case for you.... |
| 18:12:05 | pb_ | cool. |
| 18:12:37 | BBrox | pb_: the biggest problem I think is that 'mul' scheduling is non-trivial whereas code with adds and shifts can be easily inserted in any code path without problems.... |
| 18:12:49 | pb_ | regardless of that, I'll try to make it a bit more enthusiastic about "mul" at -Os, but that probably won't help much in your application. |
| 18:14:11 | BBrox | For example, two 'muls' one after the other. In that case, you provoke a pipeline stall for the second one (as there are some 'problems' with the execute and buffer pipeline stages) |
| 18:19:41 | pb_ | BBrox: right, yeah. |
| 18:20:42 | pb_ | BBrox: the scheduler does have a relatively sophisticated model of the cpu though, it knows about each individual function unit and what the issue/result delays are. So it ought to be possible for it to deal with that kind of situation. |
| 18:21:35 | pb_ | BBrox: for example, recent compilers do a pretty good job of filling the strongarm's load delay slot. I don't know how well they will do with XScale - it'll be interesting to find out. |
| 18:22:53 | pb_ | back later |
| 18:58:30 | cmarqu | re |
| 19:22:08 | BBrox | Hmmmm, is there any way to accurately count cycles on an iPAQ ? |
| 19:25:42 | mallum | BBrox: put spokey dokeys on them ? |
| 19:27:47 | BBrox | pb_[beer]: found some case where it takes one cycle to 'mul' and four when doing the GCC way... I definitely think that this could be optimized out (ie faster code + smaller code, it's a win everywhere :-) ) |
| 19:56:15 | spung | mallum: can i ask you a gtk question? |
| 20:09:48 | cmarqu | pb_[beer]: Is that helpful?: http://www.handhelds.org/pipermail/ipaq/2002-January/011238.html (also see my followup) |
| 20:14:56 | mallum | spung: I dont really know much about gtk |
| 20:15:27 | BBrox | spung: what do you want to know ? |
| 20:16:10 | spung | =) |
| 20:16:20 | spung | BBrox: it's about gtkrc files |
| 20:16:35 | spung | can i paste mine in your query? |
| 20:17:57 | BBrox | spung: ah, this is one part of GTK that I do not know about :-/ Sorry :-) |
| 20:18:23 | spung | heh |
| 20:19:04 | spung | BBrox: it's that i set a GtkScrolledView's bg_pixmap[NORMAL] to <parent> but it gets black |
| 20:19:41 | spung | BBrox: i wonder if i have to change GtkScrolledView's parent or something? |
| 20:24:51 | cmarqu | pb_[beer]: gpe-todo now complains that it can't find icon "hide" |
| 20:27:19 | mallum | BBrox: did you look at the xkeycaps code ? |
| 20:28:38 | BBrox | pb_[beer]: in case I am in bed when you come back from your beer, '692' produces 5 lines of ASM to do the multiplication whereas it takes one cycle to do a mul rx, ry, 692. Now, as you need to load the constant, the gain is only 3 cycles at most, but well, in a loop, it could even be 4 :-) |
| 20:29:00 | BBrox | mallum: looked at it quickly, did not do anything yet :-) |
| 20:31:44 | mallum | BBrox: ok np |
| 20:54:06 | BBrox_ZZzz | => book :-) |
| 20:58:05 | flypiper | mallum: finally got mmenu to compile on my notebook but still don't have menu button showing:( |
| 20:58:35 | flypiper | mallum: didn't have libpng-dev installed |
| 20:59:26 | spung | flypiper: send me your notebook, i'll compile it for you. i don't promise to send it back though |
| 21:02:25 | flypiper | Im starting to understand your nick:-( |
| 21:02:45 | pb_ | ? |
| 21:02:59 | flypiper | spung: you wouldn't want it... compaq 1530..133 |
| 21:03:05 | spung | flypiper: ohh... it's not supposed to mean anything |
| 21:03:12 | cmarqu | Hi pb_ |
| 21:03:17 | pb_ | hey cmarqu |
| 21:03:21 | spung | flypiper: heh, it's better than mine, which is non-existant |
| 21:03:49 | flypiper | finally has wireless working :) |
| 21:04:48 | moray | spung: I've got a laptop you could have; it's a 46 with a whole 4 MB of memory (and Debian installed)... |
