Re: meson advice on CFLAGS et al.

Greg Troxel <[email protected]> Wed, 22 Jul 2026 06:33:40 -0400
Newsgroups gmane.os.netbsd.devel.packages
Message-ID <[email protected]>
"Dr. Thomas Orgis" <[email protected]> writes:

> 1. Would be nice if pkgsrc worked out what the breakage is so that we
>    do not need to resort to the meson-specific move of CFLAGS etal.
>    into command-line arguments. Would mean less code in meson/build.mk.
>    Less special-casing =E2=80=A6
>
> 2. I understand that if it works now and is likely to be robust in the
>    future, that adding those args to meson/build.mk may save
>    develociraptor time in the future.
>
> But if there is a chance to work out what actually fails in
> graphics/geeqie with passing flags via environment and fix it, possibly
> inside meson, I'd prefer that =E2=80=94 but again, I understand if we don=
't
> have the personal time for that. It feels like a build system bug if
> one has to make the flags specific without any special setup justifying
> it.


Makes sense.  I think the problem is that meson sees meson-is-special as
a feature, rather than the bug that it is.  But if $SOMEONE wants to
file a bug and/or fix it, that's great.


My take is that adding the meson flags is easy, and suppressing CFLAGS
etc. is harder and also problematic (does a build use something else
too?), and thus the addition is ok for now.

I see your point that it deserves a "work around meson brokenness"
comment.