Specifically:
1. SKIP_TEST_BENCHMARK => SKIP_TEST_BENCHMARKS
I looked for --skip-test-benchmark (notice no s) and SKIP_TEST_BENCHMARK. It
seems like it is dead, especially since we have --skip-test-benchmarks.
2. Remove the code that works around SKIP_TEST_BENCHMARK typo.
* LLDB assertions are on by default, like swift assertions
* LLDB assertions can be enabled/disabled globally with the --assertions
and --no-assertions options
Partially addresses: rdar://38524846
This adds two flags to the build script to enable/disable assertions in
llbuild: --llbuild-assertions, --no-llbuild-assertions
The default value is taken from the global assertions flag.
* Implemented a presets module which includes a more robust and easy to understand parser. Moved the swift-sdks migration code to a new migration module and added testing for both the presets and migration modules. Also converted build-script to use the new presets parser.
* Switched the expansion algorithm to resolve mixins in-place (little known feature) and also changed the parser to skip all non-preset sections. Tests are included for these two behaviors.
* Re-worked the presets error hierarchy to have more descriptive and information packed exception classes. Also re-worked the PresetParser to catch duplicate preset declarations and duplicate options in a single preset. There's some special shim-code to handle the disconnect between the Python 2 ConfigParser module and the Python 3 update which adds DuplicateOptionError.
* Respect the TOOLCHAINS environment variable if set when deciding the darwin xcrun toolchain.
* Updated test/lit.cfg to respect the TOOLCHAINS environment variable when deciding the xcrun toolchain for Darwin platforms, rather than hardcoding the default.
* Fixed the default darwin xcrun toolchain logic.
These don't make sense to build separately, and indeed it's likely that PlaygroundLogger will soon depend on PlaygroundSupport.
As a result, build them as one product (playgroundsupport) in build-script.
Aside from removing the playgroundlogger product, this otherwise continues to build PlaygroundLogger and PlaygroundSupport the same way.
This is for <rdar://problem/36512531>.
build-script can already build lldb on Darwin using its Xcode project,
but it's useful to support the CMake build as well.
The CMake build allows incremental rebuilds with build-script. I expect
this to significantly cut down on iteration time.
Taking advantage of CMake also lets lldb piggyback on existing support
for sanitizers -- or exciting new build configurations we don't know
about yet -- without having to update the Xcode project file.
rdar://problem/36751944
The reason to do this is that:
1. We already have a separate leaks bot.
2. We are not leaks clean at -O, but are ASAN clean at -O.
It doesn't make sense to limit our ability to verify that on Linux, we stay ASAN
clean at -O since we are not leaks clean at -O. Once we get leaks clean at -O,
we should create a separate bot.
rdar://31276806
* Moved argument-defaults logic that should be determined outside the actual parser back into the build-script. This logic will be relocated again at a later date to a more suitable module with testing.
* Moved test-paths argument default logic into build-script until it can find a better home in the future.
* Removed old default value for stdlib_deployment_targets from the expected options.
I recently broke the out of tree build by mistake [its fixed now ; )]. This
inspired me to make it easy to test this behavior by adding support to
build-script/cmake/etc for running an external benchmark build via
AddExternalProjects.
Now I can just call build-script with --build-external-benchmarks and thats it!
It should just work! It already helped me to avoid breaking the external build
twice!
I hope that eventually we get this on a bot to make sure it keeps working [or
even added to the smoke tests ; )].
*NOTE* This is disabled by default so it will not affect normal builds.
*NOTE* This just builds the external benchmarks. There is an rpath issue that
prevents you from running them (the benchmarks have the rpath set as if they are
next to the stdlib, but they are not. This can be fixed in a few different ways,
but I do not have time to finish implementing it = (. But this commit is a good
first step and at least detects build errors.
Sometimes it's useful to be able to run tests located in specific
directories and/or files, let's enable this in `utils/build-script`
using `--test-paths` option which accepts a list of viable test locations.
Resolves: rdar://problem/32004487
* Added new OptionalTrueAction and OptionalFalseAction classes to the arguments module.
* Fixed missing return statement in help-test generator.
* Renamed all the `skip_build_*` names to the more simple and affirmative `build_*`.
* Renamed all the `skip_test_*` names to the more simple and affirmative `test_*`.
* Added documentation to the OptionalTrueAction and OptionalFalseAction classes.
* Implemented a (mostly) comprehensive test suite for the argument parser (ab)using metaclasses to dynamically generate unit-tests for each valid argument and preset.
* Fixed failing defaults test for build_ninja.
* Added new HelpOption and IgnoreOption classes for generating tests.
* Converted default value test into individual generated tests which should give better output in case the default tests fail.
* Moved the default argument application logic to the new driver_arguments module in preparation for the larger argument parsing refactor.
* Re-ordered imports in order to satisfy python linter.
This only enables the swift compiler (not its output) to be used with leaks
sanitizer on linux.
Some important notes: On Linux, right now we are not completely leak clean. I
was only able to get a -Onone build of the stdlib without triggering lsan (I was
unable to run the optimizer and the tests successfully). Additionally even at
-Onone, I had to suppress some leaks in the driver. The point of this though is
to prevent any further -Onone leaks from being committed to the tree since
-Onone leaks that are not bounded (unlike the driver bugs) could cause SourceKit
to leak memory. Since SourceKit is a long running process... such a type of leak
would be bad.
rdar://32876901
In the past if one wanted to be able to develop using multiple swift repos, one
was often stymied by build-script always using build-script-impl from the swift
directory. This includes the case where one invokes the build-script from a
hypothetical directory called swift-2.
On Darwin platforms, the Swift build scripts generally use "xcrun" to
locate various tools, and the build-script's darwin-xcrun-toolchain option
is specified to xcrun. However, with the change to use the just-built clang
to build the runtime (PR #6112), there was an issue with the linker. Clang
normally tries to run the linker from the same directory as the clang
binary, but when there isn't one, it falls back on /usr/bin/ld, which is
just an xcrun-like wrapper. Since there is no way to specify an explicit
toolchain option in that case, set TOOLCHAINS in the environment to
make it use the linker from the specified toolchain.
rdar://problem/30709330
