01:41 <jaeger> oops, I tagged the wrong person there, sorry

01:42 <jaeger> Reading comprehension was hard today :)

07:11 <beerman> Mh maybe I messed up :) if it works fine please ignore my report about it

07:20 <beerman> arch includes the file as well, and links this issue about it https://github.com/golang/go/issues/57179

07:21 <beerman> for me docker-compose failed to build without it, but I have a fork of the go port in my overlay which can compile with system installed go instead of the prebuild one

07:21 <beerman> so maybe its one of my changes triggering it, I have no idea :)

09:36 <beerman> jaeger do we still need pangox-compat? nothing depends on it directly, and pangox-compat depends on pango-compat which is a danny port i'd mark unmaintained

09:54 <beerman> jaeger vte3 is moved to contrib as well, and needs adoption for virt-manager

09:55 <beerman> vkd3d is needed for vkd3d-32 which is a dependency for both wine packages. i moved dannys wine to contrib and marked it unmaintained as well

10:00 <beerman> 55 ports remain to be looked through in opt. new ports in contrib that are in dannys name and i didn't mention are dependencies somebody should pick up, which are only vte3 and openal

14:51 <farkuhar> beerman: openal can be moved to the optional dependencies of gst-plugins-bad, if i'm reading correctly https://gitlab.freedesktop.org/gstreamer/gstreamer/-/blob/main/subprojects/gst-plugins-bad/meson_options.txt

14:52 <farkuhar> and if nobody picks up the unmaintained freealut, allegro, or sfml, then openal can be marked unmaintained too.

14:55 <farkuhar> somewhat related ... i looked at the vorbis-tools port and found that it's easy to make libao an optional dependency.

15:37 <jaeger> I can add the go file if needed, just didn't run into the same problem building here

15:51 <beerman> jaeger: go is just weird:)

15:52 <beerman> farkuhar: I'll check that out later/tomorrow. I'm out for the night

15:57 <jaeger> Agreed

17:15 <jaeger> I'm not aware of anything specific needing pangox-compat. I guess if I find something later I can always bring it back

17:23 <jaeger> Interesting... getting this from openssh on the ISO builder:

17:23 <jaeger> configure: error: *** zlib too old - check config.log ***

17:23 <jaeger> zlib is updated to 1.3

17:25 <jaeger> It also says to upgrade zlib to 1.2.3 or greater, heh

17:25 <jaeger> Probably something dumb like a failing string check

17:29 <jaeger> Yeah... stupid problem caused by the version string getting shorter

17:30 <jaeger> A quick hack of their check in configure shows:

17:30 <jaeger> ZLIB_VERSION = 1.3

17:30 <jaeger> found zlib version 1.3 (1030000)

17:30 <jaeger> n = 2

17:30 <jaeger> The check does an exit(1) if (n != 3 && n != 4)

17:30 <jaeger> n = sscanf(ZLIB_VERSION, "%d.%d.%d.%d", &a, &b, &c, &d);

17:35 <jaeger> Here's an admittedly naive quick fix: http://ix.io/4DYE

17:45 <jaeger> Back to contrib/go, go.env already exists in its footprint

17:46 <jaeger> -rw-r--r-- root/root usr/lib/go/go.env

17:47 <jaeger> Handled by 'mv $SRC/go $PKG/usr/lib/' automagically, I guess. Would be interesting to find why it's missing from your build, perhaps that overlaid version after all

18:06 <farkuhar> the way I'm reading beerman's report, his overlaid port checks for an existing go installation, and prefers that one instead of the upstream precompiled tarball (hence redefining GOROOT_BOOTSTRAP).

18:06 <jaeger> yeah

18:06 <jaeger> That was my take as well

18:08 <farkuhar> but it still puzzles my why he didn't get a footprint mismatch, if redefining GOROOT_BOOTSTRAP has the side-effect of not putting usr/lib/go/go.env into the built package.

18:09 <beerman> Your assumptions are right :) i can share what i use if it's of any interest here

18:09 <beerman> Would be nice to merge that maybe, i dunno

18:16 <jaeger> Is the overlaid port's .footprint indeed the same as contrib/go's?

18:19 <beerman> That I'd need to check when I'm back home

18:21 <jaeger> ok

18:21 <farkuhar> I guess if usr/lib/go/go.env is not in the overlaid port's .footprint, then merging it would require that we deploy PKGMK_IGNORE_NEW=yes, so that a first-time installation does not fail.

18:44 <beerman> Not home yet, but here you go for the Pkgfile https://dpaste.org/BeJNW