jaeger changed the topic of #crux-devel to: CRUX (https://crux.nu/) development channel | Logs: https://libera.irclog.whitequark.org/crux-devel/
farkuhar has left #crux-devel [#crux-devel]
dlcusa has quit [Ping timeout: 268 seconds]
Belgacem1958 has joined #crux-devel
dlcusa has joined #crux-devel
AndroUser has joined #crux-devel
Belgacem1958 has quit [Ping timeout: 260 seconds]
braewoods has quit [*.net *.split]
braewoods has joined #crux-devel
AndroUser has quit [Ping timeout: 260 seconds]
AndroUser has joined #crux-devel
AndroUser has quit [Ping timeout: 245 seconds]
groovy2shoes has quit [Remote host closed the connection]
groovy2shoes has joined #crux-devel
braewoods has quit [Quit: WeeChat 2.8]
braewoods has joined #crux-devel
jue has joined #crux-devel
jue has joined #crux-devel
<jue> beerman: the xorg-server update is a bit broken: xorg-server depends on libxcvt but you cannot install it clean cause you get file conflicts with cvt resp cvt.1 from old xorg-server, thus a forced install is needed
<beerman> mh, i have had that double checked from romster. true about the file conflict, but that shouldn't be a big issue, is it? it should work
<jue> well, with a manual forced install, a simple prt-get install libxcvt will not work
<jue> not a big issue but IMO worth a notify
<beerman> there was a notify attached to it..?
<beerman> https://crux.nu/gitweb/?p=ports/xorg.git;a=commit;h=e4dc03c70f8f5394a58f9dd6156dd9b5b4d45b6f
<jue> not to libxcvt
<jue> and not to the forced install issue
<beerman> yep, ok, noted for next time something like this occurs.
<beerman> but i had these changes double checked by romster before i commited them
<jue> btw, I got the +s for Xorg.wrap as well
<beerman> as well?
<jue> frinnst and elderk had that too
<beerman> i had no footprint mismatches on three machines myself, can't say for which tests romster did but i guess container then desktop
<beerman> guess we could set it in the Pkgfile to make sure it stays like that then, although it would make sense to check if xorg-server can make use of seatd to avoid the sticky bit
<jue> well, not sure, but I think the sticky bit should be set for Xorg.wrap
<jue> you commentef the link from Xorg->X, is that intentional?
<beerman> the meson build created that automatically now
<jue> ok, fine
<beerman> well, anyway, would you be ok to test if seatd works with xorg-server as well? I don't have an X11 desktop for months, but having one sticky bit less installed should be considered a good thing
<jue> sorry, don't us seatd, but will try to do a test later
<beerman> spooky, now i get the sticky bit on my desktop after rebuilding it ~10 times to make sure
jue has quit [Read error: Connection reset by peer]
<beerman> uhm, bye then..
jue has joined #crux-devel
jue has joined #crux-devel
jue has quit [Changing host]
<jue> beerman: as expected, got a Xorg freeze without the sticky bit
<beerman> ah, ok
<beerman> i just rebuilt on my desktop and got the sticky bit too, after who knows how many rebuilds
<beerman> will have that sorted soon then
<jue> btw, it's neccessary to rebuid all input and video drivers
<beerman> really? O_o
<beerman> Romster: ping
<beerman> i haven't had to do that with wayland, and there you still use xorg input stuff
<jue> that's a least what xorg.log says to me
<Romster> yeah reading it and it wants root for KMS stuff else it drops root https://man.archlinux.org/man/Xorg.wrap.1
<jue> for example -> (EE) vesa: module ABI major version (24) doesn't match the server's version (25)
<Romster> surely we can use setcap in a post-install to do this better than the arcane suid bit
<beerman> so much for testing beforehand :D intentions were good, execution was leaving things to be desired rofl
<Romster> i mean it drops root privilege for KMS
<jue> the version mismatch thing is the usual behaviour with new xorg-server versions
<Romster> suid stuff is weird it wont get set in lxc containers but it does on my system
<Romster> i didn't get that on the current nvidia driver in crux
<Romster> now we need to find if other video drivers are ok with the ABI change.
<jue> well, it should work for the normal user, a machine freeze after an xorg-server update isn't really nice ;)
<Romster> mine never froze :/
<Romster> i'
<Romster> i'm on it now since last night
<Romster> total lockup or just X?
<jue> total
<Romster> 21:49:00 up 17 days, 5:08, 1 user, load average: 1.38, 1.52, 1.95
<jue> btw, we had the same issue with the update to meson for xorg-server lately
<Romster> well damn 2 machines i've bumped here and they are working.
<Romster> question is do we revert or look for a fix or both
<Romster> video drivers not caught up? I did read the announcement i didn't see an ABI change mentioned.
<jue> IMO it cannot work without the +s bit for Xorg.wrap
<beerman> please don't revert.. we can just set the sticky bit and look for a cleaner solution in build
<jue> +1
<Romster> set the sticky bit but does that fix the freezing?
<beerman> yes..
<jue> sure
<Romster> easy fix then.
<beerman> reverting makes my toe nails roll up, fyi
<Romster> well i was just reading here and was like freezing the entire machine. what have we done.