| 21:05:25 | moray | spung: I think there's a 2?86 portable around here too... |
| 21:05:47 | spung | moray: well, if you're serious i'll give you my address :) |
| 21:06:24 | moray | spung: I'm really sure that there must be a better machine closer to you for free/very little |
| 21:06:37 | spung | heheh |
| 21:06:45 | spung | yeah, i guess so |
| 21:08:59 | flypiper | spung: i think that's why I love debian so much.. it runs on all my very old stuff..well |
| 21:09:37 | pb_ | cmarqu: ha, thankyou for the SDRAM information. |
| 21:09:39 | spung | flypiper: well, i use zipslack in my 386 (my first computer) |
| 21:09:56 | spung | flypiper: it was the easiest way to get linux there |
| 21:09:58 | moray | flypiper: though newer Debian versions wouldn't support 4MB machines |
| 21:10:00 | flypiper | spung: kewl |
| 21:10:23 | flypiper | moray: your right there |
| 21:10:37 | pb_ | cmarqu: I don't know what's going on with gpe-todo: it's had a "hide" icon for ages. Maybe it just isn't getting shipped in the package or something. |
| 21:13:00 | cmarqu | pb_: Yeah, in my package is only /usr/share/pixmaps/gpe-todo.png |
| 21:13:18 | pb_ | righto, I'll look at that. |
| 21:13:52 | pb_ | cmarqu: looks like hide.png is supposed to be shipped in gpe-icons. |
| 21:14:23 | cmarqu | pb_: Also, this paintbrush I remember from the older package: is that "delete"? A trashcan would be better I guess... I accidentally deleted an item on my todo list. |
| 21:15:03 | pb_ | cmarqu: yeah, that was one of Joe's innovations; I'm not quite sure where the paintbrush came from. |
| 21:15:17 | pb_ | cmarqu: it probably want to have an "are you sure" dialog, since the effect is irreversible. |
| 21:15:24 | cmarqu | Right. |
| 21:16:12 | cmarqu | pb_: The translation package creation stuff is almost there. Know an easy way to transfor "pt_PT" into "pt-pt"? |
| 21:16:44 | pb_ | tr A-Z_ a-z- |
| 21:16:58 | cmarqu | Cool, thanks. |
| 21:17:22 | cmarqu | Package gpe-icons (0.07) installed in root is up to date |
| 21:17:28 | cmarqu | I guess it's too old. |
| 21:17:59 | pb_ | well, that does seem to be the latest version. |
| 21:18:35 | pb_ | maybe we just need a new release of the icons. |
| 21:18:43 | cmarqu | Apparently. |
| 21:19:14 | cmarqu | Most packages should have a new version before 0.6, if only for the fixed sections. |
| 21:20:12 | cmarqu | pb_: I saw you were asking about sync... SyncML uses vcard and vcalendar IIRC, so we should be able to export/import these as a first step. |
| 21:20:40 | pb_ | cmarqu: okay, I'll make it my mission to generate fixed packages tomorrow. |
| 21:20:52 | pb_ | cmarqu: do we have a definitive list of what packages need updates? |
| 21:21:03 | cmarqu | pb_: Hopefully I didn't break anything... :) |
| 21:21:11 | pb_ | cmarqu: yeah, I need to read up on SyncML. |
| 21:21:21 | cmarqu | pb_: Let me grep throught the changelog... |
| 21:21:26 | pb_ | okay, cool. |
| 21:22:11 | pb_ | I know we need to rev gpe-login and/or session-scripts to get Moray's logout confirmation thing. |
| 21:22:41 | cmarqu | Ah, yeah. |
| 21:23:03 | cmarqu | Also, gpe-ownerinfo isn't shown when locking is in effect. |
| 21:23:13 | pb_ | That is in CVS now (though, moray, you need to add a ChangeLog entry for it) |
| 21:23:35 | pb_ | cmarqu: right, that still needs code changes, I think it will have to wait until post 0.6. |
| 21:23:43 | cmarqu | Okay. |
| 21:23:50 | cmarqu | News about 0.6 BTW? |
| 21:23:57 | pb_ | Nothing that I've heard. |
| 21:24:45 | cmarqu | gpe-calendar, gpe-kbd, gpe-kbd, gpe-login, gpe-soundbite, gpe-todo, gpe-edit, gpe-su, gpe-terminal, gpe-sketchbook, isgui |
| 21:24:56 | cmarqu | These have my section and icon fixes. |
| 21:25:02 | pb_ | The other thing someone needs to do is make a list of packages that need to be pushed from unstable into 0.6. |
| 21:25:02 | cmarqu | and/or actually. |
