It didn't properly set "withInlineStorage(false)" on the test value witness table, causing it to pick direct-storage value witnesses that overflow a fixed-size buffer and crash on the ASan bot. Fix the test, and add a check for a canary that fails even without ASan.
Swift SVN r30462
These will be used for reflection, and eventually to speed up generic
operations on single payload enums as well.
Progress on <rdar://problem/21739870>.
Swift SVN r30214
We incorrectly tested the uninitialized "next" pointer against MAP_FAILED, instead of the real result of mmap. Fixes rdar://problem/21659505.
Swift SVN r30030
@objc protocols aren't supported with an ObjC runtime, but we still want values of AnyObject type to be word-sized. Handle this by turning the binary "needsWitnessTable" condition into a "dispatch strategy" enum, so we can recognize the condition "has no methods, so neither swift nor objc dispatch" as distinct from either swift or ObjC protocol representations. Assign this dispatch strategy when we lower AnyObject. Should be NFC for the ObjC-enabled build.
(It would also be beneficial for the ObjC-runtime-enabled version of Swift if AnyObject weren't an @objc protocol; that would mean we could give it a canonical protocol descriptor in the standard library, among other things. There are fairly deep assumptions in Sema that AnyObject is @objc, though, and it's not worth disturbing those assumptions right now.)
Reapplying with updates to the runtime unit tests.
Swift SVN r27341
These should all apply to any Darwin platform, and the current behaviour
was breaking cross-compilation.
Introduces a SWIFT_DARWIN_VARIANTS pattern, to be used as follows:
if(SWIFT_HOST_VARIANT MATCHES "${SWIFT_DARWIN_VARIANTS}")
Also fix one place where I checked the CMAKE_SYSTEM_NAME instead of the
host variant that I recently introduced. I haven't attempted to find
the rest of the places we're doing this though.
Swift SVN r26554
If an existential type for a special protocol (not a composition) is instantiated, carry the special protocol identifier from that protocol to the existential, allowing us to easily recognize existentials with unique runtime characteristics.
Swift SVN r26436
- Add frontend and standard library build support for tvOS.
- Add frontend support for watchOS.
watchOS standard library builds are still disabled during SDK bring-up.
To build for TVOS, specify --tvos to build-script.
To build for watchOS, specify --watchos to build-script (not yet supported).
This patch does not include turning on full tests for TVOS or watchOS, and
will be included in a follow-up patch.
Swift SVN r26278
We can't do this because the object was deallocated. However, ASan should flag us if we regress on touching deallocated unowned-referenced objects though, so we don't need this. rdar://problem/19677703
Swift SVN r24894
ObjC unowned references are backed by an ObjC weak reference which will eagerly deallocate the object when it's strongly released, so in an unknown-refcount situation, we can't safely dereference the object pointer to determine its Swiftness. We can, however, look at the side table of weak references; if there's an entry for this object, then it's reliably an ObjC object (or it's some other object that got allocated in the reclaimed space for the dead object, but that's a race we fundamentally can't win with this broken design). Fixes rdar://problem/18091547 (modulo the aforementioned reallocation race).
Swift SVN r24825
Instead of directly referencing a long-dead constructor from the standard library, put a purpose-built function in StdlibUnittest to feed a Swift object reference to the test C code. rdar://problem/18498737
Swift SVN r24816
Metatype metadata now always refers to the maximally-abstracted @thick metatype representation, so metatypes are never trivial from the runtime's point of view.
Swift SVN r21975
constructs generic class metadata properly when the
superclass has an unexpected prefix.
Also fix a bug in the constructor for ClassMetadata (which is
only actually constructed this way in unit tests).
Swift SVN r20586
unexpected forematter from the superclass.
This requires a pretty substantial shift in the
generic-metadata allocation/initialization dance
because (1) we can't allocate class metadata without
knowing what the superclass is and (2) the offset
from the metadata cache entry to the address point is
no longer determined solely by the metadata pattern.
While I'm making invasive changes to metadata, fix
two race conditions in metadata creation. The first
is that we need to ensure that only one thread succeeds
at lazily creating a generic-metadata cache. The second
is that we need to ensure that only one thread actually
attempts to create a particular metadata; any others
should block until the metadata is successfully built.
This commit finishes rdar://17776354. LLDB will
need to adjust to the runtime-private metadata layout
changes.
Swift SVN r20537
a flags field, add an instance address point field, and reserve
some additional space.
This change must be coordinated with a corresponding change
to ObjC runtime bits in libarclite; without this, dynamic
subclassing features like KVO will break.
The actual contents of the new fields can change without
bothering the ObjC runtime.
Swift SVN r20183
I've put these fields on the class object for now, just
so we can at least theoretically update them. A superclass
that grew left rather than right could maybe even be made
to work with this schema, but probably not.
rdar://16705821
Swift SVN r16880
