Conversation with #inferno at Wed Apr 22 08:41:17 2009 on powerman-asdf@irc.freenode.net (irc) (09:22:32) underspecified_ [n=eric@naist-wavenet126-150.naist.jp] entered the room. (09:33:36) underspecified left the room (quit: Read error: 110 (Connection timed out)). (10:02:39) npe [n=npe@195.207.5.2] entered the room. (10:02:47) npe left the room (quit: Remote closed the connection). (10:03:05) npe [n=npe@195.207.5.2] entered the room. (10:03:08) andguent [n=andre@p4FF67C5C.dip.t-dialin.net] entered the room. (10:37:21) mjl-: morning (10:44:46) underspecified_ left the room (quit: Read error: 60 (Operation timed out)). (10:50:41) npe_ [n=npe@195.207.5.2] entered the room. (10:53:42) npe_: mjl-: morning. (11:06:26) npe left the room (quit: Read error: 113 (No route to host)). (11:10:39) gdiaz_ [n=gdiaz@dos.9grid.es] entered the room. (11:19:56) npe_ left the room (quit: Remote closed the connection). (11:20:03) npe [n=npe@195.207.5.2] entered the room. (11:31:13) gualteri [n=unknown@crespins.disca.upv.es] entered the room. (12:00:39) underspecified [n=eric@walnut.naist.jp] entered the room. (12:28:34) underspecified left the room (quit: ). (12:53:11) hotaru2k3_ [n=hotaru@cpe-66-61-20-230.neo.res.rr.com] entered the room. (12:53:33) hotaru2k3 left the room (quit: Read error: 113 (No route to host)). (12:56:09) hotaru2k3_ is now known as hotaru2k3 (13:18:28) gdiaz_: hum i just discovered the msi package has private fonts on it. . . i'll remove that download and will generate a proper msi without that fonts. . . (13:21:16) soul9: grmbl i am getting those stupid xcb errors (13:21:39) gdiaz_: what are those errors? (13:28:56) soul9: er.. (13:29:01) soul9: segfaults and the likes (13:29:08) soul9: you want me to paste an example? (13:29:19) soul9: but i'm installing latest svn to see if that was fixed.. (13:32:15) gdiaz_: there are binary packages in inferno-npe.googlecode.com (13:32:32) gdiaz_: i'll upload a new version of the windows package without the private fonts (13:50:20) soul9: okay, looks like it was fixed (13:50:22) soul9: w0000h0000 (13:50:30) soul9: also, it now builds with gcc4.x (13:50:33) soul9: or so it seems (13:50:35) soul9: powerman-asdf, i made an ebuild for inferno-subversion, http://dev.9souldier.org/?p=overlay.git;a=blob;f=dev-inferno/inferno/inferno-99999999.ebuild;h=0cedfc8d0aa99008fa05f8a8ed6299ded9673d5a;hb=HEAD (shamelessly ripped off your ebuild ;) (13:54:27) soul9: now to figure out how to add the free fonts from acme-sac (13:54:38) soul9: and then i can finally use again ircfs (13:54:40) soul9: w00t!!!! (13:54:54) soul9: all irc clients suck, and bncs suck even more! (13:56:41) soul9: huuuum (13:56:44) soul9: not totally gone away (13:56:47) soul9: gdiaz_, around? (13:57:06) soul9: http://pastebin.com/m154fa98e (13:58:39) soul9: do you think this may be because of the fonts missing (this happened when i tried starting acme for the 4th time) (14:02:09) mjl-: it does complain about fonts missing. but it shouldn't crash... (14:02:31) mjl-: but it's an octopus binary? perhaps that has different bugs from normal inferno? (14:02:49) mjl-: soul9: do you want the inferno non-free fonts? or only the acme-sac fonts (14:09:37) gdiaz_: hello (14:10:14) gdiaz_: soul9 you tried any of those packages? (14:18:06) KillerX [n=anant@gentoo/developer/KillerX] entered the room. (14:24:53) soul9: gdiaz_, which packages? (14:25:06) soul9: mjl-, no, it's just the sript i used to use for octipus (14:25:30) soul9: but since i can't figure out how to build octopus with latest inferno sources, i have stopped using it (14:25:47) soul9: since my machine now fully uses gcc-4.x and xcb (14:26:03) gualteri: maybe adding the non free fonts (pelm/unicode.8.font & lucidasans/euro.8.font) will help: http://code.google.com/p/acme-sac/source/browse/trunk/dis/setupfonts?r=102 (14:27:19) soul9: i'll try that then (14:27:47) soul9: heh (14:27:49) soul9: crash.. (15:12:54) underspecified [n=eric@softbank220043052007.bbtec.net] entered the room. (15:39:44) soul9: http://pastebin.com/m70bf76f6 (15:48:40) hotaru2k3 left the room (quit: "Leaving"). (15:49:27) gualteri: emu is still crashing in that paste, besides other things. have you tried getting a binary from http://code.google.com/p/inferno-bin/downloads/list?, a bit outdated but should work (16:08:38) soul9: i'll try (16:08:55) soul9: but my system is fairly cutting-edge, inferno usually doesn't like that very much... (16:09:39) soul9: heh (16:09:55) soul9: i love emu's non-deterministic panics (16:10:00) soul9: it completed the script now (16:14:09) soul9: well, that emu just hang on my system doing nothing (16:14:36) mjl-: which linux do you have? (16:14:50) soul9: i use gentoo (16:14:57) mjl-: ahhh (16:15:00) soul9: and this is the unstable version of gentoo (16:15:06) soul9: so it's really cutting-edge (16:15:11) mjl-: fun ☺ (16:15:22) soul9: maybe i could try a chroot of a stable debian or something (16:15:36) soul9: that could just bitrot withot really affecting inferno i guess.. (16:15:42) soul9: without* (16:15:50) mjl-: myeah, i ran emu on the latest debian stable. seemed to work. and it's relatively recent. (16:16:02) soul9: (for now ;) (16:16:12) soul9: we'll see in 2 years :D (16:18:46) gualteri: (could use gcc-config to downgrade to a less cutting-edge gcc, and compile emu with that) (16:19:39) mjl-: soul9: do you have jit on? (16:19:41) mjl-: cat /dev/jit (16:20:06) soul9: er (16:20:18) soul9: ; cat /dev/jit (16:20:18) soul9: 0; (16:20:21) soul9: mjl-, ^ (16:20:23) soul9: should i? (16:20:28) mjl-: ok. well, doesn't matter much anyway (16:20:56) soul9: gualteri, well, that doesn't work, you have to do a gcc-counfig && source /etc/profile each time you start inferno iirc (16:21:03) soul9: and on ~x86 even that doesn't work (16:23:56) mjl-: perhaps you can turn off some gcc optimizations? (16:24:05) mjl-: and see if it still breaks? (16:24:13) soul9: hum (16:24:20) soul9: i really don't use optimisations (16:24:40) mjl-: doesn't gentoo turn on all kinds of optimizations by default? (16:24:47) soul9: -fomit-frame-pointer is the only one i have. (16:24:49) soul9: no it doesn't. (16:25:02) soul9: but you can turn stupid optimizations on very easily though (16:25:07) mjl-: ok (16:25:42) mjl-: it seems a freed block is used. i assume that's a dis vm block. would be useful to have a minimal program that causes the crash. (16:25:43) soul9: gualteri, you think this is a gcc-4 problem? (16:26:04) soul9: mjl-, usually mounting and binding makes it crash very fast (16:26:43) mjl-: so try to cut it down to one program to ruN? (16:27:42) soul9: ok, i'll try (16:29:27) gdiaz_: soul9 have you tried an already compiled emu? (16:30:14) soul9: gdiaz_, i just tried on from the url gualteri gave me (16:30:25) soul9: but it just hangs, doesn't even give me a prompt (16:30:39) soul9: it's waaaaaay too old, though, fcrom 2007 (16:30:43) soul9: from 2007 (16:31:04) gdiaz_: have you tried those at inferno-npe.googlecode.com download section? (16:31:20) soul9: oh, no, i didn't know about that (16:31:22) soul9: lemme see (16:31:25) gdiaz_: those are svn binaries release 407 (16:31:45) gdiaz_: let me know if you find any problem with them, so i can fix them and upload newer versions (16:31:59) npe: gdiaz_: ha... we need to rename that project. (16:32:00) soul9: okay, i'll try it (16:32:10) npe: inferno-os-pub or something. (16:32:19) soul9: but i don't know why it would work better than the current svn (16:32:41) gdiaz_: the tgz is an slackware package but if you decompress it on / you will get /usr/lib/inferno dir with inferno inside (16:32:52) mjl-: soul9: i guess the gentoo gcc isn't trusted :P (16:33:13) gdiaz_: npe hahaha, i don't really care, but probably it would be better to have a generic name so people doesn't get confused :) (16:33:15) soul9: mjl-, lol (16:35:15) soul9: well, i'll try it (16:35:44) gdiaz_: i'm going to upload a new msi package without the private fonts (release 408 i just compiled) (16:35:53) mjl-: soul9: you can open a debugger on the emu, and see where the panic occurs (16:35:53) gdiaz_: for those using windows machines (16:36:13) mjl-: the "from X/Y" in the panic are pc's. you should be able to get a file+line number from those (16:36:29) mjl-: gdiaz_: how about adding a little script that automatically fetches the fonts? (16:36:35) mjl-: and installs them in /fonts/ (16:36:49) mjl-: so you just have to run "fetchfonts" after the install (16:36:58) mjl-: then you're not distributing them, but they'll be trivial to get (16:37:19) gualteri: mjl-, that's what setupfonts does (see backlog) (16:37:28) gualteri: just to credit the author :) (16:37:49) soul9: mjl-, you mean using gdb i could get something usable? (16:37:51) mjl-: ah, nice (16:37:59) mjl-: soul9: assuming gentoo has a working gdb ;) (16:38:09) soul9: :f (16:38:12) mjl-: but yes (16:38:17) mjl-: i don't know how gdb works (16:38:29) ***soul9 isn't fond of gdb (16:38:32) mjl-: in acid, you can do: acid /path/to/emu; and inside acid just do src($pc) (16:38:35) mjl-: with pc replaced (16:38:35) soul9: yeah, me either, really (16:38:39) mjl-: and you get what you want (16:38:46) mjl-: gdb probably does something like that, with different syntax (16:38:50) soul9: hum (16:38:58) soul9: i have acid from p9p :) (16:39:55) soul9: well, right now i'm at work, so.. (16:39:56) gdiaz_: mjl- we can write the fonts thing on the wiki at least for now, i think we can convince charles to put some more fonts in inferno and default all the programs to use those fonts. . . or may be we can convince the lucida fonts authors to release them with gpl :? (16:40:17) soul9: isn't there some fonts hardcoded in acme? (16:40:35) soul9: because even when i use acme -f otherfont it complains about not finding the fonts (16:40:38) gdiaz_: you can use acme -f $font, but yes, the default ones are in the source (16:41:11) gdiaz_: soul9 that's because some characters of the distributed fonts relays on the lucida ones which are not included (16:41:18) mjl-: to replace fonts in programs, you can still bind the ones you want on the placea ll programs use (16:41:26) soul9: ahh (16:41:52) soul9: okay, so say if i dl the free fonts from acme-sac, i need to bind those to /fonts/lucidasans? (16:41:55) mjl-: i haven't had problems with fonts. the ones from vitanuova work fine. (16:42:08) gdiaz_: i use those too without problems (16:42:20) mjl-: soul9: perhaps you need to bind a file inside /fonts/lucidasans/ too (16:42:27) mjl-: like unicode.$size or what is it (16:42:37) soul9: i see (16:42:38) soul9: :-/ (16:42:52) soul9: oh well, simplest is installing lucidasans (16:42:56) mjl-: jus (16:42:58) mjl-: jups* (16:43:12) soul9: well, this evening i'll come and maybe bug you guys some more, i can't take xchat anymore :( (16:43:18) gualteri: soul9, is uname -r >= 2.6.27-gentoo-r8?, that's why the binaries in inferno-bin won't work as in 2.6.27 clone(2) was modified, that's the reason i had to build a new emu here (16:43:28) soul9: ah (16:43:36) soul9: yes, i use a vanilla .29 kernel iirc (16:43:42) soul9: no, .28 (16:43:56) soul9: not gentoo-sources, even i don't trust gentoo-sources (16:44:00) soul9: :) (16:44:17) gualteri: that's why old emu binaries will hang, (16:44:26) soul9: ok, i see (16:44:41) gdiaz_: i have here a 2.6.21.5-smp. . . so not sure if it will work with yours (16:45:07) soul9: well, the one from the npe-inferno looks like it works great (16:45:34) soul9: i don't have much time to test right now though, i'm at work :) (16:45:36) gualteri: then you've got a good workaround the bug (16:45:39) gdiaz_: soul9 nice to hear, try also the ios script in $INFERNO_ROOT/Linux/386/bin/ios like ios wm/wm :) (16:45:51) soul9: gdiaz_, what does it do? (16:46:19) soul9: hmmm (16:46:25) soul9: i don't have that file (16:46:34) soul9: bash: Linux/386/bin/ios: No such file or directory (16:46:38) mjl-: soul9: you can get a quick approx on the function where the error occurs by matching the addresses with the output of nm(1) on the binary (16:47:40) gdiaz_: soul9 hum strange, i though i had it on the package. . . (16:48:34) soul9: :-/ (16:49:01) gdiaz_: anyway you can get it from the wiki pages in the inferno-npw site StartingInfernoCommandLine (16:49:08) gdiaz_: inferno-npe i meant (16:49:26) gdiaz_: do you have a $INFERNO_ROOT/ios/init.sh ? (16:49:32) gualteri: mjl-, for those things i just use objdump -d elf.file | less +/0xhexaddr (16:51:28) soul9: gdiaz_, hmm, i think something went wrong, i hardly have anything in the directory (16:51:42) gdiaz_: :? (16:51:57) soul9: charon, lib, keyring, namespace and .svn are my only directories (16:52:12) gdiaz_: where? on INFERNO_ROOT? (16:52:48) soul9: yes (16:53:02) soul9: oh hmmm (16:53:10) soul9: waird thing (16:53:12) gdiaz_: hum could it be something is really broken then? (16:53:20) soul9: thank god i didn't uncompress it on / (16:53:21) soul9: :) (16:53:46) gdiaz_: damn lunix packaging. . . i used and automation tool and now i know what's happening. . . (16:54:02) soul9: gdiaz_, that package is SERIOUSLY broken if it's supposed to be uncompressed on / and create dir /usr/inferno (16:54:21) gdiaz_: it is a slackware package, it is intented to be pkginstalled (16:55:01) gdiaz_: but seems if you need to uncompress it under /usr/lib/inferno already. . . (16:55:05) gdiaz_: i should put notes on that (16:55:06) soul9: well no (16:55:19) soul9: it contains /usr/inferno (16:55:20) soul9: ooooh (16:55:24) soul9: yes, it should (16:55:28) soul9: yeah, that's normal :) (16:55:29) gdiaz_: yes, but that's from the inferno namespace (16:59:08) gdiaz_: that package is dangerous. . . i'm going to do a new one. . . (16:59:38) soul9: :D (16:59:40) soul9: yeah, it is (16:59:50) ***soul9 really, really really happy he didn't extract it to / (17:02:04) gdiaz_: soul9 well, now that you have, uncompress it under /usr/lib/inferno/ and check the ios swcript, i will upload new packages as soon as i finish to build them (17:22:04) gdiaz_: soul9 in 5 minutes there will be a new .tgz, would you minf to check if now decompress it under a inferno/ dir and have the ios script? i already checked here, but the last time i checked it too XDD (17:22:19) soul9: :) (17:22:21) soul9: sure, yea (17:22:34) soul9: this one contains the ios script too, if you uncopress it correctly ;) (17:25:17) gdiaz_: ok the package is there :) (17:25:33) gdiaz_: off to a meeting. . . (17:25:36) gdiaz_ left the room (quit: "BitchX: the official sponsor of the 2002 Olympic Winter Games"). (17:31:18) soul9: ok (17:49:54) soul9: gdiaz: no win :( (17:50:15) soul9: the package now contains a /inferno directory :) (17:58:01) gdiaz_ [n=gdiaz@dos.9grid.es] entered the room. (17:58:16) gdiaz_: hello (17:58:23) gdiaz_: soul9 did you tried the package at the end? (17:59:02) soul9: yes (17:59:08) soul9: ok (17:59:08) soul9: gdiaz: no win :( (17:59:08) soul9: the package now contains a /inferno directory :) (17:59:12) fortyseventeen [n=47teen@c-67-171-113-158.hsd1.ut.comcast.net] entered the room. (18:01:46) gdiaz_: that's the the behaviour it should have, create an inferno/ dir and populate it so you can put it wherever you want (18:01:59) gdiaz_: i'll upload a .deb package in a few minutes too (18:03:35) soul9: ok, well, you said it needed /usr/inferno, so i thought it was wrong (18:03:48) soul9: but a tarball with an inferno/ is fine for me too (18:09:11) gdiaz_: yep, i said that because the old package was a slackware package which should be installed with installpkg and i supposed untaring it would create usr/lib, but something was bad. . . (18:09:24) gdiaz_: so the new tar file just has an inferno/ folder (18:09:40) gdiaz_: now i'm on a ubuntu machine doing a .deb package (18:17:42) soul9: ok (18:17:45) soul9: good luck (18:17:48) soul9: that must be horrible (18:19:14) gdiaz_: not much more than any other package like the msi for windows or the pkg for the mac. . . (18:20:13) soul9: oh, yeah, well, i guess (18:24:45) mennis: There is a really nice free font that quanstro ported for me a while back. It is missing some symbols but I don't usually see that. (18:25:25) mennis: http://www.levien.com/type/myfonts/inconsolata.html < the font (18:27:20) mennis: http://www.kix.in/plan9/mirror/sources/contrib/quanstro/subpixel/icons.tar < the port (18:28:29) gdiaz_: mennins i think the goal is to include some sort of default utf font within the official inferno distribution. . .but i guess incorporating one in