| 21:25:40 | cmarqu | Add gpe-tetris to that "fixed" list. |
| 21:26:04 | pb_ | cmarqu: okay, fine. gpe-todo 0.20 is presumably sorted already, but I'll do the others tomorrow. |
| 21:26:19 | cmarqu | pb_: All packages that the "gpe" package depends on should be moved, no? |
| 21:26:58 | pb_ | Probably. But we need a concrete list of which ones are actually out of date to give to Jamey. |
| 21:29:49 | moray | obviously it would be a pain if GPE updates are left out of 0.6 unintentionally... - presumably the previously suggested date is pretty near now |
| 21:30:06 | pb_ | right. |
| 21:32:05 | cmarqu | Ideally, we'd have a script for that. |
| 21:32:51 | cmarqu | Not totally simple. |
| 21:35:09 | moray | have we got a solution to the xstroke problem yet? or something to give a default dock setup? |
| 21:35:25 | cmarqu | If ipkg had a switch like "--dry-run", it would be rather simple. |
| 21:35:31 | pb_ | I think mallum is still working on that. |
| 21:40:27 | cmarqu | mibus asked on the ML why people can even close gpe-appmgr... why can they? |
| 21:44:47 | pb_ | cmarqu: it may very well be that they can't. None of the bug reports we had on that were especially lucid. |
| 21:45:47 | cmarqu | I see it exiting when I'm OOM. minisys shows about 80% mem used when it starts killing stuff |
| 21:46:00 | cmarqu | I guess it's just that. |
| 21:47:08 | pb_ | Could be. |
| 21:47:39 | pb_ | Certainly, if there was an OOM situation, that would explain why X and/or matchbox was going away too. |
| 21:48:47 | cmarqu | Oh, I see it's not even in ImminentZeroPointSixRelease anymore. Good. |
| 21:49:22 | pb_ | Yah. |
| 22:02:06 | mallum | pb_: yah, we gotta turn xstroke-ma into xstroke |
| 22:02:27 | mallum | pb_: I'll add a menu and dock entry and upload it as a new xstroke version |
| 22:02:32 | pb_ | excellent. |
| 22:03:12 | mallum | pb_: I'll add some icons to the mini* dockmenu entries and then I think 0.3.3 matchbox is ready |
| 22:03:16 | cmarqu | "# |
| 22:03:19 | cmarqu | mmenu needs to ship a .dock file." |
| 22:03:28 | mallum | what ? |
| 22:03:38 | cmarqu | mallum: From http://handhelds.org/z/wiki/ImminentZeroPointSixRelease |
| 22:03:51 | cmarqu | Don't know who wrote that. |
| 22:03:56 | pb_ | me. |
| 22:04:08 | mallum | whats a .dock file ? |
| 22:04:18 | cmarqu | The package cannot really ship it, can it? Not forevery user. |
| 22:04:21 | prpplague | mallum: you been testing CosmicPenguins parser? |
| 22:04:34 | pb_ | mallum: /usr/share/matchbox/mbdock/... |
| 22:04:40 | prpplague | mallum: i installed it friday but didn't have time for testing..... |
| 22:04:48 | pb_ | cmarqu: why not |
| 22:04:49 | pb_ | ? |
| 22:05:01 | mallum | pb_: oh right - yeah 0.8 has one |
| 22:05:04 | pb_ | mallum: I think we talked about that this morning. |
| 22:05:09 | pb_ | yeah. |
| 22:05:26 | moray | cmarqu: you wouldn't want to install it into user homedirs, but you could have a default |
| 22:05:36 | mallum | sorry though you meant writing to .mbdock |
| 22:05:39 | cmarqu | Ah, okay/ |
| 22:05:42 | mallum | which is bad |
| 22:06:10 | cmarqu | ISTR suggesting that at one point, but it was dismissed :) |
| 22:06:14 | mallum | prpplague: not yet - I gotta get 0.3.3 out |
| 22:06:28 | mallum | prpplague: then I'll start breaking stuff in CVS |
| 22:06:41 | pb_ | mallum: uh, yeah, I don't know exactly where the ".dock file" terminology came from. :-| |
| 22:11:05 | mallum | pb_: ok I updated the wiki page |
| 22:12:12 | moray | ok, I have a patch for gpe-appmgr that makes the window go to the back when the close button is pressed - would that be an acceptable compromise? |
| 22:13:47 | spung | moray: i like it. btw, did my messages to you go to /dev/null? |
| 22:14:32 | moray | spung: er, pretty much - now I check, I find they're in an xchat tab somewhere offscreen that way --> |
