IRC channel logs
2025-06-13.log
back to list of logs
<damo22>p4r4D0xum: hurd is ported to amd64 <lnnk>OK cool which filesystem does Hurd support and which init system are supported? <lnnk>Does Hurd has support for system d ?and does it have support for modern GPU ? <aaabbb>lnnk: no it doesn't support systemd. systemd requires linux-specific features <azert>lnnk: the Hurd is not in any way supported by redhat <azeem>what does that have to do with redhat <azert>i find working for private company without getting paid unethical, particularly when it is a giant multinational, as such I’d not support red hat os with my work <azeem>azert: yeah ok, but what does that have to with systemd <azeem>I mean, it's Free Software, released under the LGPL <aaabbb>but it requires a large number of features that are linux-specific <azert>The way I see it it’s embrace and destroy <aaabbb>systemd is funded by redhat and is intended only for use on linux. it would require massive changes to either systemd or to hurd to get them to work together <azeem>I am not sure azert realizes that in the early 2000s, one of the main Hurd developers and the main glibc developer were working at redhat <damo22>redhat isnt what it used to be since IBM bought it i guess <azeem>yeah well maybe, but I also think we're past systemd gripes <azeem>it's not going to work on the Hurd, move on <damo22>maan i need to work out this zero filled packet probem <damo22>i should have working smp system by now <azert>azeem: it’s not the 2000 anymore. <azert>now we have to deal with lots of cancer that wasn’t there 25 years ago <azert>specifically, everyone was happy if you could do stuff back then. Now they introduced all sorts of pressures for people to desist <azert>systemd is the kind of super integrated software sype <azert>supervisor you’d expect in mainframes <azeem>well yeah, I'm not interested in systemd discussions <azert>lets then talk about documentation. In the late 2000 you could still find excellent docs on the internet to implement yourself all kind of servers <azert>such thing is inexistent now <zamfofex>lnnk: The Hurd has support for the Shepherd, which is the init system used for Guix System (on both Linux and the Hurd). It is a very capable init system, and it integrates well with Guix. <zamfofex>Though note that I have had limited success using the Hurd with Guix. From what I’ve seen, Debian is a better alternative if you want to try out the Hurd. <lnnk>Which file systems are supported by Hurd? <lnnk>Cool Shepard I herd about it. <identity>and 'herd' is a command to interface with Shepherd :) <gfleury>youpi: can i move thread_key_{delete, create} and thread_{getspecific, setspecific} in one commit as it will be simple? <youpi>to get this for 2.42, it would be useful to have it before june 20th <sneek>gnucode, you have 1 message! <azeem>I keep hitting this assertion failure (since I managed to boot an apic-enabled kernel it actually prints it before the VM hangs, I assume it was the same problem before): <azeem>../../ext2fs/getblk.c:103: ext2_alloc_block: Assertion 'result >= (((((sblock_block << 10)) >> log2_block_size) + 1) + ((((off_t) ((((groups_counlt * sizeof(struct ext2_group_desc))) + block_size - 1) >> log2_block_size) << log2_block_size)) >> log2_block_size)) && result < store->size >> log2_block_size' failed. <azeem>is this "just" file system corruption and should I rebuild it, or something more interesting? <gnucode>p4r4D0xum: I'm wanting to write..."You can build this file with gcc -pthread -rt -o context-switch context-switch.c" <gnucode>is that the right command to compile that file ? <gnucode>Also, it's SUPER easy to edit the wiki using the "Edit" button on the darnassus wiki. youpi normally uploads my changes in a day or two. His response time is incredible! <youpi>azeem: I have seen such when memory pressure is high <youpi>my guess is that swapping has some troubles, making ext2fs write bogus information and step on itself <youpi>I added that assertion to try to avoid overwriting critical ext2 information <azeem>well, I now tried to run e2fsck -y -f a second time, and it keeps finding the same issues <azeem>ok, when it just hangs without ext2fs assert, there's no memory pressure, or least swap isn't used <lnnk>Which file systems are supported by hurd?if there is up-to-date documention can anyone please provide link? <lnnk>I usually use ext4 or btrfs is one of these supported? <identity>lnnk: no, ext4 and btrfs are not supported <lnnk>OK so is there a list of supported file systems? <lnnk>I can't find any documention on file systems. <lnnk>From the official websites <identity>lnnk: well it definitely also supports swapfs <identity>you probably are not getting any fancier than ext2 <azeem>lnnk: what is the issue with file systems in particular? <lnnk>I used ext4 it did not boot VM crashed. <lnnk>I tried replacing Linux kernel with hurd in Debian 12 and system crashed. <lnnk>4,99identity:99,99I though swapfs is a sudo file system ? <gnucode>lnnk: Did you install Debian GNU/Linux.... and then replace Linux with the Hurd !?? Man you are brave! <gnucode>Also, Debian GNU/Linux uses a different userland than Debian GNU/Hurd. Emacs compiled for GNU/Linux will not run on GNU/Hurd. :) <lnnk>identity: you said swapfs.but I think swapfs is a pseudo file system ?am I wrong <identity>lnnk: well, yeah, i meant that "well if you do not like ext2 you can store your files in memory and then swap them out", jokingly <azeem>lnnk: this channel is for technical discussions about the Hurd, please keep it on-topic <lnnk>OK so ext2 is supported thank you. <gnucode>azeem: lnnk did stay on topic. He was talking about the Hurd. :) Am I missing something? <gnucode>and lnnk is brave/daring because he tried to use ext4 and the Hurd! <gnucode>The Hurd only supports ext2 by the way. <gnucode>and lnnk You are brave because I believe you are trying to share the userland of the Hurd and Linux, which is not possible