09:51:00 | pb | morning |
09:52:33 | mallum | hey pb |
09:52:45 | mallum | http://handhelds.org/~mallum/matchbox/ |
09:58:24 | pb | mallum: wahey! |
09:58:55 | mallum | pb: would you grab the 0.3 tarball and just check it builds for you ? |
09:59:02 | pb | mallum: sure thing |
10:04:10 | pb | mallum: am I supposed to need to run autoconfigure.sh? |
10:06:14 | mallum | no |
10:06:19 | mallum | just configure |
10:06:43 | pb | configure: creating ./config.status |
10:06:43 | pb | config.status: creating Makefile |
10:06:43 | pb | config.status: error: cannot find input file: Makefile.in |
10:06:43 | pb | pb@chelsea:~/matchbox-0.3$ |
10:06:44 | pb | :-( |
10:07:07 | mallum | doh |
10:08:20 | mallum | eating lunch |
10:08:30 | mallum | will take a peek in a mo |
10:12:23 | mallum | arg, works fine for me :( |
10:13:24 | pb | :-| |
10:15:18 | mallum | you just do a ./configure and then get that ? |
10:16:20 | pb | yup |
10:16:29 | pb | well, ./configure --prefix=/usr, but I doubt the prefix matters |
10:16:43 | mallum | no shouldn't do |
10:16:56 | pb | let me try again, might have been a temporary NFS glitch or something |
10:16:59 | mallum | pb: let me try it on a woody machine .. |
10:18:07 | mallum | pb: I got xstroke working on powerpc last night :) |
10:18:17 | pb | cool :-) |
10:18:31 | pb | mallum: hmm, weird, it worked that time. |
10:18:45 | pb | mallum: guess NFS was to blame after all |
10:19:02 | mallum | pb: phew |
10:19:15 | mallum | pb: yes it worked ok for me on my x86 woody machnine |
10:24:05 | pb | mallum: all compiled, no problems. |
10:28:18 | mallum | pb: woohoo |
10:28:53 | mallum | pb: I was playing with X and ARM machine last night, it appears X may not support that card :( |
10:29:34 | pb | mallum: oh dear |
10:29:46 | pb | mallum: what version of X server are you using? |
10:31:52 | mallum | 4.1 |
10:31:59 | mallum | its an s3 trio64 card |
10:32:28 | pb | mallum: bizarre, S3 Trios must be the single most common type of card out there. |
10:32:54 | mallum | pb: yeah |
10:33:14 | pb | mallum: I suppose it's possible the driver didn't get ported to XF86 4, though. Can you install XF86_S3 from release 3.3? |
10:33:54 | mallum | pb: maybe the vesa driver supports them, but I dont think the vesa driver can speak through the CATS weirdo bios |
10:34:05 | pb | mallum: indeed |
10:34:20 | pb | mallum: I don't think the CATS provides any BIOS services after the kernel has booted |
10:34:24 | mallum | pb: 3.3 is not really much good to me as I need xrender stuff which is only in 4 |
10:34:46 | pb | mallum: ah |
10:34:47 | mallum | pb: the vesa driver pokes around for a bios and cant find one I think |
10:35:32 | mallum | pb: the idea behind getting X running on it would be to test GTK2 speeds, newer Xrender stuff before dumping on ipaq |
10:35:33 | pb | hmm, yes |
10:35:36 | pb | Support (accelerated) for the ViRGE, ViRGE/VX, ViRGE/DX, ViRGE/GX, ViRGE/GX2, ViRGE/MX, ViRGE/MX+, Trio3D and Trio3D/2X is provided by the "s3virge" driver. Support (accelerated) for the Savage3D, Savage3D/MV, Savage4, and Savage2000, is provided by the "savage" driver. Support for the other S3 chipsets has not yet been ported. |
10:35:47 | pb | that's from the 4.1 documentation |
10:36:17 | mallum | pb: yeaeh but I dont think trio3d is the same as trio 64 |
10:36:26 | pb | mallum: no, absolutely |
10:36:35 | pb | mallum: but it looks like it might be supported in 4.2: |
