* more filtering of tool paths.
* test it in CI
* only use tools/* and tools/*/bin
* debugging
* test removal of tools/* if tools/*/bin is available
* fix variable naming
* cleanup
* revert ports/qtdeclarative/portfile.cmake
* Update vcpkg-tool to 2022-03-24
* Hook up https://github.com/microsoft/vcpkg-tool/pull/345
* Hook up https://github.com/microsoft/vcpkg-tool/pull/442
* Update vcpkg-tool to 2022-03-25
* Analysis of failures.
* [Most recent nightly build failed](https://dev.azure.com/vcpkg/public/_build/results?buildId=69427)
* [Validation of this tool update failed](https://dev.azure.com/vcpkg/public/_build/results?buildId=69417)
## Common to both:
PASSING, REMOVE FROM FAIL LIST: chartdir:x64-windows (.\scripts\ci.baseline.txt)
PASSING, REMOVE FROM FAIL LIST: chartdir:x64-windows-static-md (.\scripts\ci.baseline.txt)
PASSING, REMOVE FROM FAIL LIST: chartdir:x86-windows (.\scripts\ci.baseline.txt)
Probably fixed by https://github.com/microsoft/vcpkg/pull/23701
PASSING, REMOVE FROM FAIL LIST: gmp:x64-uwp (.\scripts\ci.baseline.txt)
PASSING, REMOVE FROM FAIL LIST: gmp:x64-windows-static-md (.\scripts\ci.baseline.txt)
Probably fixed by https://github.com/microsoft/vcpkg/pull/23466 ?
REGRESSION: colmap:x64-windows-static-md failed with BUILD_FAILED. If expected, add colmap:x64-windows-static-md=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
I don't know exactly what changed. I observe that
* this thing depends on a *lot* of stuff
* on March 14 we didn't even attempt to build this
* the x64-windows ones are already in the baseline
so I skipped it.
REGRESSION: qtdeclarative:x64-windows. If expected, add qtdeclarative:x64-windows=fail to .\scripts\ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\1\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: qtdeclarative:x64-windows failed with BUILD_FAILED. If expected, add qtdeclarative:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
This is a reporting change: The new world order also includes host build failures which is why it's duplicated.
See also https://github.com/microsoft/vcpkg/issues/23714
See also https://github.com/microsoft/vcpkg/issues/23490
I'm nervous about baslining this because it seems most of the qt world is built on top of this port
I filed https://github.com/microsoft/vcpkg/issues/23824 about this and @Neumann-A indicated this should be fixed by https://github.com/microsoft/vcpkg/pull/23755
REGRESSION: nettle:x64-uwp. If expected, add nettle:x64-uwp=fail to .\scripts\ci.baseline.txt.
REGRESSION: nettle:x64-windows-static-md. If expected, add nettle:x64-windows-static-md=fail to .\scripts\ci.baseline.txt.
REGRESSION: nettle:x64-uwp failed with BUILD_FAILED. If expected, add nettle:x64-uwp=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
REGRESSION: nettle:x64-windows-static-md failed with POST_BUILD_CHECKS_FAILED. If expected, add nettle:x64-windows-static-md=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
Didn't analyze, probably fixed by https://github.com/microsoft/vcpkg/pull/23519 ?
REGRESSION: libgpg-error:x64-uwp. If expected, add libgpg-error:x64-uwp=fail to .\scripts\ci.baseline.txt.
REGRESSION: libgpg-error:x64-uwp failed with BUILD_FAILED. If expected, add libgpg-error:x64-uwp=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
This was broken by VS2022 update:
```
C:\Program Files\Microsoft Visual Studio\2022\Enterprise\MSBuild\Microsoft\VisualStudio\v17.0\AppxPackage\Microsoft.AppXPackage.Targets(892,25): error MSB4086: A numeric comparison was attempted on "$(TargetPlatformMinVersion)" that evaluates to "" instead of a number, in condition "'$(TargetPlatformMinVersion)' >= '10.0.17200.0'". [C:\Dev\vcpkg\buildtrees\libgpg-error\x64-uwp-rel\error-1.42-2324ddbc71.clean\SMP\libgpg-error_winrt.vcxproj]
```
REGRESSION: libmikmod:x64-osx. If expected, add libmikmod:x64-osx=fail to .\scripts\ci.baseline.txt.
