17:21 <farkuhar> beerman: here's the fork that I've tested successfully against your original scenario (ffmpeg and vulkan stuff both needing updates, and it's desired that vulkan get updated first). https://git.crux.nu:82/farkuhar/prt-get

17:22 <beerman> farkuhar: neat, that i will need to check out

17:23 <farkuhar> this fork currently does not use soft dependency resolution by default, you have to opt in by adding the line "softdeps yes" to prt-get.conf. Maybe later I will add support for a command-line switch.

17:33 <farkuhar> This fork introduces a breaking change to the cache file format, so don't run it as prt-cache! The out-of-fashion 'Packager' field has been repurposed to hold the 'Optional' metadata. In this I was guided by the cautionary words of Fun, who lamented in FS#1576 that CRUX had started to go down the path of "adding complexity but removing none".

17:52 <beerman> heh, noted

18:15 <beerman> jaeger: https://crux.nu/gitweb/?p=ports/opt.git;a=commitdiff;h=bf8127b7d9006a43443524560f1478346a029fc8 you were right, the script to clean the footprint is not needed anymore, at least my tests so far looked like it, rebuilt it a couple of times this time

18:16 <beerman> i took the liberty to note down "CRUX System Team, core-ports at crux dot nu" as the maintainer, same as llvm and friends

18:16 <beerman> which it basically is, a llvm friend

19:06 <beerman> jue: https://dpaste.org/vaSam i am getting a footprint mismatch on imlib2 in the container, is this just me?

19:10 <beerman> thats what I get on my desktop: https://dpaste.org/vaMQW

19:11 <beerman> this is where a patch like the one farkuhar just posted might come in handy as well

21:41 <jaeger> ok, good to know

21:41 <farkuhar> to be clear, the prt-get fork I posted today only uses the 'Optional:' metadata for install, update, sysup, and info. If there's any interest in making 'depends' or 'dependent' aware of optional dependencies in the same way, I could work on those subcommands next.

21:47 <farkuhar> s/install, // (I've been working too long with my perl-based rewrite, which operates in mixed install/update mode.)

22:38 <beerman> farkuhar: i guess the amount of energy spent on that depends on if jue and jaeger would agree into merging that, or what it would need to be merged. I generally do appreciate the effort you put into that

23:52 <jaeger> I would say that if you implement partial optional support, might as well go all the way

23:59 <farkuhar> jaeger: thanks for the input. I'll go ahead and extend the 'depends' and 'dependent' subcommands in the forked repo.