10:36:39 | pb | Support (accelerated) for the 964 (revisions 0 and 1), 968, Trio32, Trio64, Trio64, Trio64V+, Trio64UV+, Aurora64V+, Trio64V2, and PLATO/PX is provided by the "s3" driver (however, only models using the IBM RGB 524, Texas Instruments 3025, or an internal TrioDAC RAMDAC chip are supported). Support (accelerated) for the ViRGE, ViRGE/VX, ViRGE/DX, ViRGE/GX, ViRGE/GX2, ViRGE/MX, ViRGE/MX+, Trio3D and Trio3D/2X is provided by the "s3virge" driver. |
10:36:39 | pb | Support (accelerated) for the Savage3D, Savage3D/MV, Savage4, Savage2000, and SuperSavage, is provided by the "savage" driver. Support for the other S3 chipsets has not yet been ported. |
10:36:47 | mallum | pb: aha |
10:37:03 | mallum | pb: are the exp 4.2 debs built for ARM yet ? |
10:37:15 | pb | mallum: dunno |
10:38:10 | pb | getting lunch, back in 10 |
10:38:32 | mallum | pb: everything but arm :( -> http://people.debian.org/~branden/sid/ |
10:48:59 | pb | :-| |
10:49:06 | pb | mallum: wanna build 'em yourself? |
10:50:55 | Suppressor_g3 | hello all |
10:51:36 | mallum | pb: doing it |
10:51:43 | pb | good man |
10:51:52 | Suppressor_g3 | which the site of packages latest qtopia? |
10:52:03 | pb | #opie |
10:53:53 | pb | mallum: yesterday I was trying to figure out how to build TinyX packages for debian; it seems to be quite hard. |
10:54:35 | mallum | pb: the build rather than the packaging ? or the otherway round ? |
10:54:46 | pb | mallum: well, the combination. |
10:55:09 | pb | mallum: it doesn't seem to be possible to build both the "normal" XFree86 and the kdrive server in a single "make World" invocation |
10:55:28 | pb | mallum: so debian/rules would have to build once, change the .cf file, selectively "make clean" and then build again. |
10:55:41 | mallum | pb: I think you do a Buildserveronly in host.def and a buildkdrive too |
10:56:08 | pb | yeah |
10:56:22 | pb | I'll send email to Branden |
10:57:30 | mallum | then packaging wise I guess you'd have a xserver-kdrive |
10:57:46 | pb | yup |
10:57:53 | mallum | or maybe xserver-kdrive-fbdev or xserver-kdrive-vesa whatever |
10:58:07 | pb | well, probably xserver-h3600 in practice; it seems to be rather chummy with the touchscreen driver |
10:58:10 | mallum | pb: and that would just link /etc/X11/x to its binary |
10:58:44 | mallum | pb: yeah, they are meant to be converting it over to the rmk's tslib stuff |
10:59:18 | pb | mallum: yup, I think there is already some mechanism in place to select between different X servers, so you can go on using the 3.3 legacy servers if you need to, while still having the 4.x libs and apps. |
10:59:24 | mallum | pb: cos the ucb-whatever touchscreen does not work with tinyX as it supports new tslib /dev/ts format whilst tinyX the old |
10:59:25 | pb | mallum: dunno, you'd have to ask jg I guess. |
10:59:33 | pb | hmm, right |
10:59:41 | pb | it would be kind of cool to be able to run X on my assabet board |
10:59:58 | mallum | pb: is assabet ucbXXXX based ? |
11:00:05 | pb | mallum: think so |
11:00:05 | mallum | pb: like touchscreen ? |
11:00:15 | mallum | pb: tiny is pretty easy to patch to do it |
11:00:25 | pb | ah, ok |
11:00:49 | mallum | pb: but fontenot said that tslib was just too slow doing the filtering etc in userspace |
11:00:56 | pb | :-| |
11:01:09 | mallum | pb: Id like to patch tinyX to do click and hold for left click ;) |
11:01:17 | pb | seems strange, I can't see that it would be all that much slower in userspace than in the kernel. |
11:01:31 | pb | for right click, you mean? |
11:01:36 | mallum | pb: I think they needed to use sigio or something |
11:01:38 | mallum | pb: yeah |
11:01:42 | mallum | pb: sorry |
11:01:50 | mallum | pb: that what jg mentioned |