the inferno-npe project is a first step into that ? (18:31:56) mennis: I don't really knw what the right path (18:32:00) mennis: is. (18:33:25) mennis: If there was an inferno tool for converting tt to the p9 bitmap fonts an installer could pull the fonts from an open source font site and bitmap them. (18:38:20) mjl-: mennis: that font looks quite nice (18:41:02) npe_ [n=npe@195.207.5.2] entered the room. (18:41:39) stu8ball: ah, the O and 0 are distinct. :-) (18:42:00) mjl-: hmm, i don't like it so much in acme. it's way too blurry (18:44:25) mennis: Really, blurry? (18:45:32) mennis: I use it in drawterm and acme-sac both. (18:45:39) mennis: http://picasaweb.google.com/Michaelian.Ennis/Packetmonkey#5131021241189368818 (18:46:11) mjl-: jups, that's what it looks like here :) (18:46:22) mjl-: i need to use font 18 or 20 to get a similar size to the default font (18:47:25) mjl-: hmm, despite it bluriness, it's better than acme's default monospace font. i was comparing it against acme's normal (non-monospace) font. (18:47:59) mjl-: but the font is still quite beautiful (18:48:37) mennis: I use 18. (18:49:04) mennis: What do you think causes the blurryness? (18:49:28) mjl-: anti-aliased rendering :) (18:49:40) mjl-: i guess it's "normal", depending on what you're used to (18:49:52) mjl-: the switch from plan 9's default font to this makes it feel blurry (18:50:01) mjl-: i've always thought fonts on mac os x look very blurry too (18:50:12) mjl-: compared to e.g. the fonts my firefox on openbsd uses (18:50:23) mjl-: windows fonts seem less blurred than mac os x fonts too... (18:50:29) mennis: Oh so it's the sub-pixel rendering then? (18:50:37) mjl-: i think so (18:50:41) mennis: OK. (18:50:57) mennis: I can see that. (18:56:48) npe left the room (quit: Read error: 113 (No route to host)). (18:57:05) npe_ left the room (quit: No route to host). (19:16:39) [andguent] [n=[andguen@labor.iig.uni-freiburg.de] entered the room. (19:23:38) powerman: soul9: svn ebuild is cool. I was bothering is it will checkout full source on each install, but looks like it should cache sources in distfiles directory (in subdir svn-src/). also it doesn't clear is it will update automatically on emerge world, or I have to force update of that package manually from time to time... (19:44:25) gdiaz_ left the room (quit: "BitchX: now with 38 percent fewer calories!"). (20:33:35) fortyseventeen left the room (quit: Remote closed the connection). (20:36:47) [andguent] is now known as [[andguent]] (20:37:22) [[andguent]] is now known as [[[andguent]]] (20:45:21) [[[andguent]]] is now known as andg (20:45:40) andg left the room (quit: Read error: 104 (Connection reset by peer)). (20:46:47) aguent [n=[andguen@labor.iig.uni-freiburg.de] entered the room. (20:59:35) gualteri left the room (quit: "leaving"). (21:09:55) aguent left the room (quit: "blub"). (21:10:52) fortyseventeen [n=47teen@c-67-171-113-158.hsd1.ut.comcast.net] entered the room. (22:22:27) mennis left the room (quit: Read error: 60 (Operation timed out)). (22:33:05) stu8ball left the room (quit: Read error: 60 (Operation timed out)). (22:34:47) stu8ball [n=stuart@aurora.ossified.net] entered the room. (22:44:15) soul9: powerman-asdf, no, if you don't use paludis svn ebuilds won't get rebuilt automatically (22:44:46) soul9: and yes, it will cache to $DISTDIR/svn-src/packagename iirc (22:45:19) soul9: with paludis, you can specify wether you want to rebuild live ebuilds daily, weekly or monthly (22:48:49) powerman: soul9: yeah. and there several other issues with svn ebuilds. for example, it's harder to tune patches and other hacks used in ebuild because some of them should change for different svn versions (22:49:18) powerman: for example, this line: http://dev.9souldier.org/?p=overlay.git;a=blob;f=dev-inferno/inferno/inferno-99999999.ebuild;h=0cedfc8d0aa99008fa05f8a8ed6299ded9673d5a;hb=HEAD#l34 is not needed with current version (22:49:55) soul9: o? (22:49:57) soul9: ok (22:49:58) powerman: because Charles has fixed that (at least, he note that in changelog, but I havn't tested that myself) (22:51:34) soul9: i see