| 00:07.17 | *** join/#edev grummund (~user@unaffiliated/grummund) |
| 01:11.05 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 01:11.58 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 02:06.32 | *** join/#edev fire_ (~fire@unaffiliated/codingenesis) |
| 02:52.31 | *** join/#edev fire__ (~fire@unaffiliated/codingenesis) |
| 03:10.18 | *** join/#edev codingenesis (~fire@unaffiliated/codingenesis) |
| 04:05.49 | *** join/#edev blupp (~blupp@dslb-188-099-240-178.pools.arcor-ip.net) |
| 04:24.30 | *** join/#edev [7] (~quassel@rockbox/developer/TheSeven) |
| 04:29.41 | *** join/#edev codingenesis_ (~fire@unaffiliated/codingenesis) |
| 04:35.28 | *** join/#edev fire_ (~fire@unaffiliated/codingenesis) |
| 04:39.57 | *** join/#edev fire__ (~fire@unaffiliated/codingenesis) |
| 04:44.52 | *** join/#edev codingenesis (~fire@unaffiliated/codingenesis) |
| 04:45.00 | *** join/#edev _Lucretia_ (~munkee@pdpc/supporter/active/lucretia) |
| 05:46.02 | *** join/#edev clio (~andrej@85.159.109.222) |
| 06:09.01 | *** join/#edev mbuf (user@nat/redhat/x-xbbszpwnpxzntsfq) |
| 06:16.33 | *** join/#edev bsdfox (~Bob@c-71-197-83-152.hsd1.ca.comcast.net) |
| 06:16.33 | *** join/#edev bsdfox (~Bob@unaffiliated/bsdfox) |
| 07:36.12 | *** join/#edev landley (~landley@cpe-72-177-21-136.austin.res.rr.com) |
| 08:05.09 | Belliash | landley: ping |
| 08:22.43 | *** join/#edev codingenesis_ (~fire@unaffiliated/codingenesis) |
| 08:23.45 | *** join/#edev florian_kc (~fuchs@Maemo/community/contributor/florian) |
| 08:27.10 | *** join/#edev calculu5 (~calculus@gentoo/user/calculus) |
| 08:32.26 | *** join/#edev fire_ (~fire@unaffiliated/codingenesis) |
| 08:38.36 | Belliash | landley: we were discussing yesterday with slashbeast and SpeedEvil... about running a few instances of single program and about memory usage. Everything begun with that i wanted to launch a script in haserl what will call another haserl instance, both of them would use buybox as an interpreter, so finally there would be 2 instances of haserl and busybox running, |
| 08:38.44 | Belliash | while interpreting 2 different scripts |
| 08:39.33 | Belliash | i was wondering about the memory usage, if the 2 instances would much increase the memory usage |
| 08:41.01 | Belliash | from what SpeedEvil said, only the shared libraries only share VM space and ordinary programs do not, but on the other hand |
| 08:42.08 | Belliash | slashbeast mentioned that the memory usage shouldnt increase (the programs shouldnt be loaded twice into memory, excluding files to parse) |
| 08:43.08 | Belliash | because in fact the data is already read into memory, the inode is known and unless it was altered it supposes to share the memory |
| 08:46.10 | *** join/#edev prpplague (~danders@55.Red-88-2-50.staticIP.rima-tde.net) |
| 08:50.14 | Belliash | :( |
| 09:08.51 | *** join/#edev calculus (~calculus@gentoo/user/calculus) |
| 09:27.30 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 09:37.57 | *** join/#edev jackmitchell (~Thunderbi@195.171.99.130) |
| 09:53.00 | *** join/#edev Crofton (~balister@78.Red-88-2-49.staticIP.rima-tde.net) |
| 10:08.25 | *** join/#edev clio (~andrej@85.159.109.222) |
| 10:35.46 | SpeedEvil | Belliash: I'm basing this on looooong ago. |
| 10:36.42 | SpeedEvil | when I had a laptop with single digit megs and was using 'mergemem' to merge pages. |
| 10:36.50 | Belliash | the memory management changed few times already afaik, i just dont want to waste a memory on embedded devices |
| 10:37.14 | SpeedEvil | also |