11:01:59 | pb | bizarre, guess I ought to look at the tslib code and see what's going on. |
11:02:17 | pb | I'd have thought they could just use blocking I/O, it seems like a fundamentally single threaded process. |
11:02:32 | pb | mallum: you don't buy into the cworth "fix it in the toolkit" school of thought then? |
11:03:07 | mallum | pb: hmm, sure. But I also believe in practicality ;) |
11:03:20 | pb | heh |
11:03:41 | mallum | pb: Im sure originally jg+gang it was like 'hey look run desktop apps on your handheld' |
11:03:49 | pb | well, it'd be interesting to look at what it would take to do that in Gdk/Gtk. |
11:03:59 | pb | mallum: mmm. |
11:04:16 | mallum | pb: the only solid way to do it is in the Xserver |
11:04:46 | pb | you mean solid as in reliable, or solid as in coping with all those bare-metal Xlib programs? |
11:05:02 | mallum | pb: and as far as Im concerned its no different to big X emulating 3 buttons |
11:05:11 | pb | mallum: agreed |
11:05:47 | mallum | pb: solid as in reliable. I did it once in gdk and it difficult to balance up events ( up and downs ) which then confuses X |
11:05:55 | pb | mallum: I see |
11:05:59 | mallum | pb: and also it approx 5 line patch to tinyX |
11:06:32 | mallum | pb: also I have a touchscreen machine which has a modifier button for the 2nd mounse button, tinyX has no way to deal with that |
11:07:20 | pb | mallum: right |
11:09:34 | mallum | pb: also theres no reason why you couldn't do it in tslib ;) |
11:09:49 | pb | mallum: mmm, that's a thought. |
11:10:07 | mallum | pb: as long as the tslib spec has an allowance for a second button |
11:10:08 | pb | mallum: well, I suppose you still have the problem of the modifier button on your hardware, unless that gets exposed as a key event or something. |
11:11:05 | mallum | pb: X people couldn't complain if it was done in tslib |
11:12:15 | pb | mallum: indeed, that's a big plus. |
11:12:30 | pb | mallum: just downloaded the tslib source, looking at it now. |
11:16:36 | mallum | pb: if I put http://people.debian.org/~branden/sid/source/ in my sources.list as a src, do I just do an apt-get -b source XFree86_4.2.0 ? |
11:16:44 | pb | mallum: yes |
11:16:55 | mallum | pb: okey |
11:17:23 | pb | well, I can't see anything obvious in tslib that ought to slow it down too much |
11:17:52 | pb | and it doesn't seem to have any concept of multiple buttons at all: the event structure it gives you looks like: |
11:17:56 | pb | struct ts_sample { |
11:17:56 | pb | int x; |
11:17:57 | pb | int y; |
11:17:57 | pb | unsigned int pressure; |
11:17:57 | pb | struct timeval tv; |
11:17:59 | pb | }; |
11:18:03 | pb | :-| |
11:19:27 | mallum | :( |
11:19:53 | mallum | need to find more touchscreens with modifier buttons |
11:20:07 | mallum | I wonder if one exists with the modifier button on the pen ? |
11:21:05 | pb | never seen one like that, but it doesn't seem unlikely |
11:22:40 | mallum | pb: do you knwo rmk personally ? |
11:23:43 | pb | mallum: depends what you mean |
11:23:45 | pb | mallum: I've met him a few times |
11:23:56 | mallum | realises not just the build of X is gonna take some time - only 11k p/sec download |
11:24:21 | mallum | pb: maybe you could mention it to him. I d be too scared to post and suggest it ;) |
11:24:44 | pb | mallum: probably better coming from you to be honest, rmk tends to have a very low opinion of things I suggest |
11:25:52 | mallum | pb: :( |
11:26:28 | mallum | pb: he can get very aggressive on the arm list |
11:26:36 | mallum | pb: sometimes it is funny |
