apple / swift

…zationInfo-race

[sourcekit] Prevent race between getInitializationInfo and requests

Request handling happens on a global queue, so we need to use a
barrier_async on the same queue to ensure the initialization finishes
before requests are handled.

test: recover Driver/opt-record.swift

[test] Disable some tests failing in specific configurations pending investigation

Pending investigation: rdar://71857770

…ed stdlib

Pending investigation rdar://71857262

[test] Temporarily disable AutoDiff tests failing with optimizations

These recently started failing when optimized. Disable while we
investigate and fix.

rdar://71642726

[test] Update SILOptimizer/OSLogFullOptTest.swift

After recent changes we have eliminated one of the retain/release pairs.

rdar://71641343

Fix test/IRGen/objc_generic_class_stub.swift again

[test] Disable test failing rarely in CI until we can fix it

rdar://71468441

…ogether

[sourcekit] Make it possible to install both sourcekitdInProc and sourcekitd efficiently

We only need the C++ symbols on Darwin, so avoid the issue of different
manglings.

…e when possible

When installing both InProc and XPC versions of sourcekitd, dynamically
link the XPC service to the InProc version. This lets us install both
frameworks without wasting the disk space from having another copy of
the swift frontend statically linked.

Makes it possible to build and install both sourcekitdInProc and the XPC
service on Darwin.

[test] Fix test/SourceKit/CursorInfo/invalid_offset

This fixes a layering violation where UID handling was provided by a
dependent library, creating a dependency cycle if you tried to link the
lower level libraries dynamically.

This refactors notification handling so that the XPC service and InProc
library pass in their notification handler callback. This fixes a
layering issue where the sourcekitd libraries could not be linked
dynamically due to the missing symbols.

This test was intended to catch a crash on invalid offset, but in rare
cases it was failing spuriously, because the error message depends on
non-deterministic behaviour. It's sufficient for this test that it
doesn't crash.

rdar://63187529

[code-completion] Fix cache writer for non-decl kind

Thanks to Ben Barham for spotting this:
sizeof(~static_cast<uint8_t>(...)) is 4, but we need to write a single
byte here. Thankfully this code was probably not being hit in the
current caching scheme, which only caches declarations.

Looking at the history, this code was broken by d8fbaa0, which was
fixing an MSVC warning in this code. Unfortunately I do not have access
to the version of MSVC to check if there is still a warning here or for
what.

[gardening] Move some code between sourcekitd and libIDE

…ce3000

[sourcekitd] Fix potential race in semantic highlighting tests

In order to make range-shifting for semantic highlighting testable,
disable returning semantic information during an "open" request. This
has no real value anyway, since it only happens very rarely, and it
makes testing range shifting impossible to do deterministically.

rdar://problem/66386179