This was failing alongside inherits_ObjCClasses.swift (they were failing at the
same time).
Until MikeA has time to look at this disable this as well.
rdar://problem/50898688
...instead of crashing. Also drop the class if its generic
requirements depend on a type that can't be loaded (instead of
crashing).
rdar://problem/50125674
A number of tests exercise features only available in Apple OSes that
shipped with Swift 5.0 in the OS; this includes the following versions:
- macOS 10.14.4
- iOS 12.2
- tvOS 12.2
- watchOS 5.2
Previously these tests were restricted to running on macOS only, with
an explicit -target x86_64-apple-macosx10.14.4. To get better test
coverage, add a new %target-stable-abi-triple substitution which
expands to a triple with the correct OS version on all Apple platforms.
On non-Apple platforms, this is the same as %target-variant-triple,
but for now any test that uses this exercises Apple platform features
anyway.
One caveat is that since iOS 12.2 does not have a 32-bit slice, we
have to skip any tests that use -target %target-stable-abi-triple
on this platform. A new swift_stable_abi feature flag can be tested
with 'REQUIRES: swift_stable_abi'. To get maximum test coverage,
I split off a 'stable_abi' version of a few tests that build with both
an old and new deployment target. This allows the old deployment
target case to still be tested on 32-bit iOS.
Check the availability of decls that declare an opaque return type to ensure they deploy to a
runtime that supports opaque types.
rdar://problem/50731151
Adds a NumStoredPropertiesQueries stat.
Adds a test case for an increasing number of lazy stored class
properties. Each property requires a formal access within the
initializer. This would manifest as cubic behavior in
AccessEnforcementOpts, which scales as O(1.5) in the above stat.
These can be recreated if needed in a client library. To do this, I've
added a new ConformanceLookupKind::NonInherited, which can also be
used elsewhere in the project where we're already filtering out
inherited conformances some other way.
Note that this doesn't drop inherited conformances from the entire
serialized interface, just from the list that a class explicitly
declares. They still get referenced sometimes.
rdar://problem/50541451 and possibly others
We use the just-built clang without any -target or -sdk flags. Keep things
simple and limit the test to macOS for now.
Fixes <rdar://problem/50586614>.
The objc_loadClassref() symbol is weak linked from Clang now, so we can
un-XFAIL this test. Note that we have to pass a special flag to Clang
since objc_loadClassref() is not in the SDK's TBD files yet.
Also, change the test to use its own asserts instead of FileCheck since
there's no output unless a new libobjc is available.
Opaque result type archetypes can involve type variables, which
then get introduced into GenericSignatureBuilders and the
generated GenericSignatures. Allocate them in the proper arena
So we don’t end up with use-after-free errors.
Fixes rdar://problem/50309503.
Some tests are limited to only Linux, when they should also pass for
Android.
Additionally, InputStream.swift.gyb was disabled for Android ARMv7, but
wasn't for Android AArch64, which allow me to find the error on it and
fix it on #24521.
Finally, StringLowercasedUppercased is interesting in Android because it
checks the used ICU is correct for performing the tasks that the stdlib
needs.
The REQUIRES written one in each line were impossible to satisfy in any
platform, so the test wasn't running at all. This was changed to a
REQUIRE-ANY listing all the platforms (and adding Android). I also
needed to add runAllTest() at the end to make the test pass.
Fixes crashes in 28437-swift-typechecker-validatedecl.swift (from previous commit)
and the compiler crasher 28861-gpdecl-getdepth-generictypeparamdecl-
invaliddepth-parameter-hasnt-been-validated.swift.
Android is one of those platforms that define the symbol EOF as a
negative number. Trying to turn it into an UInt8 will only crash the
test. Move the transformation after we have checked for EOF first. In my
opinion, in other platforms, having the EOF attached to the getline
result was also invalid, and that's probably why the two test that use
EOF were sending an extra newline.
Even with this, the tests do not pass on Android because it deadlocks
waiting for input/output from the child process, which never happens. I
haven't find why this happens, but only happens if the last test of the
suite closes stdin. I will try to fix that in another PR.
Until the issue is fixed. Tracked by:
<rdar://problem/49791522> Swift CI Test failures: IRGen/autorelease_optimized_armv7/aarch64.
https://bugs.swift.org/browse/SR-10474
