IRC channel logs

2020-05-10.log

back to list of logs

***Server sets mode: +nt
<Posterdati>youpi: thanks!
<Posterdati>hi
<Posterdati>please help: I downloaded the install image, but on qemu-system-x86_64 the emulation is very slow!
<Posterdati>thanks
<Posterdati>I'd like to make sbcl run on top of hurd, has anyone tried that?
<Pellescours>Posterdati: add the -enable-kvm option
<Posterdati>Pellescours: ok
<Posterdati>Pellescours: better now!
<Posterdati>Pellescours: thanks
<Posterdati>what is this error: getty[xxxx]: Bad syntax
<Posterdati>?
<youpi>it's fixed in latest upgrades
<Posterdati>I upgraded and then appeared this problem
<youpi>which version of the hurd package do you have?
<youpi>the latest is 1:0.9+git20200314-1
<Posterdati>I downloaded the DVD image from the last url of the channel topic
<Posterdati>debian-sid-hurd-i386-DVD-1.iso
<youpi>the hurd package I said
<youpi>not the hurd image
<youpi>dpkg -l hudr
<Posterdati>ok
<Posterdati>1:0.9.git20191228
<Posterdati>1:0.9.git20191228-1
<youpi>so that's not latest, do upgrade
<Posterdati>ok
<Posterdati>apt-get dist-upgrade
<Posterdati>ok
<Posterdati>it is downloading 41 MB of stuffs
<Posterdati>I did dist-upgrade before adding repos in the apt source file :)
<Posterdati>ok
<Posterdati>is it a 32bit os?
<youpi>for now yes
<youpi>please read the faq
<Posterdati>ah
<Posterdati>I'm reading the one on savannah
<Posterdati> https://www.gnu.org/software/hurd/contributing.html
***janneke_ is now known as janneke
<Posterdati>nice it freezes during upgrade
<Posterdati>cannot install vim too :(
<gnu_srs> Posterdati: Maybe it is network problems. Which NIC do you use on the qemu command line?
<Posterdati>network's working
<Posterdati>I have held broken packages
<gnu_srs>e1000 is known to work well. rtl8139 is flaky.
<Pellescours>Posterdati: vim is installed on the debian image, last time I did the dist-upgrade, it broke vim
<Posterdati>broken!
<Pellescours>doing a simple "upgrade" upgrades hurd
<Posterdati>nice
<Posterdati>is X usable with qemu?
<Pellescours>I think yes, but I didn't test it personally
<Posterdati>startx returned an error: unable to determine if running on a console
<youpi>please read https://www.debian.org/ports/hurd/hurd-install
<Posterdati>youpi: already read
<youpi>Posterdati: it talks about allowing to start X whatever the console state
<Posterdati>that solution did not work for me
<Posterdati> https://unix.stackexchange.com/questions/478742/error-when-trying-to-use-xorg-only-console-users-are-allowed-to-run-the-x-serve
<Posterdati>this worked for me
<Posterdati>but X still won't work
<youpi>which solution ? going root?
<youpi>how did you configure in the dpkg-reconfigure menu ?
<Posterdati>no, console command
<youpi>I mean that: You need to run dpkg-reconfigure x11-common xserver-xorg-legacy to allow any user to start Xorg, because the X wrapper does not
<youpi>know about the Hurd and Mach consoles.
<youpi>in the X.Org section
<youpi>I don't ses what else you have been looking at
<Posterdati>I did the dpkg-reconfigure stuff, but still got errors
<youpi>well, the dpkg-reconfigure stuff is *exactly* the /etc/X11/Xwrapper.config modification
<youpi>again, what did you choose in the dpkg-reconfigure menu?
<Posterdati>anybody
<youpi>I'm lost in what you did
<youpi>when you wrote "this worked for me", what exactl in that page did work for you?
<youpi>I see a lot of stuff there
<Posterdati>I changed to anybody
<youpi>ok, so again that's *exactly* what dpkg-reconfigure is about
<Posterdati>the error about console gone, but X won't work
<youpi>so I don't know how you did it, but apparently it's incoherent
<youpi>"won't work" doesn't mean it didn't have effect
<Posterdati>yes but there are more problems here
<youpi>did you read the rest of the hurd-install page about the window manager?
<Posterdati>it is not matter of window manager
<Posterdati>it is an X error (EE)
<Posterdati>more than one
<youpi>there are thousands of differnet possible X errors
<youpi>if you don't show what you actually get, there is no way for us to help you
<youpi>only guess
<Posterdati>ok
