IRC channel logs

<Viking667>yo. I'm having a problem getting my head around Hurd. I'm brand new, and I'm a dvorak user. I logged in, but can't figure out how to get my consoles to use dvorak keymap. Any clues aside from the rather confusing keyboard(5) manpage?

<Viking667>(and I found out that setupcon doesn't appear to work and complains about only working under Linux or FreeBSD)

<Viking667>Meh, gotta go to bed, it's getting too late.

<luckyluke>Hi Viking667 , I'm not an expert of the keyboard config, but just out of curiosity, how did you log in? I mean, via ssh, serial, on-screen console...

<luckyluke>btw, if you're on debian-hurd, did you try dpkg-reconfigure keyboard-configuration?

<youpi>he's not connected any more

<youpi>for information, dvorak is a layout variant

<youpi>and variants are not supported by the hurd console yet, contribution welcome

<luckyluke>ah ok, thanks for the update (I'm using the matrix bridge currently, apparently it missed the logout notifications)

<luckyluke>on another topic, I was looking at irq on the x86_64, is it still worth to use PIC? or could we require APIC (and stop if not available)?

<youpi>IIRC it was concluded that on x86_64 one really needs to use the APIC

<youpi>they have it anyway so it'll be fine to depend on it there

<luckyluke>ok

<luckyluke>I noticed that the current PIC configuration doesn't work on x86_64 for irqs configured from userspace (e.g. rump), I'll try to set up APIC then

<luckyluke>but without smp (for now)

<zamfofex>Does anyone know whether Debian has some kind of patch for coreutils? My build on Guix is failing on the check ‘tests/df/unreadable.sh’. It’s the only test marked as “FAIL”, though there are some others marked as “XFAIL” (though I’m not sure what the distinction is). See: <https://git.savannah.gnu.org/cgit/coreutils.git/tree/tests/df/unreadable.sh>

<kilobug>zamfofex: it probably does, you can find them in the debian/patches subdirectory when you apt-get source it

<zamfofex>Is there any way to view them online? I don’t have Debian readily available.

<zamfofex>Is this where I should be looking? <https://sources.debian.org/src/coreutils/8.32-4.1/debian/patches/>

<youpi>yes

<zamfofex>It seems none of them have anything to do with ‘df’. The failing test seems to be able to go through the ‘df unreadable || fail=1’ line without setting the ‘fail’ flag, but then sets it on the ‘timeout 10 df fifo || fail=1’ line. On both lines, it warns: ‘df: Warning: cannot read table of mounted file systems: No such file or directory’. Seems I have some investigation to do, I suppose.

<zamfofex>Seems like it uses ‘_PATH_MOUNTED’ from glibc, which is set to ‘/etc/mtab’ on Linux and to ‘/var/run/mtab’ elsewhere, but at least on Guix, there is no ‘/var/run/mtab’ on GNU/Hurd, and instead it is on ‘/etc/mtab’ like in Linux.

<zamfofex>Though on Linux, it doesn’t use ‘_PATH_MOUNTED’, interestingly enough, but rather hardcodes ‘/proc/self/mountinfo’ directly (to which ‘/etc/mtab’ is a symlink).

<zamfofex>Do you reckon I should submit a patch to glibc, coreutils, or Guix?

<youpi>the complex thing will rather be to work out which way it's supposed to get fixed :)

<zamfofex>Yes, exactly! 🙃 I guess for now I might be able to “fix it” by simply making ‘/var/run/mtab’ a symlink to ‘/var/run/mtab’ on my VM, but I’d still like to be able to fix it properly at some point.

IRC channel logs

2022-09-13.log