| 10:37.19 | Belliash | but every solution has its pros & cons |
| 10:37.35 | SpeedEvil | http://www.selenic.com/smem/ |
| 10:37.44 | SpeedEvil | tool to measure it |
| 10:38.08 | Belliash | i decided to try one, as it suits more my needs, but this might influent on memory usage |
| 10:38.11 | SpeedEvil | http://www.complang.tuwien.ac.at/ulrich/mergemem/ |
| 10:39.43 | *** join/#edev Posterdati (~antani@host45-237-dynamic.6-87-r.retail.telecomitalia.it) |
| 10:39.47 | Posterdati | hi |
| 10:40.00 | Posterdati | is anyone know the cc110l? |
| 10:40.35 | Belliash | mergemem looks goo, but im not sure if this will be useful on embedded devices, when we are talking about small programs, like haserl. Dont forget that it also needs to fit into flash, so additional binary means bigger image |
| 10:41.58 | SpeedEvil | Belliash: see above |
| 10:42.00 | Belliash | s/goo/good/ |
| 10:42.37 | SpeedEvil | I started out using it on a laptop with 5 megs |
| 10:43.21 | Belliash | 5 megs of what? ram? rom? |
| 10:44.21 | SpeedEvil | ram |
| 10:44.54 | SpeedEvil | I don't think it was ported to 2.4 kernels those |
| 10:44.58 | SpeedEvil | though |
| 10:52.59 | *** join/#edev g1powermac (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 10:52.59 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 11:17.27 | *** join/#edev gustavoz (~gustavoz@host187.190-139-30.telecom.net.ar) |
| 11:58.15 | *** join/#edev cbrake (~cbrake@oh-69-34-21-229.sta.embarqhsd.net) |
| 12:36.03 | *** join/#edev codefire_ (51531411@gateway/web/freenode/ip.81.83.20.17) |
| 12:36.07 | codefire_ | hi all |
| 12:36.13 | codefire_ | i am new here |
| 12:36.33 | codefire_ | is it the right place to get help about using JFFS2 with summary on embedded system ? |
| 12:41.45 | *** join/#edev Crofton (~balister@55.Red-88-2-50.staticIP.rima-tde.net) |
| 14:38.58 | *** join/#edev BabakKia (~babak@mail.pemicro.com) |
| 14:46.18 | *** part/#edev BabakKia (~babak@mail.pemicro.com) |
| 14:47.06 | *** join/#edev gustavoz (~gustavoz@host187.190-139-30.telecom.net.ar) |
| 15:22.33 | *** join/#edev chouimat (~mathieu@206.191.34.211) |
| 15:22.33 | *** join/#edev chouimat (~mathieu@kde/developer/chouinard) |
| 15:30.22 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 15:30.22 | *** mode/#edev [+o GPSFan] by ChanServ |
| 16:43.35 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-fef8de00-126.dhcp.inet.fi) |
| 16:55.01 | *** join/#edev SpeedEvil (~quassel@tor/regular/SpeedEvil) |
| 16:57.06 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 16:57.06 | *** mode/#edev [+o GPSFan] by ChanServ |
| 17:03.44 | *** join/#edev bsdfox (~Bob@unaffiliated/bsdfox) |
| 18:00.01 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 18:11.21 | *** join/#edev florian (~fuchs@sign-4d091bca.pool.mediaWays.net) |
| 18:11.22 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 19:10.35 | *** join/#edev landley (~landley@cpe-72-177-21-136.austin.res.rr.com) |
| 19:16.49 | landley | belliash: hello. (Reading backscroll...) |
| 19:17.02 | Belliash | hello, landley |
| 19:17.22 | Belliash | i have also sent you an e-mail, 'cos i was not sure if you will manage to read that |
| 19:18.20 | landley | slashbeast, belliash: ordinary programs totally share their program image. It's mmap(MAP_PRIVATE), so only dirtied pages get a per-process copy. |
| 19:18.45 | landley | Your stack, heap, and environment space get allocated per program, but if you're really clever and running something like "true" you can get that down to about 3 pages I think. |
| 19:18.45 | SpeedEvil | I was confused |
