IRC channel logs
2024-04-04.log
back to list of logs
<gnucode>hmmm. I'm probably not knowledgeable enough to help out there. <Pellescours>reminder for me to inspect (and if someone can confirm it’s apparently a bug): check if implementation of proc/info.c:S_proc_getprocargs is correct. The displayed command line for console (in `ps aux`) does not represent the real command line <Pellescours>example vga appear twice while it should be "vga --font-width=9" <white-wolf>solid_black, less than 10k line of code considering to a small kernel ? <white-wolf>mach is coded in C object oriented ? with what norme of C ? <white-wolf>there have dvorak international with death key in hurd ? <Pellescours>hurd need a lot of stuff, drivers to run in userland. the current support is very limited for example <Pellescours>with my recent patch that use libxkbcommon for keyboard layout, dvorak international with dead kzy should be supported, it's just not available yet in debian (unless you compile hurd latest and configure the console command line) <white-wolf>ok i have an society project, blues softwares. can pay developpers to add features. it's totaly desinteresed. see https:www.blues-softwares.net i balance my project to hurd, need actualize my wiki <white-wolf>if you ok with my idear for hurd, when i create my society i pay developpers to add stuff to hurd, peraps support of xeon and risk V 128 bits <white-wolf>i'm sure that propriarty of xeon processor not use in another kernel <Pellescours>support for proc architecture is done in the kernel (gnumach), i know someone is trying to add riscv support <white-wolf>how many developpers you need realy to upon hurd ? <anatoly>128 bit riscv doesn't exist in hardware and is a matter of distant future <damo22>white-wolf: please introduce yourself on the mailing list so our maintainer can discuss with you <damo22>sorry i havent read my emails today <white-wolf>damo22, an acces to dev-hurd@gnu.org is more easy to speach techinal with dev <white-wolf>if you havn't my money become to trading activity... i'm futur trader, but i am handicaped too and money don't motive me, juste death proprietary softwares <white-wolf>ACTION smoke again, that is was very difficulte to get 8h am <white-wolf>imagine GNU/Hurd in international space station, or in moon colony <white-wolf>but need thinking about buisness plan for rentabilize my french society... may be pro version, anonymous version, peraps for nord corean or chinese. or russian. an customer service or else, need think about <white-wolf>sorry if my english isn`t correct, i learn only by irc <white-wolf>damo22, anatoly how can i have acces to dev-hurd@gnu.org ? <damo22>white-wolf: we dont use dev-hurd <damo22>you are welcome to post on bug-hurd, its a fairly low traffic list anyway <damo22>white-wolf: the foundation that looks after GNU is called Free Software Foundation fsf.org <damo22>it may have philosophies in your language <white-wolf>what address can use to subscribe by my mail client, please ? <damo22>also the philosophy of the GNU project is here: <damo22>there is an icon on that page on the top that can change to many languages <white-wolf>ok damo22 i read, not all but read more, i save the link <white-wolf>can i do a copy on my wiki www.blues-softwares.net ? <white-wolf>the donation to the fsf is deductible to the french taxe ? <solid_black>I'm not going to run tokei, but more than 30M according to a quick web search <solid_black>most of that is drivers, sure, but not having to put drivers in kernel mode is eactly one of the points of microkernels <solid_black>anatoly: I haven't pushed into the alpine hurd repo for a while :| <solid_black>maybe I should go back to that, now that it should be possible to build and run aarch64-gnu <solid_black>Pellescours: I'm not seeing the 'ps aux' bug? do you have more details? <solid_black>Pellescours: it's "AArch64", "aarm64" is not a thing :D <solid_black>it is sometimes (mistakenly) called "arm64", for example in Linux <solid_black>also it would be great if anyone would help me with this <solid_black>as for FSF (and I need to word this very carefully): FSF is certainly directly related to GNU and GNU Hurd, and money donated to FSF will certainly go towards advancing/protecting free software in general, but, I don't think any of that many