<youpi>and my guess is that you don't have a window manager
<youpi>if that doesn't please you as an answer, then actually properly ask the question
<Posterdati>I installed wmaker
<youpi>did you specify it on the startx command?
<Posterdati>and did startx /usr/bin/wmaker
<youpi>ok so at least it seems you have done what is documented
<youpi>but you still haven't actually said what error you aregetting
<Posterdati>I'm pasting Xorg logs
<Posterdati> https://paste.debian.net/1145968/
<youpi>ok so there's an issue withthe keyboard driver
<youpi>had you started with that, we'd have save ourself a lot of time
<Posterdati>then?
<youpi>then we have to fix it
<youpi>(really, running X is not really fancy, it's just X, you get the same with Linux)
<Posterdati>ok
<Posterdati>well I'd like to use emacs under X
<Posterdati>is more funny
<Posterdati>than in console
<janneke>i think someone (civodul?) suggested that it's possible to use file system attributes to store something like "settrans -c /servers/socket/1 /hurd/pflocal" in the file system
<janneke>how would i do that?
<civodul>hey!
<civodul>janneke: by default /hurd/ext2fs uses a custom hack to store "passive translator" settings in the file system
<civodul>it happens automatically when you do "settrans foo /hurd/bar" (without '-a')
<civodul>but /hurd/ext2fs has an option to use extended attributes (xattr) instead of the custom hack
<civodul>but xattr is marked as experimental
<civodul>so it's xxattr in a way
<janneke>hehe
<civodul>there's a command-line option in /hurd/ext2fs
*janneke looks
<janneke>civodul: that would be: --x-xattr-translator-records Store translator records in extended attributes (experimental)
<janneke>
<janneke>hmm, i was hoping to do that from the guix cross build (i.e., from gnu/linux)
<janneke>so i may need to port that, hmm
<janneke>...or maybe not; the --x-xattr-translator-records is about /reading/; setting xattr's might be fine
<civodul>janneke: the new image API might be able to store xattrs
<civodul>do we need passive translators settings?
<janneke>civodul: i have no idea ...
<janneke>i want to get rid of the need to run bash to do: "settrans -c /servers/socket/1 /hurd/pflocal"
<janneke>i already got rid of runsystem, init.c, runttys, runsystem.hurd
<janneke>hmm, i find the documentation hard to read
<Posterdati>is there a need for hurd developer? (I read the announce on savannah)
<janneke>it says "they are completely different things"
<janneke>yet, it seems that a "passive translator" is the function/command that starts an active translator
<janneke>ah, no that seems not to be the case at all
<janneke>hmm, it could just be "too straightforward to document"
<janneke>something like "setfattr --name=gnu.translator --value=/hurd/pflocal /servers/socket/1"
<civodul>yes
<janneke>unsure whether the f in setFattr should have been an X because it sets extended attrs
<janneke>obvious things can really puzzle me, at times :-)
<civodul>setfattr is the command to set xattrs (in the 'attr' package)
<janneke>hehe
<janneke>"we can't make it faster, but we can make it more confusing"
<janneke>aka what's in a name
<janneke>anyway, ty civodul; /me goes to play & try some things
<civodul>heh :-)
<youpi>stressing the xattr support to make sure it works before we enable it by default would be useful, yes
***Emulatorman___ is now known as Emulatorman
*janneke hears encouragement :-)
<janneke>hmm ... setfattr: 1: Operation not supported
<janneke>how helpful
*janneke downloads "attr" source
<janneke>on gnu/linux, this works: /gnu/store/0g0j5j7m52mcfwazd8scip4qb7r001m9-attr-2.4.48/bin/setfattr -n user.foo -v bar 2
<janneke>but it seems that "gnu.translator" is rejected
<janneke>i cannot imagine such a string would be checked...but...
<janneke>hmm, it's a kernel thing?
<janneke>hmm, possibly a mount option
<janneke>i'm afraid that setting "gnu.translator" cannot is refused by a regular linux kernel :-(
<janneke>hmm, the "#linux" channel here is actually about gnu/linux :-(
<janneke>i found something interesting: https://lists.gnu.org/archive/html/bug-hurd/2004-02/msg00108.html
<janneke>i'm afraid the code wasn't upstreamed to linux