<beerman> well, +1 point then for why wayland is much better than xorg :)
<Romster> well we got the xwayland compatibility for those few apps that haven't got wayland yet
<beerman> which still needs the xorg-server package but yeah, its mainly just steam, some java crap, and ancient gtk stuff nobody should care about anymore
<jue> there are some old patches in xorg-server ...
<beerman> i was planing to use prtwash when i tag the 3.7 branches
<beerman> to clean out all things left over, but we could do that maybe monthly and automated too..?
<jue> has you seen frinnst' report about failing ati driver?
<Romster> i am doing a full rebuild of all xorg ports right now
<Romster> yes i saw that
<jue> as a general rule: if you need me as a tester for critical stuff, don't hesitate to ask me ;) I'm always reachable via email at least
<Romster> ok, i just tested beerman's x11 side of it while beerman did the wayland side.
<beerman> yep, we tried to get every use case but clearly something went wrong
<Romster> well we learned to get some more testers and to also check if it breaks anything on recompile too, the later we both should of checked sooner.
<beerman> about the log: no i haven't, i don't follow #crux for a reason
<jue> I know, that's why I pasted the link to the log here :)
<beerman> yep. ok, well, maybe frinnst needs to change up the tactics to report things like that?
<Romster> i do follow it and i quoted that to beerman, i'm onto it but testing all the xorg incase of more breakage
<beerman> so is somebody already on his way to setup the sticky bit? else i will do it now
<jue> no
<beerman> ok, on the way
<beerman> jue: guess you didn't test wayland yet? :)
<jue> no, sorry, I'm really short in time at the moment, hope that will change later this year
<beerman> thumbs pressed that you get everything sorted, whatever is taking your time
<beerman> no worries, you will just love it and never look back once you get to it❤️
<beerman> funny, explicitly setting the sticky bit doesn't show in my lxc footprint.. <_<
<beerman> pushed xorg-server, it wouldn't take chmod u+s for some reason, but 4755 worked just fine.. no idea whats up with that
<beerman> and actually should be 4555, i need another ☕️
<beerman> i'll just update and bump the release number, so that everybody gets to rebuild it with the suid bit explicitly set and the correct permissions on the file..
<beerman> jue: btw, mozjs78 is currently broken for me with rust 1.56 because of an incompatible crate, also the url changed: https://spidermonkey.dev/
<beerman> a temporarily downgrade to 1.55 fixes that, and firefox had the same problem that romster and me patched, maybe that would be possible with mozjs78 as well
<jue> ahh, thanks, will look
<beerman> be wary, cargo doesn't like it when you change local crates, gentoo has a function inside their firefox ebuild that will nuke that hashsum file completely..
farkuhar has joined #crux-devel
<beerman> https://github.com/sepen/crux-ports-sepen look who is getting back in business (a little)
<jue> wrt the freetype update -> https://dpaste.com/8ENL359JJ
<jue> stumbled over that while trying to update imagemagick
<beerman> let me guess, revdep didn't report it?
<jue> no
<jue> it's a problem of the missisng la file
<beerman> yeah i know.. should've included that one or the other one liner, i have that saved as an shell alias
<beerman> didn't think twice yesterday, i have been running this for months now, just added the backpatch for chromium
<beerman> i build ungoogled-chromium and electron, and it affects qt6-webengine as well as soon they reach m94 chromium sources
<beerman> qt5-webengine gets still free patches from the kde project, but it would be a great move to ditch qt5 as soon as possible... the situation is gruesome, only paying customers get official updates to the qt5 branch anymore
<jue> it's a bit of a pain, normal user will probably not be able to solve the freetype issue without some help
<beerman> it probably should be in the wiki/handbook
<beerman> would make it easier to link to a single webpage, or manpage if we so desire, instead of everytime something like this happens having to write a novel about it
<jue> well, we should think twice if we do such changes, I understand that you're goal is to avoid autotools as much as possible, but in that case it wasn't really necessary
<beerman> to be honest my goal was to get rid of the dups from my overlay, getting rid of autotools was just a goodie that i didn't really think about anymore
<jue> :)
<beerman> granted, maybe not super necessary, but it was a good thing for me, and that needs to count for something too, every once in a while at least ;) chromium, webengines, electron, people might cringe (lol) but i do need these in my daily workflows so yeah
<beerman> and with these patches (e.g. subpixel hinting) everybody gets to enjoy nicer fonts (if they configure fontconfig correctly)
<beerman> nicest blog post i read was titled "i stared into fontconfig and it stared right back at me"... man i wish we were further with that already, but good fonts on linux takes a few extra steps
ivandi has quit [Ping timeout: 260 seconds]
ivandi has joined #crux-devel
<Romster> fixed up xorg issues
<beerman> 🎉
<Romster> i can take a look at patching mozjs78
ivandi has quit [Ping timeout: 268 seconds]
ivandi has joined #crux-devel
jue has quit [Ping timeout: 260 seconds]
jue has joined #crux-devel
jue has joined #crux-devel
jue has quit [Changing host]
prologic has quit [Quit: ZNC - https://znc.in]
prologic has joined #crux-devel
jue has quit [Ping timeout: 268 seconds]
groovy2shoes has quit [Remote host closed the connection]
groovy2shoes has joined #crux-devel
<ivandi> beerman: Thanks for xwayland. The pixman dependency has to changed to xorg-libpixman or removed, xorg-server already depends on xorg-libpixman. There is no pixman port.
<beerman> ivandi: good catch