would go to any of the Hurd developers or contributors <solid_black>SeernityOS Kernel (including AK, but excluding LibDeviceTree, LibEDID, LibELF, LibVT, LibCrypto, LibPartition) is 174 KLoC <Pellescours>solid_black: if you check the command line of the hurd-console you have "vga vga" instead of "vga --font-width=9", what is strange it’s that the issue only appear for the console process. It seems only to appear for the console <solid_black> /bin/console --daemonize -d current_vcs -c /dev/vcs -d vga -d pc_kbd --keymap us pc_kbd -d pc_mouse --protocol=ps/2 pc_mouse <Pellescours>real command line should be "/bin/console --daemonize -d current_vcs -c /dev/vcs -d vga --font-width=9 -d pc_kbd --keymap fr --repeat=kbd -d pc_mouse --protocol=ps/2 --repeat=mouse" while I have "/bin/console --daemonize -d current_vcs -c /dev/vcs -d vga vga -d pc_kbd --keymap fr pc_kbd -d pc_mouse --protocol=ps/2 pc_mouse" <solid_black>what does the command line look like from inside the console process? <Pellescours>I just realized, the only arguments replaced are the long form argument (--) of the plugins of the console <Pellescours>when I add an "echo $DAEMON_ARGS" in /etc/init.d/hurd-console and do a "service hurd-console status" If have the first line <Pellescours>--daemonize -d current_vcs -c /dev/vcs -d vga -d pc_kbd --keymap us pc_kbd -d pc_mouse --protocol=ps/2 pc_mouse <Pellescours>--daemonize -d current_vcs -c /dev/vcs -d vga --font-width=9 -d pc_kbd --keymap fr --repeat=kbd -d pc_mouse --protocol=ps/2 --repeat=mouse <solid_black>before declaring it the proc server issue, see what arguments /bin/console actually got <Pellescours>--daemonize -d current_vcs -c /dev/vcs -d vga vga -d pc_kbd --keymap fr pc_kbd -d pc_mouse --protocol=ps/2 pc_mouse <solid_black>Pellescours: in gnumach, mainly getting more serious about drivers & device tree <solid_black>currently things are very basic and somewhat hardcoded for the "virt" machine <solid_black>if you would build it on your machine, test that it works, that would already be great <solid_black>so far I know it works for me and azert, but not for luckyluke <Pellescours>I found the cause of the gibberish I was having. I did not setup the vcs <anatoly>solid_black I was wondering if you have few more commits/changes which you can push. Not that I'm asking if you've done more work on it <gnucode>well I should probably start another qoth <gnucode>What's funny is I thought I had started working on it, and apparently not... <almuhs>hi. I've just find a problem meanwhile i try to install gcc <almuhs>pruebas@debian-hurd:~$ sudo apt install gcc <almuhs>Building dependency tree... Done <almuhs>Reading state information... Done <almuhs>You might want to run 'apt --fix-broken install' to correct these. <almuhs>The following packages have unmet dependencies: <almuhs> cpp-13 : Depends: cpp-13-i686-gnu (= 13.2.0-23) but it is not going to be installed <almuhs> cpp-i686-gnu : Depends: cpp-13-i686-gnu (>= 13.2.0-11~) but it is not going to be installed <almuhs> gcc-13 : Depends: gcc-13-i686-gnu (= 13.2.0-23) but it is not going to be installed <almuhs> gcc-i686-gnu : Depends: gcc-13-i686-gnu (>= 13.2.0-11~) but it is not going to be installed <almuhs>E: Unmet dependencies. Try 'apt --fix-broken install' with no packages (or specify a solution). <youpi>did you try apt --fix-broken install <almuhs>i tried to run apt previously from smp mode, and the VM keeps freezed, so maybe it corrupted the installation <almuhs>oops: pruebas@debian-hurd:~$ gcc hilos.c -o hilos.c -pthread <almuhs>i had to "apt install gcc" after --fix-broken-install <Pellescours>FYI I noticed that htop does not display the process list anymore. I don’t have time to invest in it <almuhs>i have a pending fix: the stat table shows always cpu 0, even when this thread is running in other cpu <almuhs>gnumach shows the real cpu which is running the process, but must be some error in proc or procfs, that avoid this value will be written correctly in the stat file of the process <almuhs>i use the old patch of damo22 which modify the scheduler to be able to run in smp mode <almuhs>but it not stable. When I make operation which requires intensive using of harddisk, it usually freeze <Pellescours>Yes, there is a problem, when you stress memory or disk with rumpdisk (not related to SMP)