| 22:16:04 | spung | moray: well, xchat sucks in that aspect |
| 22:16:40 | mallum | pb_: hmm, I wonder whats the easiest way of adding a question mark app to a defualt docked app |
| 22:16:46 | moray | I'd turned on beeping on PMs before, but apparently not saved the options before the powercut the other day |
| 22:16:55 | mallum | ack, I mean empty brand new dock after install |
| 22:17:42 | mallum | pb_: is there some 'global' gpe postinst script that can do a echo "monolaunch ..." > ~/.mbdock ? |
| 22:17:55 | pb_ | mallum: not really. |
| 22:18:17 | cmarqu | mallum: ~ would be /root/ always... :/ |
| 22:18:43 | mallum | yeah I know .... |
| 22:19:18 | moray | mallum: if you want to do things that way it would (perhaps) be better to have a session script that created an .mbdock if none exists. I'd still prefer the dock to look in some location for a default if none exists, though... |
| 22:19:19 | pb_ | mallum: we might be able to hack something into Xsession: test for ~/.mbdock empty or non-existent, and fire up monolaunch. |
| 22:20:06 | moray | I'd say we really need to have mmenu in the dock by default, or people will just nag with confused questions the whole time |
| 22:20:34 | pb_ | hmm, really? Even if they have gpe-appmgr for launching? |
| 22:20:42 | moray | probably xstroke, xkbd too; maybe even the clock and volume control... |
| 22:20:47 | spung | moray: indeed |
| 22:21:17 | mallum | moray, pb_: yeah session script would make sense |
| 22:21:30 | spung | mmenu | xkbd | xstroke --------- volume | battery | clock |
| 22:21:46 | spung | that'd be a nice default |
| 22:21:55 | pb_ | mallum: I must admit, I quite like moray's idea of having the dock search for a default session if ~/.mbdock doesn't exist. |
| 22:22:32 | moray | pb_: well, I could maybe be persuaded we don't need the menu, but at present I'd prefer to have it there so people can use it if they want - it's obviously very easy to remove it |
| 22:22:44 | mallum | pb_: yeah except mbdock dont know what you've got installed |
| 22:23:17 | pb_ | hmm, true. |
| 22:23:25 | moray | mallum: well, we could have some other gpe package install the default if you really wanted - but to be honest I doubt it does much harm to try to run things anyway |
| 22:23:45 | pb_ | okay, well, any volunteers to add this to xsession.d? |
| 22:24:18 | mallum | pb_: how about a gpe-newuser-setup type script ? |
| 22:24:31 | mallum | pb_: which sets some nice defulats for the new user ? |
| 22:24:37 | moray | pb_: I could in a moment (just checking this appmgr thing) |
| 22:24:48 | moray | can we check for a .gpe existing as the 'new user?' test? |
| 22:24:52 | mallum | pb_: it could be scripted with (gtk|x|*)dialog |
| 22:25:03 | moray | mallum: well, maybe later ;) |
| 22:25:27 | pb_ | moray: I don't think checking for ~/.gpe is going to be completely reliable. |
| 22:25:34 | moray | pb_: but is anything? |
| 22:25:56 | moray | pb_: if you e.g. check for .mbdock, you'll really irritate people who try to delete it :) |
| 22:26:00 | pb_ | moray: well, as far as installing a default .mbdock goes, I would just do that if ~/.mbdock doesn't exist and matchbox is installed. |
| 22:26:15 | mallum | I got to go to bed |
| 22:26:16 | moray | (see above) |
| 22:26:25 | pb_ | moray: people who want an empty dock can set it to zero length, or not run mbdock in the first place. |
| 22:26:48 | mallum | I'll dream about it ( lucidly ) |
| 22:26:50 | mallum | night all |
| 22:26:52 | pb_ | night |
| 22:26:54 | moray | night mallum |
| 22:36:36 | prpplague | you guys got an ipkg for sqlite libs somewhere? |
| 22:40:48 | cmarqu | prpplague: http://ipkgfind.handhelds.org/details.phtml?package=sqlite&official=&format= |
| 22:46:30 | prpplague | cmarqu: hmm wonder why i didn't find it earlier.... |
| 22:46:48 | cmarqu | Well, it's late, at least here :) |