<Romster>
PSA: DO NOT UPGRADE TO UBUNTU 21.10 IF YOU USE ZFS. DO NOT install Ubuntu 21.10 with ZFS. It *will* corrupt your filesystem irreparably. There’s a warning in the release notes – but only after 2800 words >_> https://bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1906476
tilman has joined #crux
<SiFuh>
Wouldn't EVER install Ubuntu
rlittl01 has quit [Ping timeout: 260 seconds]
Belgacem1958 has joined #crux
<Belgacem1958>
Hello with an i7 and 8GB of ram, my pc freezes when compiling qtwebengine. Besides, this compilation has never succeeded for me (failure). What would be the origin of this gel in my configuration?
<SiFuh>
What's the temperature like during the build?
<Belgacem1958>
I feel that the pc heats up abnormally during long compilations.
<SiFuh>
I gave you the URL to a pre-built version
<SiFuh>
Just in case
<Belgacem1958>
ce même package pose des problèmes (temps de compilation) aux utilisateurs de gentoo. C'est d'ailleurs lui qui m'a forcé à abandonner gentoo. Un binaire pour Crux serait une bonne idée.
<Belgacem1958>
Thank you SiFuh : just use pkgmk -i to install it?
<Belgacem1958>
Sorry for the french message. this same package causes problems (compilation time) for gentoo users. He was the one who forced me to give up gentoo. A binary for Crux would be a good idea.
belgacem195846 has joined #crux
<SiFuh>
Belgacem1958: pkgadd <filename>
<belgacem195846>
SiFuh : Thank you, all well done
<Romster>
it uses a lot of resources to compile. perhaps your cpu cooler is cloged and/or thermal paste needs redoing
<Romster>
you can also change the number of JOBS in /etc/pkgmk.conf
<belgacem195846>
Romster : I have this in my file : # export JOBS=$(nproc)
Belgacem1958 has quit [Remote host closed the connection]
belgacem195846 has quit [Quit: Client closed]
Belgacem1958 has joined #crux
Belgacem1958 has quit [Client Quit]
<dlcusa>
SiFuh, re: video, where's the video to expose the not-so-well-educated on the right? You know it will be at least as disheartening.
<dlcusa>
Still, a signature's a signature, and a vote's a vote (when it's really a vote, of course, and I'm sure they all are in the USA, at least--not that I could prove it, of course).
<SiFuh>
dlcusa: He's in California. They are all right, California is actually a red state :-P
pedja has joined #crux
GazL has joined #crux
<GazL>
Hey folks. I wanted to make some bug reports, but I can't see an obvious way to signup for your flyspray bug-tracker. How would I go about that? Or would you prefer I email someone?
<SiFuh>
GazL: no worries, I also never signed up for it.
<SiFuh>
If you have a bug just send here
<SiFuh>
Someone will see and assist you for the flyspray thing when they see it
<GazL>
Fair enough.
<SiFuh>
GazL: what's the issue?
<GazL>
first one, really trivial: mkfonscale and mkfondir are the wrong way around in rc.fix. fontscale is used as input fo fontdir.
farkuhar has joined #crux
<GazL>
Next one, The new lvm package has a reference to systemd-run in /lib/udev/rules.d/69-dm-lvm.rules, which we obviously don't have.
<GazL>
I have a couple more, but they fall into suggestion/wishlist sort of area.
<SiFuh>
Because scalable font files do not usually include the X font name, the file "fonts.scale" can be used to name the scalable fonts in the directory. The fonts listed in it are copied to fonts.dir by mkfontdir. "fonts.scale" has the same format as the "fonts.dir" file, and can be created with the mkfontscale(1) program.
<SiFuh>
Oooh cool, learn somethign new
<GazL>
:)
<SiFuh>
jue is not here for now but the lvm2 port is his
<GazL>
Okeydoke. I'll see if he picks it up from this log, if not I can always drop him an email.
<SiFuh>
Yeah I do that some times
<GazL>
I'm not sure whether this next one counts as a bug or not, but it's something I don't like, so I'll mention it anyway... ;)
<SiFuh>
Just mention everything
<SiFuh>
Pretty sociable channel here
<GazL>
:) yeah, I noticed that. :)
<GazL>
Ok, here it comes...
<SiFuh>
Yeah ignore of the flood of hugs and oves and gossip
<GazL>
The Perl Pkgfile uses the same directory for sitelib and vendorlib, which means that @INC has a duplicate entry.
<GazL>
I know you don't use vendor_lib, but IMO it would be cleaner to have an empty dir than a duplicate entry in the search list
<SiFuh>
GazL: that one is for #crux-devel
<SiFuh>
But they will see it here
<GazL>
I thought #crux-devel was for private dev chat.
<SiFuh>
In a way yes
<SiFuh>
But you can still join and mention 'core' based stuff
<SiFuh>
Most of the dev team is here anyways and they do see everything
* GazL
nods
rlittl01 has joined #crux
<GazL>
Ok, last one is more of a request. I use groups heavily, includeing group passwords. I had to rebuild shadow without the --disable-shadowgrp and reincluding gpasswd to get the features I needed.