REGRESSION: libmikmod:x64-osx failed with BUILD_FAILED. If expected, add libmikmod:x64-osx=fail to /Users/vagrant/Data/work/2/s/scripts/azure-pipelines/../ci.baseline.txt.
Broken between [2022-03-16](https://dev.azure.com/vcpkg/public/_build/results?buildId=68947) and [2022-03-18](https://dev.azure.com/vcpkg/public/_build/results?buildId=69051). Unfortunately I don't see obvious reasons why. Nothing else depends on this and nobody has noticed in 2 weeks, so I'm baslining it for now. (Will investigate shortly...)
## Only broken in tool update:
REGRESSION: mesa:x64-windows failed with BUILD_FAILED. If expected, add mesa:x64-windows=fail to C:\a\2\s\scripts\azure-pipelines/../ci.baseline.txt.
```
-- Downloading https://gitlab.freedesktop.org/mesa/mesa/-/archive/mesa-21.2.5/mesa-mesa-21.2.5.tar.gz -> mesa-mesa-mesa-21.2.5-1.tar.gz...
-- Extracting source /Users/vagrant/Data/downloads/mesa-mesa-mesa-21.2.5-1.tar.gz
-- Applying patch swravx512-post-static-link.patch
-- Applying patch swr-msvc-2.patch
-- Applying patch swr-llvm13.patch
-- Applying patch radv-msvc-llvm13-2.patch
-- Applying patch d3d10sw.patch
-- Using source at /Users/vagrant/Data/buildtrees/mesa/src/esa-21.2.5-2df234d2b1.clean
Traceback (most recent call last):
File "<string>", line 1, in <module>
ModuleNotFoundError: No module named 'mako'
CMake Error at ports/mesa/portfile.cmake:85 (message):
Python package 'mako' needs to be installed for port 'mesa'.
Complete list of required python packages: setuptools;mako
Call Stack (most recent call first):
ports/mesa/portfile.cmake:91 (vcpkg_get_python_package)
scripts/ports.cmake:145 (include)
```
Looks like this is being tracked by https://github.com/microsoft/vcpkg/pull/23089 ; perhaps that we don't have as aggressive a recycling strategy for macos boxes as we do for the others has let different machines give different results?
## Only broken without tool update:
REGRESSION: chromium-base:x64-osx. If expected, add chromium-base:x64-osx=fail to .\scripts\ci.baseline.txt.
This one has been constantly flaky; I baselined it.
REGRESSION: libxml2:x64-osx. If expected, add libxml2:x64-osx=fail to .\scripts\ci.baseline.txt.
This port uses vcpkg_from_git and the upstream server was down during the build.
* Restore chartdir to the baseline, I thought https://github.com/microsoft/vcpkg/pull/23732 had been merged.
* [fuzzylite] fix unix build
* update version scheme
* version
* unsupported on UWP
* remove from CI baseline
* version
* fix typo
* version
* windows only
* version
* Add fuzzylite:x64-osx
* Fix typo
* cr
* version
* Fix string replace on static
* version
* Don't rename artifacts
* version
* Trigger some test ports from vcpkg.cmake changes
* Bump test port version
* Revert "Bump test port version"
* Use old toolchain version for parent hashes
* Add ABI hashing note to vcpkg.cmake
* Test vcpkg.cmake with cmake instead vpckg-ci-paraview
PASSING, REMOVE FROM FAIL LIST: intelrdfpmathlib:x64-linux
PASSING, REMOVE FROM FAIL LIST: ocilib:x64-windows-static
PASSING, REMOVE FROM FAIL LIST: skia:arm64-windows
PASSING, REMOVE FROM FAIL LIST: yato:x64-windows-static-md
REGRESSION: argtable3:arm-uwp
No repro. Maybe fixed by https://github.com/microsoft/vcpkg/pull/23655
REGRESSION: infoware:x64-windows-static
```
-- Found Git: C:/agent/externals/git/cmd/git.exe (found version "2.30.2.windows.1")
Cloning into 'pciids'...
