* [libusb] support linux
* Add Supports
* Using vcpkg_configure_make instead
* Revert to install headers in windows
* Fix failures in libftd1
* Update treehopper
* Add failed ports in ci baseline file
* Revert to remove azure-kinect-sensor-sdk:x64-windows-static=fail in baseline file
* Resolve conflicts
Co-authored-by: PhoebeHui <v-phma@microsoft.com>
* [libb2] Add supports and install pkg-conifg on macOS VM
* Add blank line to the end of the json file
* vcpkg x-format-manifest --all
* Remove failed ports in ci baseline file
* [kfr] Add new port
* [kfr] Modernize and update to lastest
* [kfr] Add features
* [kfr] Fail fast on arm-uwp, arm64-windows
* [kfr] Fail fast on linux
* [qt5] update qt to 5.15 and make ports depend on qt5-base[core].
* [qt5] update and remove patches
* [baseline] skip qt5-canvas3d due to deprecation
* [qwt] update to 6.1.4 and fix build issues
* [qt5-webengine] remove old patches
* [qwt] fix static patch
* [qt5-location] add a message that the port has a sprious failure
* [qt5-webengine] fix filename of patch
* [qt5-webengine] add build.patch (ci is probably fast than me in checking for more errors)
* [qt5-winextras] depend on qt5-base[core]
Co-authored-by: Jack·Boos·Yu <47264268+JackBoosY@users.noreply.github.com>
* [qt5-base] update latest feature (latest=core)
* [baseline] remove qt5-webengine for nonstatic windows since CI can handle it now
* [qt5-base] make qt5-base fail to get logs.
* [angle] include missing egl sources.
* [qt5-base] patch egl detection logic.
* [baseline] remove qt-x11extras:x64-linux=fail
* [qt5-base] need to be explicit about xcb to build qt5-x11extras
* [CI] add necessary system libraries to linux CI
* [qt5-base] remove comment to retry on linux
* [qt5-base] add static_init patch for MSVC
* [qt5-base] refined workaround to not regress linux and osx due to
* [paraview] remove unnecessary patches with newer qt versions
* [vcpkg/baseline] add webengine back to skip
* [vtk] fix missing include regression using patch from upstream
* [vtk] fix patch
* [qt5-base] replace with patch from upstream
* [paraview] use hdf5 1.10 api.
* [paraview] bump control
* Add more packages to Linux machines.
* [qt5-base] forgot to add the patch
* Delete 6943.diff
Co-authored-by: Jack·Boos·Yu <47264268+JackBoosY@users.noreply.github.com>
Co-authored-by: Billy Robert O'Neal III <bion@microsoft.com>
- New upstream version 10.35
- Update patches:
pcre2-10.35_fix-space.patch
pcre2-10.35_fix-uwp.patch
- Add patch:
pcre2-10.35_fix_postfix_for_debug_Windows_builds.patch
- Drop patch (fixed upstream):
fix-arm64-config.patch
- Since 10.35, pkgconfig files are created using CMake
* fix debug suffix of pkgconfig files under Windows
* add vcpkg_fixup_pkgconfig()
- Remove bin and debug/bin in case of static build
- Update ci.baseline.txt
Remove "pcre2:arm-uwp=fail", because arm-uwp passes now
Add "unicorn-lib:arm-uwp=fail"
This changes our PR builds to treat 'fail' in the ci.baseline.txt as 'skip' instead of using tombstones.
We currently have large numbers of spurious failures that get enshrined in PRs through no fault of a PR author, removing the tombstones concept will fix those by allowing the user to retry. This does mean we accept some risk of not detecting when a port is 'fixed', but that failure is reasonable for us to handle after we see it in CI, but that seems worth it given that it lets us get rid of the tombstone concept.
This also helps out the binary caching feature, because we don't have to figure out how to productize tombstones.
==== Changes Related to manifests ====
* Add the `manifests` feature flag
* This only says whether we look for a `vcpkg.json` in the cwd, not
whether we support parsing manifests (for ports, for example)
* Changes to the manifests RFC
* `"authors"` -> `"maintainers"`
* `--x-classic-mode` -> `-manifests` \in `vcpkg_feature_flags`
* reserve `"core"` in addition to `"default"`, since that's already
reserved for features
* Add a small helper note about what identifiers must look like
* `<license-string>`: SPDX v3.8 -> v3.9
* `"feature"."description"` is allowed to be an array of strings as well
* `"version"` -> `"version-string"` for forward-compat with versions
RFC
* Add the `--feature-flags` option
* Add the ability to turn off feature flags via passing
`-<feature-flag>` to `VCPKG_FEATURE_FLAGS` or `--feature-flags`
* Add CMake toolchain support for manifests
* Requires either:
* a feature flag of `manifests` in either `Env{VCPKG_FEATURE_FLAGS}`
or `VCPKG_FEATURE_FLAGS`
* Passing the `VCPKG_ENABLE_MANIFESTS` option
* The toolchain will install your packages to
`${VCPKG_MANIFEST_DIR}/vcpkg_installed`.