11:26:40 | pb | indeed |
11:26:47 | mallum | pb: he does do alot of work on that list though ... |
11:27:46 | pb | yeah |
11:28:08 | pb | and he mostly seems to have given up styling himself "THE developer", which I guess is a step in the right direction :-) |
11:28:23 | mallum | ? |
11:28:54 | pb | see his sig: e.g. http://lists.arm.linux.org.uk/pipermail/linux-arm-announce/2001-March/000043.html |
11:33:07 | mallum | ooo-err |
11:34:10 | mallum | and THE is capitalised |
11:34:21 | pb | yup, that's the best part :-) |
11:36:24 | mallum | was he making a statement, or just being a twat ? |
11:36:31 | pb | I've never been able to figure it out. |
11:36:42 | mallum | I think he likes ac alot |
11:36:48 | pb | yeah |
11:38:07 | mallum | he is having an affair with tesla ;P |
11:38:15 | pb | heh |
11:39:45 | mallum | I find all there diarys very entertaining I must say |
11:40:33 | pb | I haven't read any of them recently, maybe I should have a look |
11:41:22 | mallum | I imagine alan cox being quite nice in person, as long as you dont mention doing any audio filtering in kernel space |
11:41:44 | pb | yeah, I should think so |
11:41:51 | pb | I've never actually spoken to him in person |
11:44:42 | pb | oh dear |
11:44:46 | pb | Jul 12 12:53:11 localhost kernel: Unable to handle kernel paging request at virtual address 00001000 |
11:54:02 | mallum | ouch |
11:54:23 | pb | trying to get my wireless LAN card working |
11:54:23 | mallum | pb: why do people sometimes put a pad 'element' in a struct ? |
11:54:34 | mallum | pb: what card is it ? |
11:54:44 | pb | mallum: either space for future expansion, or aligning to a known boundary |
11:54:47 | pb | mallum: some kind of prism2 |
11:55:05 | pb | mallum: it calls itself "ACTIONTEC", "PRISM Wireless LAN PC Card", "0381", "RevA" |
11:55:42 | mallum | pb: is it lying about its voltage ? |
11:56:06 | pb | mallum: hard to tell, the single sleeve apparently doesn't report voltage information |
11:56:29 | mallum | pb: the dwl650 have a lot of problems due to broken firmware |
11:56:42 | mallum | pb: I could never get the prism2ap drivers to work with that |
11:56:50 | pb | mallum: but it's a 3V card, and I'd expect it would have exploded with smoke and fire by now if it was going to go wrong. |
11:57:07 | mallum | pb: they used to with newer orinoco driver, if you passed some vcc option in |
11:57:17 | mallum | pb: but the card was pure evil |
11:59:09 | pb | ah, hmm, yeah, the orinoco driver seems to be working a bit better |
11:59:16 | pb | Jul 12 12:11:58 localhost kernel: eth0: Error -110 setting multicast list. |
11:59:21 | pb | I guess that's harmless enough |
11:59:40 | mallum | pb: yeah |
12:00:14 | mallum | pb: I got a cheap sitecom? prism2 card from PC world for 50 quid and that worked strait off as an AP no problems at all |
12:00:47 | pb | yeah, I have a Netgear prism2 in my other machine and that "just worked" with wlan-ng. |
12:01:47 | pb | dunno if it's an ARM specific failure, or something about the actiontec card; I can't put the actiontec into the i386 machine because the latter is 5V only, but I guess I could try the netgear in the ipaq. |
12:24:00 | mallum | Err http://people.debian.org sid/source/ xfree86 4.2.0-0pre1v1 (tar) 3 |
12:24:00 | mallum | Error reading from server Remote end closed connection |
12:24:03 | mallum | :( |
12:24:07 | mallum | its started again |
12:24:40 | pb | sigh |
12:25:48 | mallum | will that be mirrored anywhere ? |
12:26:03 | pb | don't think so |
12:26:14 | mallum | oh no wait it picked up where it left off :) |