<GazL>
I don't mind doing that, but it'd be nice if I didn't have to. :)
<SiFuh>
That'll be a #crux-devel thing too
<SiFuh>
Understand though. I had somethings I needed in just the default ISO's kernel and fortunately jaeger was kind enough to intrduce it
<SiFuh>
There may be a reason that it was disabled, or it may have just been disabled long ago and never considered again
<GazL>
I suspect they just decided it was little used and unneeded, but gpasswd is a pretty standard way of adding/removing users to/from a group, so it's IMO valuable.
<GazL>
admittedly, group passwords aren't used all that much.
<GazL>
Belgacem1958: take a look at the manpage for prtwash. It may have an option to suit your requirements.
<GazL>
speaking of which... for the CRUX devs, a prtsweep -a -t seems to identify a lot of patch files that are no longer used in the opt collection but are still in git. Might be worth a cleanup at some point.
<GazL>
correction prtwash, not sweep.
<Belgacem1958>
Gazl : good, thank you for this help
<cruxbot>
[contrib.git/3.6]: fcft: 2.4.5 -> 2.4.6
<farkuhar>
Belgacem1958: if you're using a common directory for saving pkgmk downloads and doing the work of compilation, then prtwash and prtsweep in their original form are not as fast as a custom script that looks for old files in $PKGMK_SOURCE_DIR and $PKGMK_PKG_DIR.
<farkuhar>
But if your builds are just writing to /usr/ports (which is the default), then prtwash and prtsweep will satisfy your needs just fine.
<Belgacem1958>
Farkuhar : i'm buiding in /home/pkgmk/
<Belgacem1958>
Manually deleting is safe?
<farkuhar>
Building in /home/pkgmk is a sane choice. If /usr/ports is not being cluttered up by building your packages, then prtwash and prtsweep won't be much use. Just go through /home/pkgmk and delete manually what you don't want to keep, that should be safe.
Moth has quit [Remote host closed the connection]
Moth has joined #crux
<Belgacem1958>
Farkuhar : greatly appreciate your kindness and your precious help.
<farkuhar>
There is a rewrite of the prtwash and prtsweep scripts that is better suited to your (rather common) setup, and these revised scripts also respect the new pkgmk "renaming tarballs" feature. Check them out at http://sdf.org/~jquah/cruxports/prt-utils/
<GazL>
farkuhar: what's the "renaming tarballs" feature? I've not come across that yet.
belgacem195893 has joined #crux
<farkuhar>
It's not yet available in CRUX 3.6.1, but it was on the TODO list for 3.7. jue committed an updated pkgutils source a few weeks ago.
<GazL>
I wanted my sources to go in individual subderctories so used PKGMK_SOURCE_DIR=/var/cache/pkgmk/$name but the problem was that it didn't like it when there were missing directories, so I had to patch pkgmk and I added a PKGMK_CREATE_MISSIND_DIR option to enable it.
<GazL>
Ahh, I see. thanks.
<farkuhar>
GazL: you might want to check out Flyspray issue 923, which has a more complete history of the various proposals to deal with naming and saving downloads.
<GazL>
thankyou. Will do.
<GazL>
Ahh, yes, I've hit that issue before. Looks like a useful feature.
<GazL>
Ok read that, but finding it hard to follow which of the suggested approaches was chosen in the end. Did you end up going with the associative array idea?
<GazL>
... or the renames=() approach?
belgacem195893 has quit [Quit: Client closed]
CrashTestDummy2 has joined #crux
CrashTestDummy3 has quit [Ping timeout: 260 seconds]
AndroUser has joined #crux
Belgacem1958 has quit [Read error: Connection reset by peer]
<farkuhar>
GazL: the updated source that jue committed was based on the renames=() array, one of the earlier counterproposals from Romster.
_0bitcount has joined #crux
AndroUser has quit [Remote host closed the connection]
AndroUser has joined #crux
<GazL>
Ok, thanks. I think I prefer that one, as it seems much simpler.
AndroUser has quit [Ping timeout: 265 seconds]
bdax has joined #crux
ivandi has quit [Quit: WeeChat 3.3]
ivandi has joined #crux
ivandi has quit [Quit: WeeChat 3.3]
<stenur>
fwiw i have a very specific port cleanup thing; i have separated source and package dirs; the second part that goes via "prt-get listinst" to find stale directories may be of interest, but i doubt it; http://ix.io/3CBA
<stenur>
just today separated my port-up.sh script to first download, by passing --margs=-do in a first pass; so ugly! what a mess! but only the output of course, other than that just works.
<stenur>
but 'am too lazy to drive pkgmk -d in the first pass myself
<stenur>
That port-trim.sh interestingly takes many seconds to run, even on a 4 processor 8 GB RAM with 1.5 GB/sec NVME SSD and /bin/sh
<stenur>
Now BTRFS subvolume(s) on Luks2, but the latter really is no thing. I can scrub at ~1.5 GB/sec!! Oh i love that box.
<stenur>
A nice Sunday everybody!
ivandi has joined #crux
bdax has quit [Quit: bdax]
<GazL>
Goodnight all. Have fun.
GazL has left #crux [exit stage left.]
ivandi has quit [Quit: WeeChat 3.3]
ivandi has joined #crux
_0bitcount has quit [Quit: Leaving]
Belgacem1958 has joined #crux
Belgacem1958 has quit [Quit: Client closed]
lisardman has joined #crux
Belgacem1958 has joined #crux
Belgacem1958 has quit [Client Quit]
farkuhar has left #crux [Abendessen]
crash_ has quit [Remote host closed the connection]