-
Notifications
You must be signed in to change notification settings - Fork 6.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update vcpkg-tool to 2023-02-16. #29664
Update vcpkg-tool to 2023-02-16. #29664
Conversation
You probably also have to merge microsoft/vcpkg-tool#906 to get a green ci. |
3ddb698
to
eb5a8e4
Compare
eb5a8e4
to
c3c71c3
Compare
The new CI regressions are expected, due to unblocking of port builds by no longer pulling in unsupported features via unsupported features and ports. This is the work which started with #28619, lead to microsoft/vcpkg-tool#846, but was ignored in the course of merging microsoft/vcpkg-tool#856 which unblocked even more ports. As a quick mitigation, I suggest to compare with the CI build before this PR, and then add expected failures to the CI baseline. |
Verified this update resolves all the issues with the new xbox supports term. |
…CPKG_POLICY_SKIP_DUMPBIN_CHECKS now that checking for msvc no longer requires dumpbin.
Yep! Also the dumpbin dependency removal resulted in post-build checks being more accurate in a few places.
I'm verifying for each case. |
This never worked on Windows but was guarded before by failed dependencies on arm.
…rts.ps1. microsoft#29540 and microsoft#29481 forgot to rerun generate-ports.ps1. Note that the default port-version was bumped because most everything depends on boost-build that change touched.
…rather than pegtl, if that port happened to get installed first.
Upstream has: ``` // in buildtrees\harfbuzz\src\6.0.0-21fda9578f.clean\src\hb-ft.h #if defined(WINAPI_FAMILY) && (WINAPI_FAMILY != WINAPI_FAMILY_DESKTOP_APP) #define generic GenericFromFreeTypeLibrary #endif ``` resulting in ``` ../src/6.0.0-21fda9578f.clean/src/hb-ft.cc(985): error C2039: 'generic': is not a member of 'FT_FaceRec_' C:\Dev\vcpkg\installed\x64-uwp\include\freetype/freetype.h(1077): note: see declaration of 'FT_FaceRec_' ../src/6.0.0-21fda9578f.clean/src/hb-ft.cc(1010): error C2039: 'generic': is not a member of 'FT_FaceRec_' C:\Dev\vcpkg\installed\x64-uwp\include\freetype/freetype.h(1077): note: see declaration of 'FT_FaceRec_' ``` so this problem is an upstream issue with not supporting non-desktop Windows.
``` warning: The App Container bit must be set for Windows Store apps. The following DLLs do not have the App Container bit set: D:\packages\vcpkg-gfortran_x64-uwp\debug\bin\libgcc_s_seh-1.dll D:\packages\vcpkg-gfortran_x64-uwp\debug\bin\libgfortran-5.dll D:\packages\vcpkg-gfortran_x64-uwp\debug\bin\libquadmath-0.dll D:\packages\vcpkg-gfortran_x64-uwp\debug\bin\libwinpthread-1.dll D:\packages\vcpkg-gfortran_x64-uwp\bin\libgcc_s_seh-1.dll D:\packages\vcpkg-gfortran_x64-uwp\bin\libgfortran-5.dll D:\packages\vcpkg-gfortran_x64-uwp\bin\libquadmath-0.dll D:\packages\vcpkg-gfortran_x64-uwp\bin\libwinpthread-1.dll ``` This previously succeeded in CI but the resulting program could not be distributed.
[vcpkg-gfortran] Also not supported on UWP.
This previously succeeded in CI but the resulting program could not be distributed because it would fail the WACK. (And we could not detect the problem before because it passed skip-dumpbin-checks) |
``` D:\installed\x64-uwp\include\asio/detail/impl/win_iocp_io_context.ipp(535): error C2039: 'VerSetConditionMask': is not a member of '`global namespace'' D:\installed\x64-uwp\include\asio/detail/impl/win_iocp_io_context.ipp(535): error C3861: 'VerSetConditionMask': identifier not found D:\installed\x64-uwp\include\asio/detail/impl/win_iocp_io_context.ipp(538): error C2039: 'VerifyVersionInfo': is not a member of '`global namespace'' D:\installed\x64-uwp\include\asio/detail/impl/win_iocp_io_context.ipp(538): error C3861: 'VerifyVersionInfo': identifier not found ```
Be aware that we had the discussion somewhere already. |
I don't necessarily want to relitigate something that has been litigated before but I see little reason in claiming to people that something is available only for it to blow up in their face when they try to ship. Do you have a link to the previous discussion which may have landed on mitigations? |
seems like the result was that uwp is already broken but it was never marked as broken. |
Yeah there I said
so it seems like this is just documenting that it's broken rather than actually breaking anything. |
That is because the lapack dependency on uwp is fulfilled via clapack. |
find_package(blend2d CONFIG REQUIRED) | ||
target_link_libraries(main PRIVATE blend2d::blend2d) | ||
|
||
Also, define BL_STATIC before any blend2d includes. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Shouldn't this be burned in in some way?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, this is bad. But this PR isn't making that worse.
[lapack-reference]: |
…2-15 # Conflicts: # versions/m-/monkeys-audio.json
Hm I don't understand why it is necessary to bump the port version of every boost port. |
#29540 and #29481 touch |
Actions can no longer approve itself.
…2-15 # Conflicts: # ports/openimageio/portfile.cmake # ports/vtk/pegtl.patch # versions/m-/monkeys-audio.json # versions/o-/openimageio.json # versions/v-/vtk.json
All VTK changes removed due to 8c0fe4a landing. |
https://github.com/microsoft/vcpkg-tool/releases/tag/2023-02-16