jaeger changed the topic of #crux-arm to: CRUX-ARM 3.7 Released! - http://crux-arm.nu/Documentation/ReleaseNotes3-7 | Logs: https://libera.irclog.whitequark.org/crux-arm/
pitill0 has quit [Quit: leaving]
<pitillo> fakeroot 1.35 seems to be borked on ARM
<pitillo> is someone using fakeroot to build ports on X86?
<pitillo> s/X86/X86-64b
<beerman> not me, sorry
<jaeger> I use it to build everything
<jaeger> I haven't built anything on arm recently, though
<pitillo> I thought there were problems on arm64 but I can reproduce it on 32b too
<pitillo> I need to revert to 1.34 or faked freezes
<jaeger> I'll give it a test here as well
<jaeger> Hrmm, it seems ok on my rpi5
<jaeger> Running a sysup to test it a bit
<pitillo> here I've confirmed on the rpi3, pine64 and the odroidxu4... all with the same behavior
<jaeger> When does it freeze? As soon as the build starts?
<pitillo> yes, just before downloading (if I'm right, when fakeroot execs faked)
<jaeger> And how are you invoking it? From prt-get.conf?
<pitillo> yes
<jaeger> OK, same here. No freezes so far.
<pitillo> makecommand sudo -H -u pkgmk fakeroot pkgmk
<jaeger> oh, different invocation
<pitillo> no idea then
<jaeger> makecommand fakeroot pkgmk
<jaeger> I build as my own user (jaeger) rather than a dedicated pkgmk user
<pitillo> ummmm here I'm using pkgmk user
<pitillo> interesting
<jaeger> Maybe try logging in as that pkgmk user and running fakeroot directly? See if the sudo call is interfering somehow
<pitillo> I'll do and check
<jaeger> For reference:
<jaeger> makecommand fakeroot pkgmk
<jaeger> addcommand sudo pkgadd
<jaeger> runscriptcommand sudo sh
<jaeger> removecommand sudo pkgrm
<pitillo> here I'm only using the makecommand
<pitillo> with any user, fakeroot freezes
<pitillo> just invoking fakeroot, stops
<jaeger> weird, I cannot currently reproduce it
<jaeger> I will test again after the sysup, maybe some other library is causing it
<jaeger> Not that faked is linked against much of anything, just glibc it seems
<beerman> when using sudo there like that, wouldn't it prompt between every other step after the timeout for the password over and over again?
<jaeger> Does any fakeroot call work, like 'fakeroot ls'?
<jaeger> yes, unless sudo is configured to allow commands without password
<beerman> okay, sure. I think I was just puzzled because I never thought about putting sudo/doas in there, I just have passwordless ports configured, via doas.conf
<beerman> btw, I found some time to dig further into mesa 24.1 (on x86_64 though), these are my changes so far: mesa https://nullvoid.de/rpaste/bRInCbul.txt mesa-32 https://nullvoid.de/rpaste/TIxP96R4.txt
<beerman> still testing to make sure it works good, but seems so
<jaeger> pitillo: ok, after sysup finished, still fakeroot 1.35 works as expected here.
<pitillo> I'm thinking about that could be something related to old kernels, as far as it only depends on glibc
<jaeger> 6.1.70-v8-16k+ is what I'm on here
<pitillo> 4.19.79-v8+ on the rpi3... pine64 and odroid are 3.x if I'm right
<r0ni> yall making me wonder if my issues with the rootfs script were fakeroot related now... i've no time to check right now, but i know i was using fakeroot when i had failures