fatal: unable to access 'https://github.com/pciutils/pciids/': Could not resolve host: github.com
CMake Error at CMakeLists.txt:111 (message):
Cloning/pulling pciids repository from https://github.com/pciutils/pciids
failed with 128.
The pci.ids file, downloadable from https://github.com/pciutils/pciids or
http://pci-ids.ucw.cz, is required for building infoware, and cloned
automatically from that GitHub repository by default.
To use a local copy, set INFOWARE_PCI_IDS_PATH to its location.
```
Looks like this port needs to be fixed to work with artifact caching.
REGRESSION: libgpg-error:x64-uwp
```
"D:\buildtrees\libgpg-error\x64-uwp-rel\error-1.42-2324ddbc71.clean\SMP\libgpg-error.sln" (Rebuild target) (1) ->
"D:\buildtrees\libgpg-error\x64-uwp-rel\error-1.42-2324ddbc71.clean\SMP\libgpg-error_winrt.vcxproj" (Rebuild target) (2) ->
(_GetPackageFileExtensions target) ->
C:\Program Files\Microsoft Visual Studio\2022\Enterprise\MSBuild\Microsoft\VisualStudio\v17.0\AppxPackage\Microsoft.AppXPackage.Targets(892,25): error MSB4086: A numeric comparison was attempted on "$(TargetPlatformMinVersion)" that evaluates to "" instead of a number, in condition "'$(TargetPlatformMinVersion)' >= '10.0.17200.0'". [D:\buildtrees\libgpg-error\x64-uwp-rel\error-1.42-2324ddbc71.clean\SMP\libgpg-error_winrt.vcxproj]
```
This looks like a VS2022 regression.
REGRESSION: chromium-base:x64-osx
This port is still having the strange behavior where building it once fails but building it again succeeds O_O
REGRESSION: libmikmod:x64-osx
This is emitting a bunch of spew about standard library symbols.
REGRESSION: qtapplicationmanager:x64-windows
Should be fixed by https://github.com/microsoft/vcpkg/pull/23638
* Update to 3.4.2
* Disable ATL for uwp and for arm64
* Fix linking on uwp
* Fix pc file fixup
* Update versions
* Declare dependency on atlmfc.
Co-authored-by: Billy Robert O'Neal III <bion@microsoft.com>
* Update Linux VM CUDA to 11.6.
* Update Intel OneAPI
* Update Visual Studio to 2022, and remove deployment of v140 and v141 compilers.
* Update CUDA to 11.6 on Windows.
* Update pools.
* cix ci.baseline.txt
Co-authored-by: nicole mazzuca <mazzucan@outlook.com>
* Remove pixel from ci baseline
* don't build examples
* [pixel] update to master
* copy license from portfile.cmake
* version
* license
* version
* fix typo
* version
* use version-date
* version
* change version date to today
* version
* disable crt warnings on uwp
* version
* use vcpkg provided libs before system frameworks on macOS, remove now unnecessary mod to giflib
* fix references
* fixes from review
* apply fixes from review - remove warnings too!
* [giflib] add license
* fix references
Co-authored-by: Victor Romero <viromer@microsoft.com>
* add vcpkg-tool-lessmsi
* add vcpkg-tool-python2
* adjust vcpkg_find_acquire_program
* use python2 dependency in qt5-base
* remove python2 from qt5-base
* add python2 to qt5-webengine
* use python3 instead of python2
* add some deps on the new vcpkg-tool-python2
* Apply suggestions from code review
Co-authored-by: nicole mazzuca <83086508+strega-nil-ms@users.noreply.github.com>
* add license field
* fix test detection
* error on failure to find python2
* move logs to correct position.