| 19:19.06 | landley | SpeedEvil seems to be talking about windows. |
| 19:19.08 | SpeedEvil | shouldn't comment on memory management when fitted |
| 19:19.12 | SpeedEvil | tired |
| 19:19.16 | SpeedEvil | naah |
| 19:19.22 | SpeedEvil | x |
| 19:19.58 | landley | My email is still a bit... unpolished. But I'm getting to messages within a week now. :) |
| 19:20.11 | SpeedEvil | congrats! |
| 19:20.15 | chouimat | hi ladies |
| 19:20.23 | SpeedEvil | has problems that way too. |
| 19:20.27 | slashbeast | landley: still thunderbird? |
| 19:20.48 | landley | belliash: most of the point of PIC and PIE code is to group the data the dynamic linker needs to change into an array, so it's not dirtying pages all over the executable. |
| 19:21.10 | landley | Statically linked programs should just mmap the executable image and run with no local dirty pages. |
| 19:21.20 | landley | slashbeast: finally got balsa to sort of work, with caveats and a brick. |
| 19:21.34 | landley | And a python program to do the mail filtering. |
| 19:21.43 | slashbeast | landley: isn't balsa dead project? |
| 19:21.58 | Belliash | landley: thanks a lot for clarification :) |
| 19:21.59 | slashbeast | Current revision - 2.3.28 - released on 2009-01-18 |
| 19:22.02 | landley | slashbeast: no, their list has ~3 people on it who respond to bug reports. |
| 19:22.31 | landley | Um, they had a release last week with some of my bugs in it, and the version I'm running is from this year. You're looking at the old website. |
| 19:22.48 | slashbeast | the frist one from G |
| 19:22.50 | landley | It's got a _ton_ of rough edges that need smoothing out, and its environmental dependencies are crazy. |
| 19:23.06 | landley | (I should be able to compile without 3 separate spellchecker packages, thank you.) |
| 19:23.22 | slashbeast | the 'GNOME' in title worries me a bit |
| 19:23.35 | landley | http://pawsa.fedorapeople.org/balsa/ |
| 19:23.48 | landley | Worries me too, but thunderbird was worse. |
| 19:24.34 | slashbeast | well, first the GNOME in title, next is non-optional dep on systemd |
| 19:24.43 | slashbeast | Everything goes that way |
| 19:25.01 | slashbeast | I still refuse to accept the existence of systemd. |
| 19:25.10 | landley | Where'd you see systemd? I'm running an upstart system and balsa works with it. |
| 19:25.24 | slashbeast | I am talking about future. ;-) |
| 19:25.26 | landley | I need to figure out what init to do in toybox. I have to read through the android code. |
| 19:25.34 | landley | slashbeast: ubuntu is fighting systemd and so far android is ignoring it. |
| 19:25.39 | slashbeast | make app depend on specified kit/DE, next make it depend on weird init system. |
| 19:25.59 | landley | Of the two, android is more important. Samsung finally did one of the "turn the phone into a pc" USB thingies I've been talking about for years now... |
| 19:26.20 | landley | http://www.gizmag.com/samsung-galaxy-smart-dock/24820/ |
| 19:26.23 | SpeedEvil | whichdiddn'y really well |
| 19:26.26 | SpeedEvil | sell |
| 19:26.42 | landley | It's not generic enough. |
| 19:27.02 | landley | What we _need_ is a USB3 hub with built-in peripherals, and an android software package you can install from the android market onto an arbitrary ice cream sandwich phone. |
| 19:27.05 | SpeedEvil | yeah |
| 19:27.20 | landley | What they gave us is a galaxy-specific dock, but it shows there's interest in the concept. |
| 19:27.37 | landley | "This is how the technology will genericize" doesn't mean they'll do it this year. Anybody remember the Tandy PC back in the 1980's? |