12:26:26 | mallum | and 14k per sec now - wooohooo ! |
12:28:37 | mallum | pb: is gpe what working now ? |
12:28:49 | pb | mallum: think so, want to try it out? |
12:30:13 | mallum | pb: yeah will do |
12:30:27 | mallum | just build some matchbox 0.3 ipkgs :) |
12:31:10 | mallum | Im gonna leave xft on by defualt for these |
12:31:17 | mallum | and make some nonxft alternates |
12:31:50 | pb | cool |
12:31:52 | mallum | pb: hmm ipkg does not do suggests or recommends ? |
12:32:03 | pb | don't think so |
12:32:29 | pb | nor does dpkg, as such -- it's just dselect etc that looks at those. |
12:32:39 | mallum | ok |
12:45:12 | mallum | pb: uploading 0.3 ipkg |
12:45:16 | mallum | pb: seems fine |
12:45:22 | mallum | pb: couple of minor theme issues |
12:45:34 | pb | cool |
12:45:40 | mallum | pb: but other than that seems cool |
12:46:09 | mallum | pb: Im gonna write some release notes then get it in unstable |
12:47:08 | pb | great |
12:48:18 | mallum | pb: wehay gpe-what docked :) |
12:48:26 | mallum | pb: now 'what' does it do ? |
12:49:03 | pb | mallum: open up, say, gpe-calendar or gpe-todo |
12:49:10 | mallum | pb: I see :) |
12:49:31 | mallum | pb: I click it then a widget and it tells me stuff |
12:49:35 | pb | yup |
12:49:42 | mallum | pb: how does it work ? |
12:51:54 | pb | mallum: tooltips, basically |
12:53:00 | pb | mallum: when you click on the gpe-what button in the dock, it sets a magic property on the root window; all the other apps notice this and put themselves into "help mode". |
12:54:01 | pb | mallum: the application-side stuff is all implemented in libgpewidget, it's completely transparent to the programmer. |
12:55:35 | mallum | pb: cool |
12:55:56 | mallum | pb: you could maybe use the tray_message() call to send a large amount of help too ? |
12:56:08 | mallum | pb: that messaging is part of the spec BTW |
12:56:58 | pb | mallum: at the moment, all the display is done by the client application, gpe-what itself is literally just a button that toggles a property. |
12:57:16 | pb | mallum: I think the Gtk tooltips architecture does provide for longer help messages, but none of the apps set them at the moment. |
12:57:31 | mallum | pb: sure |
12:57:53 | pb | I'll have a look at tray_message. |
12:58:33 | mallum | pb: Im thinking about extending a little ... |
12:59:37 | mallum | pb: oh if gpe-what sticks a file in /usr/share/matchbox/mbdock - it'll get included in the menu |
12:59:45 | pb | ah, cool |
12:59:56 | mallum | pb: the file format is simple |
13:00:54 | mallum | pb: though currently I dont dynamically reload it :( |
13:01:05 | mallum | pb: Id better add that to the 3.1 todo |
13:05:31 | pb | good plan |
13:06:49 | mallum | X has started building .... |
13:06:56 | pb | yay |
13:07:03 | mallum | how many hours ? |
13:07:04 | mallum | 12 |
13:07:07 | pb | about that I think |
13:07:34 | mallum | ouch |
13:07:51 | pb | you could check buildd.debian.org, the logs have total build time at the end. |
13:17:03 | mallum | If I have too packages matchbox_0.3 and matchbox_0.3noxft - ipkg is gonna think the noxft one is newer isn't it ? |
13:17:24 | pb | yes |
13:17:31 | pb | you need to put the "noxft" before the _ |
13:17:40 | pb | ie, have matchbox_0.3 and matchbox-noxft_0.3 |
13:17:43 | mallum | yeah |
13:17:47 | mallum | thats what I thought |
13:17:56 | mallum | should I change the name in control too ? |
13:17:59 | pb | yes |
13:18:31 | pb | with big debian, matchbox-xft would say "Conflicts: matchbox; Provides: matchbox" or something; dunno if there is a direct equivalent with ipkg |
13:19:48 | mallum | oh dear; |
13:19:51 | mallum | mallum@debian:~/Projects/X4.2/xfree86-4.2.0$ df -h |
13:19:51 | mallum | Filesystem Size Used Avail Use% Mounted on |
16:03:07 | mallum | pb__: you there ? |
16:04:38 | pb__ | mallum: kind of |
16:05:09 | mallum | pb__: I sent an announce to familair lists and it dont seen to have shown up :( |
16:05:18 | pb__ | I think the lists are a bit slow at the moment |
16:05:32 | pb__ | sometimes taking up to 12 hours or so |
16:06:15 | mallum | hmm, okey |
16:06:33 | mallum | I hope too many people dont upgrade before the mail ;( |
16:07:57 | mallum | brb... |
19:25:26 | lars_g | Greetings everyone. |
19:25:30 | lars_g | :) |
19:26:59 | pb__ | yo |
19:29:39 | lars_g | pb__: Yo :) |
19:30:13 | lars_g | it's quite lonely here today |
19:45:01 | mallum | hey |
19:46:27 | pb__ | hey mallum |
19:48:32 | mallum | hey pb__, X still a building ... |
19:48:41 | pb__ | :-| |
19:48:47 | pb__ | is building glibc |
19:49:27 | mallum | what for ? |
19:49:36 | mallum | as in what device ... |
19:49:39 | pb__ | arm |
19:49:44 | pb__ | CATS, more specifically |
19:49:58 | pb__ | mostly just to find out whether libc 2.3 actually works on arm at the moment |
19:50:21 | mallum | aha |
19:50:24 | pb__ | and I installed your matchbox packages, once I figured out that ipkg considers "0.3rc1" to be newer than "0.3" (duh) |
19:50:38 | mallum | and you are in charge of that port ? |
19:50:43 | pb__ | yes |
19:50:49 | pb__ | I guess so, anyway |
19:52:01 | mallum | ;) |
19:52:04 | pb__ | is also building new dillo packages for debian |
19:52:09 | mallum | Any probs with matchbox ? |
19:52:17 | pb__ | mallum: no, seems to be working pretty well so far |
19:52:50 | pb__ | I did notice that if you rotate to landscape mode, put the dock slider near the right hand edge, and then rotate back to portrait, the slider is still off the side of the screen. But I guess you could file this under "don't do that, then". |
19:53:15 | mallum | The slider is unaccessable ? |
19:53:23 | mallum | The slider will disapear if its not needed |
19:53:58 | pb__ | mallum: yeah |
19:54:25 | mallum | yeah thats whats probably happening |
19:54:36 | mallum | I think I need to rethink how the slider should work |
19:54:41 | mallum | I have some other ideas |
19:55:01 | pb__ | mallum: I've also been noticing that miniapm seems to just "go away" occasionally, not sure exactly what's going on there. |
19:56:04 | pb__ | mallum: and, right at the moment, I seem to have a mystery "<unnamed>" entry in my dock that I can't get rid of. |
19:56:41 | mallum | pb__: old xstroke ? |
19:57:01 | pb__ | mallum: don't think so, I'm running your ma3 or whatever it was. |
19:57:09 | mallum | pb__: some timee the dock apps get clipped by the oom killer thingy |
19:57:12 | pb__ | mallum: and xstroke itself is definitely visible separately in the dock. |
19:57:21 | mallum | pb__: old mmenu ? |
19:57:32 | mallum | pb__: I can remember seeing it ... |
19:57:43 | pb__ | ah, might be an old mmenu |
19:57:44 | mallum | pb__: looking at ~/.mbdock may give a clue |
19:58:10 | pb__ | I seem to have crashed my ipaq by swapping sleeves, gonna have to reboot |
19:58:31 | pb__ | sigh |
20:00:26 | pb__ | oh well, the mystery thing has gone away now, whatever it was. |
20:02:45 | mallum | I really need to set up a mini bugzilla |
20:02:58 | mallum | or something to track bugs, feature requests etc |
20:03:04 | pb__ | mallum: you can create components in the gpe bugzilla if you want |
20:03:19 | pb__ | mallum: or indeed a new top-level project |
20:03:46 | mallum | cool, thanks |
20:03:57 | mallum | bugzilla always seem overkill to me :( |
20:04:02 | pb__ | yeah |
20:04:07 | pb__ | but it seems to work pretty well |
20:04:32 | mallum | I tried to set up this 'roundup' but it dont work :( |
20:04:46 | pb__ | let me know when you've set up a bugzilla account, and I'll turn on your admin privs. |
20:04:56 | mallum | hh.org bugzilla keeps mailing me bugs I've closed |
20:05:05 | mallum | cool, thanks again |
20:05:11 | mallum | I may have to take you up on your offer |
20:05:21 | mallum | will sort it next week now though |
20:06:03 | pb__ | sure |
20:06:18 | pb__ | ../sysdeps/arm/dl-machine.h:428: `_dl_rtld_map' undeclared (first use in this function) |
20:06:20 | pb__ | :-( |
20:10:39 | mallum | doh |
20:11:45 | pb__ | hmm, this is weird, I do "apt-get update" and: |
20:11:49 | pb__ | Err http://www.uk.debian.org unstable/main Packages |
20:11:49 | pb__ | 416 Requested Range Not Satisfiable |
20:12:05 | pb__ | well, I suppose nothing should surprise me about www.uk.debian.org, but still. |
20:12:08 | mallum | oo a collectable |
20:12:30 | pb__ | now using http.us.debian.org, sigh |
20:12:44 | pb__ | groan |
20:12:48 | pb__ | gcc: Internal compiler error: program cc1 got fatal signal 11 |
20:12:48 | pb__ | make[2]: *** [/var/tmp/build/libc-2.3/elf/rtld.os] Error 1 |
20:14:50 | pb__ | fed up with enlightenment, gonna try metacity on my desktop for a bit :-) |
21:28:15 | mallum | metacity rocks |
21:28:28 | mallum | lots of nice themes at sunshineinabag.co.uk |
21:28:41 | mallum | matchbox dock runs fine it it too |
21:28:56 | mallum | or just run gnome-panel |
21:29:01 | mallum | or full blown gnome |
21:30:29 | pb__ | cool |
21:39:45 | pb_ | hmm |
21:39:49 | pb_ | having some kind of X session grief here |
21:40:54 | mallum | pb_: what with ? |
21:41:28 | pb_ | I'm not sure exactly. Metacity keeps crashing, for one thing. |
21:41:50 | pb_ | ... and GNOME doesn't seem too keen on remembering my choice of WM, so I keep getting enlightenment back every time I log in. |
21:42:14 | pb_ | I just blew away my ~/.gnome directory and restarted X, in hopes that this would fix things. Fingers crossed, no problems so far. |
21:42:25 | mallum | is x-window-manager set to metacity ? |
21:42:36 | mallum | I think you can export WINDOWMANAGER=metacity too |
21:42:37 | pb_ | mallum: well, that's not really how it works in big debian :-) |
21:42:59 | pb_ | it's supposed to be configurable per user, so x-window-manager isn't all that appropriate. |
21:43:34 | mallum | mallum@ti:~/src$ ls -la /usr/bin/x-window-manager |
21:43:34 | mallum | lrwxrwxrwx 1 root root 34 Apr 2 17:48 /usr/bin/x-window-manager -> /etc/alternatives/x-window-manager |
21:43:48 | mallum | I think the X startup scripts use it as a fallback |
21:43:51 | pb_ | right, that gives you the default for the "Debian" session |
21:44:34 | mallum | I dont run a display manager |
21:44:46 | pb_ | oh right, I use gdm |
21:44:49 | mallum | so I just have metacity in the end of my ~/.xinitrc |
21:46:00 | pb_ | I wonder if I might be in some kind of GNOME 1.4 vs GNOME 2 hell. |
21:46:14 | mallum | but I set sids machine up with gdm and metacity ( sid ) |
21:46:27 | mallum | she has gnome1.4 gdm |
21:46:33 | mallum | and gnome2 panel |
21:46:35 | mallum | it works ok |
21:46:38 | pb_ | same as me, i think |
21:46:44 | mallum | one sec I'll check what I did on that |
21:47:01 | pb_ | well, maybe it's working ok now here. |
21:48:48 | mallum | I have ; |
21:49:06 | mallum | deb http://ftp.uk.debian.org/debian/ ../project/experimental main |
21:49:14 | mallum | too for more gnome2 stuff |
21:49:25 | pb_ | maybe I should try adding that too |
21:53:13 | mallum | hmm I cant actually tell how I set sids machine up :( |
21:53:21 | pb_ | oh well |
21:53:24 | mallum | the session entry is just set to 'last session' |
21:53:46 | mallum | but I've put stuff in her .xsession file ( but not the wm ) |
22:01:51 | mallum | pb_: I got a mail from nils saying that matchbox isn't letting apps using shm extension run |
22:02:23 | pb_ | mallum: hmm |
22:02:38 | pb_ | mallum: what does he say happens? |
22:05:31 | mallum | he gets a bad access error |
22:05:36 | pb_ | hmph |
22:05:44 | pb_ | seems weird |
22:05:44 | mallum | from the app |
22:06:00 | pb_ | I wouldn't have thought the wm would get involved there at all. |
22:06:38 | mallum | although he said it worked with ion, searching google for it, seems like it could be related to alignment memory stuff and only happen some times |
22:07:05 | pb_ | does he give an example of an app that has the problem? |
22:07:16 | mallum | mplayer |
22:07:49 | mallum | I cant get mplayer to run in xnest though, so its a pain to debug on the ipaq |
22:08:25 | mallum | I need to set an Xerror handler in the mplayer source, recompile it, set a breakpoint to that with gdb .... |
22:08:53 | pb_ | crumbs, I just tried to run xawtv in Xnest and all hell broke loose |
22:10:21 | mallum | Xnest is like an xserver from 1985 |
22:10:59 | pb_ | yeah, and xawtv seems to be some kind of poster child of bad engineering |
22:11:51 | pb_ | it informed me that my X server "does not support DGA", which is fair enough I suppose, but then proceeded to start capturing into an apparently-random area of the screen anyway. |
22:12:44 | mallum | I just mplayer with -vo sdl and it Xnest crashed |
22:12:49 | pb_ | sight |
22:12:51 | pb_ | sigh even |
22:13:25 | mallum | I'll see what Nils comes back with ... he sounded like he wanted to figure out what it was |
22:19:00 | mallum | Hmm, I've got mplayer working via -vo x11, but Im getting a badacess error, matchbox or no matchbox |
22:43:27 | pb_ | hmm |
22:43:40 | pb_ | well, I conceded defeat and set x-window-manager to metacity :-} |
22:43:51 | mallum | :( |
22:44:09 | mallum | I think you can do an export WINDOWMANAGER too |
22:44:15 | pb_ | worse, metacity now crashes immediately when I log in |
22:44:27 | mallum | arg |
22:44:34 | mallum | does metacity run in xnest ? |
22:44:41 | pb_ | though if I open an xterm and run it manually from in there, it seems to be fine after that |
22:44:58 | pb_ | -- so, I guess something about the session startup is spooking it out. |
22:45:20 | mallum | pb_: it does attempt to connect to a session manager I believe |
22:45:39 | mallum | this may be useful; |
22:45:43 | mallum | http://cvs.gnome.org/lxr/source/metacity/README |
22:46:58 | pb_ | hmm, just thought to check my .xsession-errors |
22:47:15 | pb_ | but sadly it doesn't seem to be getting updated by the gnome stuff. |
22:48:43 | mallum | :( |
22:51:07 | pb_ | aha, .gnome-errors! |
22:51:23 | pb_ | how embarrassing, it seems I didn't fix x-window-manager quite as thoroughly as I thought I had... |
22:51:29 | pb_ | SESSION_MANAGER=local/kc:/tmp/.ICE-unix/336 |
22:51:30 | pb_ | /usr/bin/gnome-wm: exec: x-window-manager: not found |
22:51:30 | pb_ | Gnome-Message: gnome_execute_async_with_env_fds: returning -1 |
22:51:30 | pb_ | Gnome-Message: gnome_execute_async_with_env_fds: returning -1 |
22:51:31 | pb_ | :-/ |
22:52:44 | mallum | eek |