12:00 <r0ni> can anyone explain what "prt-get update -fr port $(prt-get quickdep other-port)" is doing exactly?

12:01 <r0ni> i'm assuming all deps in each port is being rebuilt against each other ?

12:03 <r0ni> was having an issue upgrading python 3.12 on 3.8, so I did a "prt-get update -fr python3-setuptools $(prt-get quickdep python3) and well things are building, but will it work is the question lol

12:08 <farkuhar> r0ni: From what I recall of the official prt-get, the "update" command does not sort the targets by dependencies, but instead builds them in the order they appear on the command line. But the substitution $(prt-get quickdep python3) guarantees that the latter targets are all sorted by dependency, so they'll be rebuilt in that order, after python3-setuptools is rebuilt.

12:10 <farkuhar> What the official prt-get does not yet do is customize pkgadd flags for each built package, so the "update" command will run "pkgadd -u" on every built package, even if it's a new dependency added since the previous version. If that dependency had not been installed before, then pkgadd -u will fail.

12:12 <farkuhar> Not likely to be an issue for python3, but something to keep in mind if you try adapting "prt-get update -fr port $(prt-get quickdep other-port)" to other targets.

13:03 <r0ni> well while python3 builds no modules will. even trying to build setuptools complains its missing "distutils" so I assume my python3 package has missed a key element

13:04 <r0ni> but i think I understand a little better on what that does now, thx farkuhar

13:11 <r0ni> ahh well i see a flag set in the port is depreciated as of 3.12 so maybe I'll begin there

14:01 <farkuhar> Fun opened FS#1576 to address the footprint mismatches caused by the presence of optional dependencies, but his idea could also be repurposed to reduce the number of overlays needed to accommodate different architectures, or different versions of python (3.10 versus 3.12). Fun took a more convoluted approach than PR#2 (https://git.crux.nu/tools/pkgutils/pulls/2) but maybe that level of sophistication is needed to avoid false positives.

14:09 <r0ni> i wonder if my woes will be fixed by just adding a build triplet to the pkgfile, lets hope it was really that simple

14:12 <r0ni> looks promising. oh well must sign off for today, more 3.8 will have to wait