I've been trying to rebuild (the previously working) cross-build environments: i686-w64-mingw32 & x86_64-w64-mingw32 Hitting a brick wall with cross-i686-w64-mingw32/mingw64-runtime: checking _mingw_mac.h usability... no checking _mingw_mac.h presence... no checking for _mingw_mac.h... no configure: error: Please check if the mingw-w64 header set and the build/host option are set properly. configure: error: ./configure failed for mingw-w64-crt The _mingw_mac.h header file is present in the cross build environment include directory... Also with =cross-i686-w64-mingw32/binutils-2.35.1-r1: mv: cannot stat '/tmp/portage/cross-i686-w64-mingw32/binutils-2.35.1-r1/image/usr/x86_64-pc-linux-gnu/i686-w64-mingw32/include': No such file or directory mv: cannot stat '/tmp/portage/cross-i686-w64-mingw32/binutils-2.35.1-r1/image/usr/x86_64-pc-linux-gnu/i686-w64-mingw32/lib/*': No such file or directory rm: cannot remove '/tmp/portage/cross-i686-w64-mingw32/binutils-2.35.1-r1/image/usr/x86_64-pc-linux-gnu/include': No such file or directory rm: cannot remove '/tmp/portage/cross-i686-w64-mingw32/binutils-2.35.1-r1/image/usr/x86_64-pc-linux-gnu/lib': No such file or directory - these lines looks a bit suspect! Reproducible: Always Steps to Reproduce: 1. crossdev -C -t i686-w64-mingw32 2. crossdev -t i686-w64-mingw32 Actual Results: environment is built, but with errors for: cross-i686-w64-mingw32/mingw64-runtime (and cross-i686-w64-mingw32/gcc - stage 2) Expected Results: environment is built, without build errors. The cross-i686-w64-mingw32 target was building successfully back in Oct 2020... I've not been able to pin down what is causing the build failures now...
Created attachment 678292 [details] emerge --info
Created attachment 678295 [details] cross-i686-w64-mingw32-binutils.log
Created attachment 678298 [details] cross-i686-w64-mingw32-mingw64-runtime-headers.log
Created attachment 678301 [details] cross-i686-w64-mingw32-gcc-stage1.log
Created attachment 678304 [details] cross-i686-w64-mingw32-mingw64-runtime.log
Created attachment 678307 [details] cross-i686-w64-mingw32-gcc-stage2.log
> * EXTRA_ECONF='--enable-threads=posix' Where does it come from? gcc's pthreads can only be enable after mingw64-runtime provides pthread stubs. But it can't be enabled until gcc-stage2 is built.
(In reply to Sergei Trofimovich from comment #7) > > * EXTRA_ECONF='--enable-threads=posix' > > Where does it come from? Typically added because it's needed build dxvk, some wikis/overlays/ebuilds will ask users to set it. e.g. https://wiki.gentoo.org/wiki/DXVK I use =posix as well but haven't run crossdev script in a long time (portage been keeping my mingw toolchain updated without issues, formerly was with gcc 9.2.x but now upgraded to 10.2.0-r3), so won't confirm any issues. (on a side-note, =posix is _not_ needed for USE=mingw on wine, and there's many convenient ways to get dxvk prebuilt dlls)
(In reply to Sergei Trofimovich from comment #7) > gcc's pthreads can only be enable after mingw64-runtime provides pthread > stubs. But it can't be enabled until gcc-stage2 is built. And right, this makes sense. I did not add this until I had the toolchain working myself.
(In reply to Ionen Wolkens from comment #8) > (on a side-note, =posix is _not_ needed for USE=mingw on wine, and there's > many convenient ways to get dxvk prebuilt dlls) Useful to know... But I do like to be able to build dxvk as well. ;-)
I think slyfox has a better handle on this than I do. I'll just watch from the sidelines.
Let's dupe against older bug to provide some reasonable way to enable pthreads. *** This bug has been marked as a duplicate of bug 631460 ***
