Re: FireFox91

["David H. Gutteridge" <david%gutteridge.ca@localhost>]

On Wed, 2 Nov 2022 at 07:21:18 +0000, Todd Gruhn wrote:
> I did what Benny said -- FireFox91 finished -- IT WORKS. Thanks
>
> But
>
> _REENTRANT -D_REENTRANT -D_REENTRANT -D_REENTRANT -D_REENTRANT
> -D_REENTRANT -D_REENTRANT -D_REENTRANT -D_REENTRANT -D_REENTRANT
> -D_REENTRANT -pthread
> -I/usr/pkgsrc/www/firefox91/work/build/dist/include/cairo
> -Werror=switch  -MD -MP -MF .deps/Unified_cpp_webrender_bindings1.o.pp
>  Unified_cpp_webrender_bindings1.cpp
> In file included from Unified_cpp_webrender_bindings1.cpp:11:
> /usr/pkgsrc/www/firefox91/work/firefox-91.13.0/gfx/webrender_bindings/WebRenderAPI.cpp:
> In member function 'mozilla::wr::WrClipChainId
> mozilla::wr::DisplayListBuilder::DefineClipChain(const
> nsTArray<mozilla::wr::WrClipId>&, bool)':
> /usr/pkgsrc/www/firefox91/work/firefox-91.13.0/gfx/webrender_bindings/WebRenderAPI.cpp:1052:51:
> error: 'ROOT_CLIP_CHAIN' is not a member of 'mozilla::wr'
> 1052 |       mCurrentSpaceAndClipChain.clip_chain != 
> wr::ROOT_CLIP_CHAIN) {
[...]
> FireFox91 is not dead yet -- how much does this matter ??

This is simply a bug in the packaging. When cbindgen was updated to
0.24.x, its code generation changed significantly here. This meant the
old assumption in the Firefox code about cbindgen <= 0.23 generation
was no longer valid. Now we're carrying a patch that will only work
with cbindgen >= 0.24. That needs to be reflected in the package as a
minimum dependency for anyone who isn't rebuilding packages en masse,
so they are explicitly forced to have the newer version of cbindgen.
I've now done so:
http://mail-index.netbsd.org/pkgsrc-changes/2022/11/02/msg262896.html

Regards,

Dave