execute_one_impl_action is checking for the validity of its arguments, when it
can be enforced by the structure by using smaller focused methods that trickle
down into a common function.
This structure will eventually allow different builders to participate in a
script invocation. Each builder will be told to build/test/install, and they
will not have to deal with anything else. The content of these smaller methods
will end up as part of the builder that uses build-script-impl.
The calculation of the plaforms to build/test were only used by HostSpeficicConfiguration, but were done in BuildScriptInvocation, and to make things work, the full BuildScriptInvocation object was passed down onto HostSpecificConfiguration. The changes in this commit move the code into HostSpecificConfiguration, and only pass `args` into it to let itself calculate the platforms to build/test.
This will allow in the future that HostSpecificConfiguration can to be removed from the main script file, can be tested and can be used by several parts of the build script process. During a transition period, it will be useful for the builder that delegates to build-script-impl, but in the end, it might only be useful for the Swift builder.
As proposed in #23822, the do_* methods should only be using the * parts. Because all building should have the same names, this commit removes the prefix from all the method instances.
Move the calculations of platforms_to_skip_build, platforms_to_skip_test, platforms_archs_to_skip_test, and platforms_to_skip_test_host to their own independent functions. Each function deal with one of them and they are pure functions, which should improve readability a little.
This reverts commit b9e17a7656.
Unfortunately, this caused conflicts on the Apple internal CI.
Reverting this for now. This can be split up into smaller patches to
make it easier to merge.
Now one can on Darwin/Linux build the benchmarks via swiftpm from build-script by passing in:
```
build-script $NORMAL_ARGS --install-swift --install-swiftpm --install-llbuild --toolchain-benchmarks --swiftpm --llbuild
```
This is done using the infrastructure that BenL added for sourcekit-lsp.
LLDB's CMake build forces these values when building on MacOS. It seems
saner to me to set them in build-script than to force them in the CMake
configuration.
In pch-bridging-header-deps test, the PathSanitizingFileCheck is used to
check against YAML escaped paths. This works for Unix paths, since the
slash doesn't need to be escaped, but doesn't work for Windows paths,
because the path separator is escaped, and the replacement done by
PathSanitizingFileCheck only looks for exact matches.
To avoid changes in how Darwin/Linux execute the tests, this commit adds
two options in PathSanitizingFileCheck: Windows compatibility makes the
given paths match both forward and backward slashes; in the additional
YAML compatibility is enabled, escaped backward slashes will also be
matched.
The Windows compatibility is enabled for all the test in case the tests
are running on Windows (change done in lit.cfg), while the YAML
compatibility is only enabled for those tests that might need it (like
the PCH bridging header one). This is in order to not allow possible
empty path components in tests that do not deal with YAML escaped paths.
This will allow using HostSpecificConfiguration from other parts that
are not the main script in the future.
Additionally, a lot of code that was part of BuildScriptInvocation has
been moved into HostSpeficiaConfiguration because nobody else needed it,
and HostSpecificConfiguration was highly coupled with the Invocation
object (through one of its arguments). This should make the
Configuration object a better citizen in other scenarios.
* Revert "Merge pull request #23791 from compnerd/you-know-nothing-clang"
This reverts commit 5150981150, reversing
changes made to 8fc305c03e.
* Revert "Merge pull request #23780 from compnerd/math-is-terrible"
This reverts commit 2d7fedd25f, reversing
changes made to 0205150b8f.
* Revert "Merge pull request #23140 from stephentyrone/mafs"
This reverts commit 777750dc51, reversing
changes made to 0c8920e747.
Merging the CoreFoundation build into the Foundation build results in
the build tree layout to change. Adjust the parameters for that to
allow a migration.
This commit implements SE-0246, by adding conformance to Real to the Float, CGFloat, Double, and Float80 types, implemented either in terms of the system's C math library, existing standard library functionality, or LLVM intrinsics. It includes basic test coverage for these new functions, and deprecates and obsoletes *some* existing functionality in the Platform overlay. We still need to make a decision about how to handle the remaining "tgmath" functions, because obsoleting them is technically a source-breaking change (if users have unqualified names like "exp(1)", it's fine, but it would break users who have used qualified names like "Darwin.exp(1)".)
Fundamental decision to be made; should implementation hooks present like `Float.exp(_ x: Float) -> Float` or like `Float.Maths.exp(_ x: Float) -> Float`? Having the intermediate namespace to group them as in the second is definitely nicer, but it requires a little bit of extra machinery, and much more importantly, there doesn't seem to be any way to make access to the static `Maths` associatedtype transparent.
env is not portable in systems that do not have env, like for example Windows. However, passing the environment is supported by shell.run using the env parameter, which should be portable.
For some reason shell.run doesn't merge the environment with the current environment by default, so we do it externally to not modify other possible usages of the function.
Passing an empty string to this option either as
"--extra-swift-args=''" or "--extra-swift-args ''" would cause
build-script-impl to fail to parse some arbitrary number of other
options.
This is impossible to debug from the build log because all of the
build-script-impl options still show up in the log. So, cutting and
pasting the build-script-impl command actually works.
Turns out this is why build-script build and install stages never
worked for me, beyond running cmake.
The manipulation of host-test and skip-android-host was a little
different than the equivalent skip-ios-host and similar variables. These
changes make them closer and allows executing only the compiler tests,
but skip the test that need an Android device to run.
- Disables the upload command of the tests if the subset is the
non-executable tests. The non-executable test do not need to be
uploaded, and in the case of Android, a device doesn't need to be
connected, so trying to connect to one will fail.
- Fix a problem where the swift_interpreter feature was removed without
first checking if it was really added.
- Only enable the host tests (the compiler tests) in the Android CI
preset (there's no device attached to that server, but currently only
the Linux tests were being executed, which doesn't make a lot of
sense).
- Move the decision about which platform support device/host tests into
the platform themselves, which allows Android to have device/host
tests. Also modify a little bit the logic around enabling/disabling
the test suite to allow running only the host tests of a platform.
- Fix the suffix name for the target of non-executable tests in a couple
of places.
