00:21 <farkuhar> jaeger: between prtwash and prtsweep, only prtwash makes any distinction between the current version of a package and the older versions. Does your automation rely on prtwash, or does it use `oldfiles -p` instead?

04:43 <jaeger> I've used both but currently prtwash

05:11 <jaeger> prtwash is failing, though

05:11 <jaeger> ... keeping file ports/core/libarchive/libarchive#3.7.2-1.pkg.tar.xz.

05:11 <jaeger> ... keeping file ports/core/libarchive/libarchive#3.7.2-2.pkg.tar.xz.

05:14 <jaeger> Some of its version comparison logic maybe isn't working like it should be

05:14 <jaeger> Also failing on firefox-bin 119.0-1 and 119.0.1-1 for example

05:14 <jaeger> Too tired to dig into that tonight, though

05:38 <cruxbot> [contrib.git/3.7]: wireshark: 4.0.10 -> 4.2.0; new dependency: speexdsp

10:36 <farkuhar> jaeger: try this one-line fix, http://ix.io/4LDr

12:36 <farkuhar> I can't remember why the Perl rewrite of prtwash opted for a more inclusive regex to identify the current version of a package (ignoring $release). The bash implementation always used both $version and $release. It's not as if maintainers only bump $release when they want to force a rebuild; sometimes there are non-trivial changes to the port's configuration.

16:55 <jaeger> I'll give it a look, replaced it with a quick bash script before sleeping last night

16:58 <Guest9824> SiFuh: im having issues getting cryptsetup decrypt partition inside custom initramfs.

16:58 <Guest9824> with sata drives it worked correctly on a thinkpad. setting it up on different hardware that uses nvme .. might be giving issues.

16:59 <Guest9824> the kernel is built with inbuilt CONFIG_BLK_DEV_NVME

17:00 <Guest9824> it sees the drives but .. something awful happens. been searching for hours now.

17:00 <Guest9824> '# Keyslot 0 priority 1 != 2 (required), skipped.' , '# Keyslot 0 (luks2) open failed with -22', 'No usable token is available', 'Keyslot open failed.'

17:03 <Guest9824> i am not sure if this has anything to do with argon2id pbkdf.

17:06 <SiFuh> Guest9824: "No usable token is available". You are using a key and it can't find it

17:06 <Guest9824> SiFuh: i am not using a key, i am using a password.

17:07 <Guest9824> from live system it can successfully decrypt the volume, but not from initramfs.

17:07 <Guest9824> so it must be something that is missing within the initramfs

17:07 <SiFuh> Guest9824: My sentence is missing a question mark.

17:07 <Guest9824> no, password is used. let me luksDump it and send output

17:09 <SiFuh> Guest9824: do you have lvm2 in your initramfs?

17:09 <Guest9824> no lvm2

17:09 <Guest9824> plain block device /dev/nvme0n1p2

17:09 <Guest9824> correct and wrong password return same debug and verbose information

17:14 <Guest9824> SiFuh: https://oshi.at/BsDpG

17:15 <Guest9824> this is what i have enabled in 6.1.62 kernel

17:15 <Guest9824> CONFIG_BLK_DEV_NVME, CONFIG_NVME_MULTIPATH, CONFIG_NVME_HWMON, CONFIG_CRYPTO_WP512, CONFIG_CRYPTO_XTS, DAX, CONFIG_BLK_DEV_DM, CONFIG_DM_CRYPT

17:21 <SiFuh> What is the configuration of your initramfs?

17:21 <Guest9824> moment, rewriting it now

17:30 <ukky> Guest9824: You can enable debug mode: cryptsetup luksOpen --debug <device> <mapper_name>

17:31 <Guest9824> ukky: the above errors are from debug mode

17:36 <Guest9824> SiFuh: https://oshi.at/FKmyw

17:36 <Guest9824> here is the initramfs and syslinux configuration

17:36 <ukky> You mention CONFIG_CRYPTO_XTS, but did not mention CONFIG_CRYPTO_AES

17:37 <Guest9824> ukky: correct. CONFIG_CRYPTO_AES is enabled by default

17:39 <Guest9824> .... aaaand my gpu died again

17:43 <ukky> And CONFIG_CRYPTO_USER_API_SKCIPHER?

17:46 <ukky> Guest9824: what does this command print?: ldd bin/{cryptsetup,mount,umount}

17:47 <SiFuh> I have never seen an APPEND line like that before for SYSLINUX

17:48 <Guest9824> ukky: you may be right about CONFIG_CRYPTO_USER_API_SKCIPHER, will build it as soon as gpu starts working again.

17:48 <Guest9824> ukky: cryptsetup and busybox print 'not a dynamic executable', they're both static and working from initramfs

17:48 <Guest9824> SiFuh: hmm?

17:50 <Guest9824> i see, this may be something from my previous lilo installation

17:50 <Guest9824> ukky: SiFuh: thank you both for helping me out, i'll retry soon with modified APPEND and CONFIG_CRYPTO_USER_API_SKCIPHER

17:50 <Guest9824> need to wait a bit so that gpu starts working..

17:50 <Guest9824> APPEND root=/dev/mapper/crypt cryptdevice=/dev/nvme0n1p2:crypt ro

17:50 <Guest9824> this should be better

17:52 <ukky> What init system do you use? Usually you need /sbin/init in switch_root command

17:52 <Guest9824> ukky: here is the init; https://oshi.at/FKmyw

17:52 <Guest9824> exec switch_root /chroot /init $(cat /proc/cmdline)

17:53 <SiFuh> APPEND root=/dev/mapper/crypt rw rd.auto=1

17:53 <ukky> I did not mean initramfs /init

17:53 <Guest9824> mind explaning?

17:55 <ukky> Kernel will start init process (PID1) after finishing. Usually, it is /sbin/init. But you can change its location via init=<full_path>

17:57 <ukky> Your 'exec switch_root /chroot /init' will start /init when kernel finishes loading. Do you have /init on your root (encrypted) device?

17:58 <Guest9824> no, i do not have /init on my encrypted root device.

17:58 <Guest9824> i usually kept init on initramfs only

17:58 <Guest9824> will note this down

17:58 <ukky> Then you have to change it /sbin/init

17:58 <Guest9824> ah... i see

17:59 <ukky> Your /init is just a shell script.

17:59 <ukky> /sbin/init is PID1 init process from SysVinit package

18:00 <Guest9824> i usually did use /sbin/init but.. not sure why i changed it. /sbin/init was a symlink to ../init

18:00 <Guest9824> thank you for explaining this.

18:00 <Guest9824> unable to work on it now, the gpu is dead; https://oshi.at/MgsK/Msuc.jpg

18:01 <ukky> No spare graphics card?

18:02 <Guest9824> sadly no :-(

18:04 <SiFuh> PCIe power cables not plugged in?

18:04 <Guest9824> they're plugged in. the gpu is dying. i have already revived it twice at 182 C degress with addition of flux for the chip.

18:05 <Guest9824> once revived, it works for 2-3 months and starts dying again.

18:05 <Guest9824> think i'll get a new one these days, tired of these issues

18:33 <SiFuh> https://alexjonesgame.com/ <-- Heh

21:26 <cruxbot> [contrib.git/3.7]: rust-bin: 1.73.0 -> 1.74.0

22:03 <jaeger> crash_: ok, got a new ISO uploaded