* add dep to qtwebengine
* adjust pybind cmake config.
* fix shiva
* move python2 to manual-tools
adjust search path
generate a details.cmake for vcpkg_find_acquire_program
* format manifest
* use version-date
* version stuff
* replace version-string
* more version stuff
* fix search path and add supports expression
* version stuff
* shiva version
* remove arm windows from support
* reuse output_path
use lessmsi is a host dep so use the correct path
* version stuff
* Fix x86 regression
* version stuff
* try fixing qtinterfaceframework
* version stuff
Co-authored-by: Alexander Neumann <you@example.com>
Co-authored-by: nicole mazzuca <83086508+strega-nil-ms@users.noreply.github.com>
Co-authored-by: Victor Romero <viromer@microsoft.com>
Co-authored-by: Victor Romero <romerosanchezv@gmail.com>
* [python3] Update to Python 3.10.2.
* Version nonsense
* Update Python tool to 3.10.2.
* [python3] Update version file
* [python3] Update versions again
Co-authored-by: Victor Romero <viromer@microsoft.com>
* Update vcpkg-tool to 2022-03-08.
* Back off required git version to something actually in the build lab.
* Remove hiding of the output of --parent-hashes.
* Update 7zip reference as indicated in https://github.com/microsoft/vcpkg-tool/pull/413
* Add triplet to name of main CI task
* Fix osx bootstrap task name
* Hide output from switch to parent hash
* Bump test port version
* Hardcode osx+linux triplet names
* Improve logging order
* Add error handling to each tool download
* Revert "Bump test port version"
* Don't prefetch 7zip
* Use host tools by default.
* use GLOB_RECURSE and filter the result
* be positiv ;)
* give me debug output
* Revert "give me debug output"
This reverts commit 17737bceda8ed0dbe1a60c2fd831a55e53707dcd.
* remove unnecessary if(IS_DIRECTORY)
Co-authored-by: Alexander Neumann <you@example.com>
* update cmake to 3.22.0-rc3 to verify catalogue compatibility and be ready for final release
* update to cmake 3.22 final release
* [vcpkgTools] update to cmake 3.22.1
* update test port
* update cmake to 3.22.2
and git for windows to 2.35.1.2
* fix missing uuid on osx in ignition-transport.
* modernize
* fix hash
* fix qtbase regression
* try fixing the pc file search
* hidaoi give trace logs
* give me trace output
* bit of pkgconf changes. Need system search paths for pkgconf to work correctly.
* fix underscore
* remove extra x
* revert port changes
* add license to cmake test port
* fix arm regression due to not finding taglib.
* version stuff
Co-authored-by: Stefano Sinigardi <stesinigardi@hotmail.com>
Co-authored-by: Billy Robert O'Neal III <bion@microsoft.com>
Co-authored-by: JackBoosY <yuzaiyang@beyondsoft.com>
Co-authored-by: Alexander Neumann <you@example.com>
* [get_cmake_vars] add -isysroot flag independednt of crosscompiling
Even without crosscompiling it is possible to specify a VCPKG_OSX_SYSROOT or
a SDKROOT this should be popagated to autotools and meson builds as well.
* [meson] add -arch and -isysroot to linker flags in case of osx crosscompile
* [meson] avoid DEPRECATION warning about flags in the [properties] section
* [meson] don't set OSXROOT evironment varibale
* [meson] don't set MACOSX_DEPLOYMENT_TARGET environment variable
cmake appends the -mmacosx-version-min flag c/cxx flags, overriding the
value set via VCPKG_C_FLAGS. By setting the environment variable, the
VCPKG_C_FLAGS value was used for meson builds. Now the same value is
taken for both.
* [meson/make] Add CMAKE_OSX_DEPLOYMENT_TARGET to VCPKG_DETECTED_CMAKE_C_FLAGS
This is done in the same way in CMake internally