* Add MSBuild `vcpkg integrate install` support for manifests
* Requires `VcpkgEnableManifest` to be true
* `vcpkg create` creates a port that has a `vcpkg.json` instead of a
`CONTROL`
* argparse, abseil, 3fd, and avisynthplus ports switched to manifest
from CONTROL
* Add support for `--x-manifest-root`, as well as code for finding it if
not passed
* Add support for parsing manifests!
* Add a filesystem lock!
==== Important Changes which are somewhat unrelated to manifests ====
* Rename `logicexpression.{h,cpp}` to `platform-expression.{h,cpp}`
* Add `PlatformExpression` type which takes the place of the old logic
expression
* Split the parsing of platform expressions from checking whether
they're true or not
* Eagerly parse PlatformExpressions as opposed to leaving them as
strings
* Add checking for feature flag consistency
* i.e., if `-binarycaching` is passed, you shouldn't be passing
`--binarysource`
* Add the `Json::Reader` type which, with the help of user-defined
visitors, converts JSON to your internal type
* VcpkgArgParser: place the switch names into a constant as opposed to
using magic constants
* In general update the parsing code so that this ^ works
* Add `Port-Version` fields to CONTROL files
* This replaces the existing practice of
`Version: <my-version>-<port-version>`
==== Smaller changes ====
* small drive-by cleanups to some CMake
* `${_VCPKG_INSTALLED_DIR}/${VCPKG_TARGET_TRIPLET}` ->
`${CURRENT_INSTALLED_DIR}`
* Remove `-analyze` when compiling with clang-cl, since that's not a
supported flag (vcpkg's build system)
* Add a message about which compiler is detected by vcpkg's build
system machinery
* Fix `Expected::then`
* Convert `""` to `{}` for `std::string` and `fs::path`, to avoid a
`strlen` (additionally, `.empty()` instead of `== ""`, and `.clear()`)
* Add `Strings::strto` which converts strings to numeric types
* Support built-in arrays and `StringView` for `Strings::join`
* Add `operator<` and friends to `StringView`
* Add `substr` to `StringView`
* SourceParagraphParser gets some new errors
* [hdf5] Update HDF5 to 1.12 and use github live-clones instead of binary download
* [hdf5] bump control
* [hdf5] fix szip and typo about linkage
* [hdf5] add missing package namespace.... they somehow missed to set it?
* [hdf5] retry without "
* [field3d] fix regression and update. Remove windows as unsupported since the cmakelists tries to support windows
* fix typo
* remove ws
Co-authored-by: NancyLi1013 <46708020+NancyLi1013@users.noreply.github.com>
* linebreak first patch
Co-authored-by: NancyLi1013 <46708020+NancyLi1013@users.noreply.github.com>
* ws change
Co-authored-by: NancyLi1013 <46708020+NancyLi1013@users.noreply.github.com>
* [field3d] update copying installation
* [hdf5] use official github repo
* clean baseline
Co-authored-by: NancyLi1013 <46708020+NancyLi1013@users.noreply.github.com>
Co-authored-by: Jack·Boos·Yu <47264268+JackBoosY@users.noreply.github.com>
* [cgns] add port cgns
* [paraview] add new port
* [paraview] correct dependencies; disable catalyst
* [VTK/ParaView] Make ParaView build
* [paraview] add missing patch
* [paraview] add cgns patch
* [paraview] add qt plugin patches back in. Remove in the future
* [vtk] slightly improved patches
* [vtk] don't disable findPython
* [paraview] add more features
* [paraview] remove X11Extras from OSX
* [vtk] make vtk workcorrectly with python and prepare for paraview with python
* [paraview] Python fixes
* [paraview] small bin -> lib replacement fix for static builds
* [vtk] fix patch path
* update baseline to exclude paraview for the time being on osx and linux.
requires a fix in qt5-base linkage of libpq.
