[prev in list] [next in list] [prev in thread] [next in thread]
List: netbsd-port-sparc
Subject: Re: NetBSD 8.2 Package refresh fails
From: Greg Troxel <gdt () lexort ! com>
Date: 2020-06-26 17:45:43
Message-ID: rmi5zbd67qw.fsf () s1 ! lexort ! com
[Download RAW message or body]
Riccardo Mottola <riccardo.mottola@libero.it> writes:
>> lintpkgsrc -i
> Scan Makefiles: 16777 packages
> Version mismatch: 'pango' 1.44.7 vs 1.44.7nb2
> Version mismatch: 'windowmaker' 0.95.8nb6 vs 0.95.8nb8
That seems minor.
> this, together with the fact that no installed package has BUILD_HOST
> 7.0 appears to tell me I have a consistent system.
> So, the mystery of pango vs. hb.h header remains.
Sounds like you have a consistent build; if the binaries were from a
different branch the diffs would be huge.
> I tried something rough. I went into
>
> /home/pkg-workdir/devel/pango/work/.buildlink
>
> did a more ./include/harfbuzz/hb.h
>
> I changed in=C2=A0 pango-coverage.h
>
> #include <hb.h>
>
> to #include <harfbuzz/hb.h>
>
> then I had to do the same in pango-font.h and in pangofc-fontmap.h
>
> I wonder if it is a code bug or more an issue with includes.
>
> That way it compiles for a while, then dies in:
>
> /home/pkg-workdir/devel/pango/work/pango-1.44.7/pango/pangocairo.h:26:19:
> fatal error: cairo.h: No such file or directory...
> Again I changed that to <cairo/cairo.h>
>
> after that... it compiled!
there are certainly a lot of bugs out there.
Basically, a package will install headers where that package says is the
right place, and packages that use it have to cope.
What I don't understand is why this breaks sometimes but not usually.
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic