* [vcpkg] Add `vcpkg export` to E2E tests. Enable E2E tests on MacOS.
* [vcpkg] Fix export --raw --output-dir=/path/ by changing directory to new export root
Co-authored-by: Robert Schumacher <roschuma@microsoft.com>
Co-authored-by: Billy Robert O'Neal III <bion@microsoft.com>
* [vcpkg format-manifest] initial convert-control attempt
TODO: manifest comments! we should keep $directives
* Finalize x-format-manifest
First, fix Json::parse -- "\c", for any c, was incorrectly parsed.
It would emit the escaped character, and then parse the character, so
that `\b` would give you { '\b', 'b' }.
Second, canonicalize source paragraphs as we're parsing them. This found
an error in qt5 -- The `declarative` feature was listed twice, and we
now catch it, so I removed the second paragraph.
Add PlatformExpression::complexity to allow ordering platform
expressions in a somewhat reasonable way.
Notes:
- We allow `all_modules` as a feature name for back-compat with
paraview
- In order to actually convert CONTROL to vcpkg.json, we'd need to
rename the qt5 `default` feature.
- We need to add support for $directives in x-format-manifest
* fix qt5 port
* format
* fix compile
* fix tests for canonicalization
* Clean up code
* add error message for nothing to format
* add extra_info field
* add `const X&` overloads for `Object::insert[_or_replace]`
* fix compile
* simple CRs
* add tests
* format
* Fix mosquitto port file
also unmerge a line
* fail the tests on malformed manifest
* fix format_all
* fix coroutine port-version
* format manifests
* stage checking formatting
this means that the port CI won't run if the formatting failed
* fix invalid names
* add formatting of manifests
* fix dependsOn name
* fix a thing
* CRs
* oops, typo
* Added python script to generate all packages file list, added to azure pipeline to build at the end of the run
* fixed an issue that causes some leading slashes were being removed
* Header Database now includes hpp files for C++ header files
* Changed Header Database to include all files under \include\ folder
* Apply suggestions from code review, changing condition from eq to ne
Co-authored-by: Billy O'Neal <bion@microsoft.com>
* Update last condition from eq to ne
Co-authored-by: Billy O'Neal <bion@microsoft.com>
This changes our PR builds to treat 'fail' in the ci.baseline.txt as 'skip' instead of using tombstones.
We currently have large numbers of spurious failures that get enshrined in PRs through no fault of a PR author, removing the tombstones concept will fix those by allowing the user to retry. This does mean we accept some risk of not detecting when a port is 'fixed', but that failure is reasonable for us to handle after we see it in CI, but that seems worth it given that it lets us get rid of the tombstone concept.
This also helps out the binary caching feature, because we don't have to figure out how to productize tombstones.
* [vcpkg] Implement --x-write-nuget-packages-config= setting for `install` and `x-set-installed`.
* [vcpkg] Add end-to-end testing suite for install, remove, and binary caching
* [vcpkg] Define `$TestingRoot in end-to-end-tests.ps1
* [vcpkg] Address CR comments
Co-authored-by: Robert Schumacher <roschuma@microsoft.com>
* [vcpkg] Add disk space report to PR/CI
Example output:
```
Disk Label Size Free Space
---- ----- ---- ----------
C: Sabrent 1907 GiB 1239 GiB
D: Dev 447 GiB 383 GiB
E: Samsung 960 Pro 1908 GiB 1084 GiB
H: Rocket 3815 GiB 863 GiB
R: 0 B 0 B
S: 0 B 0 B
```
* [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
* [vcpkg] Use a crypto RNG to generate admin passwords
* Apply code review comments from Stephan and Casey.
* Extract functions into create-vmss-helpers.psm1.
* Put Wait-Shutdown back.
* [vcpkg] Remove powershell from the 'run vcpkg ci' path to reduce hangs from msys components.
* The BatchScript task uses filename rather than filePath
* cmd hates BOMs
* cmd hates forward slashes
`files.h/files.cpp`:
* Add end and else comments to all macros.
* Add "remove_all_inside" function which empties a directory without actually deleting the directory. This is necessary to handle the case where the directory is actually a directory symlink.
* Change remove_all to use std::remove when VCPKG_USE_STD_FILESYSTEM is set; this will engage POSIX delete support available on current Win10.
`commands.ci.cpp`: empty "installed".
`*/initialize_environment.*`: No longer clean the directories outside the tool.
`ci-step.ps1`: Remove unused console output tee-ing.
* [vcpkg] Bump Linux VM memory size and do all operations on the temporary disk.
* [llvm] disable optional dependencies to libxml2 and zlib
* [vcpkg] Reduce quotas on the share to something reasonable.
* [mpir] Skip on Linux because it conflicts with libgmp.
Co-authored-by: yurybura <yurybura@gmail.com>
* Adds scripts to generate scale sets for testing Linux.
* Note workaround for https://github.com/microsoft/azure-pipelines-agent/pull/2929
* Switches Windows validation to 'Spot' VMs.
* Opens the git port 9418.
* Removes provisioning of the no longer used 'logs' file share.
* Changes Azure region to 'westus2', which is cheaper.
* Adds +x to all the scripts in scripts/azure-pipelines.
* Use 'xml-results' for all platforms instead of 'raw xml results' on Windows.
Included is a script that sets up all vcpkg's Azure infrastructure for Windows PR tests, and several updates to baselines. The baseline updates are generally caused by an updated copy of the MSVC++ compiler caused by updating the VMs, but some are caused by missed failures only detected now because this did a cleared out archives directory first.
Some of the build infrastructure isn't what I'd call 'pretty' (e.g. we're split into more scripts and such than I'd like) but this mirrors how our existing PR system works.
It is expected that the existing vcpkg Windows PR system will hate these baseline updates so we'll need to merge this, then remove that (duplicate) workflow immediately afterwards, then delete all the Windows VMs powering the old infrastructure.