* [paraview]try setting LD_LIBRARY_PATH for unix systems
* [paraview] a bit of cleanup
* update baseline
* fix typo in baseline
* [paraview] remove unnecessary patch
* [vtk] replace custom patch with upstream patch
* [paraview] cleanup in the portfile.
* Update ports/paraview/CONTROL
Co-authored-by: Jack·Boos·Yu <47264268+JackBoosY@users.noreply.github.com>
* [tinyfiledialogs] Re-trigger CI test
* [msix] Fix source hash
* [geographiclib] Re-trigger CI test
* [soundtouch] Re-trigger CI test
* update baseline
* [plplot] Re-trigger CI test
* [dmlc] Fix issues with parallel configuration
* [tclap] Re-trigger CI test
* [itpp] Re-trigger CI test
* [mpg123] Re-trigger CI test
* [qwt] Re-trigger CI test
* [soundtouch] Re-trigger CI test
* [pthreads] Re-trigger CI test
* Enable compilation of libpcap port on x86-windows and x64-windows
As winpcap and libpcap install the same headers, this two
port have been marked as not not compatible, and cannot be
installed together.
* Update ci.baseline.txt
* Add libcrafter failing ports to ci.baseline.txt
* [BitSerializer] Update to version 0.9
* [bitserializer] Update to version 0.10
* [bitserializer] Enable all features by default
* [bitserializer] Update hash of tagged commit and remove "rapidyaml-archive" from default list of features
* [bitserializer] Add alias packages for deprecated BitSerializer's libraries
* Skip bitserializer-rapidyaml for arm and osx
* Update scripts/ci.baseline.txt
Co-authored-by: Pavel_Kisliak <Pavel_Kisliak@epam.com>
Co-authored-by: PhoebeHui <20694052+PhoebeHui@users.noreply.github.com>
* [licensepp] Add new port
* [cryptopp-pem] Update to a newer version that will make lincensepp happy
* Revert "[cryptopp-pem] Update to a newer version that will make lincensepp happy"
This reverts commit 726e3ce3d2.
This version can not compile with cryptopp.
* [licensepp] Finish it
* [licensepp] Fix interface include directory
* [licensepp] Fixup cmake targets
* [licensepp] -O3 is not recognised by msvc
* [licensepp] Fix _iterator_debug_level value mismatch
* [licensepp] Update patch file
* [licensepp] Fail fast on UWP
* [licensepp] Simplify patch file
* [licensepp] Tabs -> Spaces
* [avisynthplus] Upgrade to 3.6.0
AviSynth+ also supports linux and macOS now
* [avisynthplus] Allow x64-linux build fail on CI
AviSynth+ requires C++17 to build correctly. C++17 std::filesystem is not available in gcc 7.5 used in CI.
* [Arrow] Update to 0.17.1
* Remove arrow:x64-linux=fail from ci.baseline.txt.
Add explicit tool dependencies on Flex and Bison for Linux and OSX.
* Revert arrow dependency on Flex/Bison, it's Thrift that needs them and its portfile is already fine.
* Use vcpkg_fail_port_install(ON_ARCH x86 arm arm64) instead of custom check.
Remove thrift:x64-osx=fail from ci.baseline.txt (we know arrow depends on it, and arrow:x64-osx has been shown to work in 3rd party project).
* Disable using pkg-config files to locate dependencies in arrow
This is incompatible with vcpkg as these files refer to paths in the
packages directory rather than the installed directory, so this only
works if the packages haven't been cleaned.
* Mark thrift:x64-osx as still failing until a proper solution for Bison can be found.
* Update ports/arrow/portfile.cmake
Co-authored-by: Adam Reeve <adreeve@gmail.com>
Co-authored-by: NancyLi1013 <46708020+NancyLi1013@users.noreply.github.com>
* [vcpkg] Remove do-nothing Set-Content from Windows azure-pipelines.yml.
* [vcpkg] Fix OSX CI by ensuring the downloads directory exists in advance, and extract common command line parameters with powershell splatting.
* [tensorflow-cc] Prevent hang building tensorflow-cc asking to configure iOS.
* Skip ignition-msgs5:x64-osx
* [google-cloud-cpp] Consolidate all google-cloud* packages
Starting with the v1.14.0 release all google-cloud* packages and the
googleapis package are kept in a single repository and released
simultaneously.
* Update CI baseline.
google-cloud-cpp used to inherit failures from googleapis, which
indicates these are (probably) problems in the generated protobuf/grpc
code.