| 19:27.55 | calculus | motorola had docks to, with a keyboard, touchpad, bigger screen, and speakers |
| 19:27.58 | calculus | too* |
| 19:28.12 | landley | It took a decade for the PC to turn into something vaguely vendor-neutral with proper standards compliance (and standards to be compliant to). |
| 19:28.18 | landley | And that's counting from 1982, not from 1975... |
| 19:28.39 | landley | Really what we need is a generic USB3-HDMI adapter, with drivers in vanilla linux. |
| 19:29.03 | SpeedEvil | sighs at markets. |
| 19:29.03 | landley | Too bad the HDMI consortium has its head so far up its own ass... |
| 19:29.04 | chouimat | hehehe |
| 19:29.12 | SpeedEvil | 4:3... |
| 19:29.29 | landley | SpeedEvil: doesn't matter, HDMI is HDTV, and HDTV is generic mass produced. |
| 19:29.40 | chouimat | landley: don't expect it soon since they breathing toxic air anyway, they will all die before it happen |
| 19:29.52 | landley | Computers have ONCE AGAIN come full circle to the point we're using televisions as display devices. |
| 19:29.57 | landley | Welcome back Jack Trameil. |
| 19:29.57 | calculus | landley: I know how much you want to upgrade gcc, but it looks like clang can't compile the kernel either, http://comments.gmane.org/gmane.linux.kernel/1385743 |
| 19:30.41 | landley | calculus: gcc 4.7 couldn't compile the kernel either. Didn't stop people from fixing things at both ends. |
| 19:31.19 | calculus | apparently it stops at the kernel end when it comes to clang |
| 19:31.35 | landley | Yup. This is nothing new. |
| 19:31.43 | landley | I have perl removal patches I maintain locally to build without perl. |
| 19:32.04 | calculus | have you tried getting those upstream recently? |
| 19:32.22 | landley | You regression test each new kernel starting with about -rc3, find the new insanity they've done, patch it out, post the patch to linux-kernel so you can feel smug about them ignoring it, and move on. |
| 19:32.27 | landley | calculus: I really should. |
| 19:32.51 | landley | I'm nominally Documentation maintainer but could NOT get a kernel.org account to jump through the security hoops they locked the barn door with after the horse escaped last year... |
| 19:32.55 | chouimat | calculus: seen this? http://git.linuxfoundation.org/?p=llvmlinux.git;a=shortlog |
| 19:32.59 | landley | I should set upa git tree on landley.net and push patches through that instead. |
| 19:33.54 | calculus | chouimat: nope, didn't realize that this was a linux foundation effort either |
| 19:33.57 | landley | But first I'm tryign to get an aboriginal linux release out, and I'm having a uclibc build bug that ONLY happens when I build all targets in parallel. It builds fine with a single-target build. ARGH! |
| 19:34.14 | landley | I need to confirm it's not a toybox bug, which means I need to figure out what it's doing. Same old... |
| 19:34.37 | landley | calculus: the linux foundation doesn't like the FSF any more than anybody else does. They're just too politic to say it out loud. |
| 19:35.05 | landley | LF is caught between google/android, Red Hat, and the FSF. |
| 19:35.26 | landley | Well, Red Hat vs Ubuntu anyway. |
| 19:36.24 | landley | I really, really miss kernel-traffic. |
| 19:40.02 | *** join/#edev ccole (~cole@coledd.com) |
| 20:00.20 | *** join/#edev Posterdati (~antani@host45-237-dynamic.6-87-r.retail.telecomitalia.it) |
| 20:47.00 | *** join/#edev mpk (mp@aggr.com) |
| 22:49.33 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |