Commit Graph

21 Commits

Author SHA1 Message Date
John McCall
3d9ac50e2c Teach MetadataReader how to skip artificial subclasses,
and improve RemoteAST to provide corresponding APIs.
2017-02-16 15:41:50 -05:00
John McCall
a8e9ad3d0d Gate the swift-remoteast-test tests on the target matching the host.
swift-remoteast-test uses the JIT to execute code in its own process.
Even if we theoretically taught the interpreter to perform some sort
of remote interpretation, that would not be sufficient to make
swift-remoteast-test work.  Therefore I've introduced a new lit
feature check specifically for swift-remoteast-test, but set it by
default to be equivalent to swift_interpreter.
2016-11-08 11:38:23 -08:00
Dmitri Gribenko
d175b3b66d Migrate FileCheck to %FileCheck in tests 2016-08-10 23:52:02 -07:00
Joe
3938d5682a [SE-0095] [Runtime], [Demangler], & AST printer updated to new composition syntax
- All parts of the compiler now use ‘P1 & P2’ syntax
- The demangler and AST printer wrap the composition in parens if it is
in a metatype lookup
- IRGen mangles compositions differently
    - “protocol<>” is now “swift.Any”
    - “protocol<_TP1P,_TP1Q>” is now “_TP1P&_TP1Q”
- Tests cases are updated and added to test the new syntax and mangling
2016-07-19 12:01:37 -07:00
Joe
a6dad0091b [SE-0095] Initial parsing implementation for '&' composition syntax
This commit defines the ‘Any’ keyword, implements parsing for composing
types with an infix ‘&’, and provides a fixit to convert ‘protocol<>’

- Updated tests & stdlib for new composition syntax
- Provide errors when compositions used in inheritance.
Any is treated as a contextual keyword. The name ‘Any’
is used emit the empty composition type. We have to
stop user declaring top level types spelled ‘Any’ too.
2016-07-19 12:01:02 -07:00
John McCall
0273167b5d Use the known type parameters to determine static offsets to stored
properties of classes with generic layouts.

Previously we were falling back on accessing them via the field
offset vector even when we knew everything about the type.

As a minor benefit, this allows RemoteAST to also determine offsets
for members of classes with generic layout.

Half of the test changes are IR type-name uniquing; I'm going to
explore mangling these with the full type where possible.
2016-04-29 15:33:05 -07:00
John McCall
7366917787 Test that classes work.
Oh look, apparently they don't for generic classes.
2016-04-28 16:39:56 -07:00
John McCall
f33c84fb3d RemoteAST: implement getOffsetOfMember for structs and classes. 2016-04-28 16:29:20 -07:00
John McCall
83dc7cffe8 RemoteAST: implement member-offset calculations for tuples. 2016-04-27 19:00:30 -07:00
John McCall
fc8636f9c4 Test RemoteAST for various existential types. 2016-04-26 13:58:38 -07:00
John McCall
92cf4a6d39 Teach MetadataReader how to handle ObjCClassWrapper metadata. 2016-04-26 13:04:02 -07:00
John McCall
627e2325d3 Read tuple labels in MetadataReader and preserve them in RemoteAST.
The metadata system doesn't actually unique based on labels
correctly, so the test case has to play some games.  That's
something that will be easier to fix when there are fewer
clients poking at the internals of the metadata runtime.
2016-04-25 22:33:41 -07:00
John McCall
56052cfe61 Apparently the right way to look up imported declarations is to just ask the importer to look things up. 2016-04-25 19:03:08 -07:00
John McCall
812557ca24 Reorganization within RemoteAST. Basic error propagation.
Initial stabs towards reading foreign class metadata, although
these do not yet successfully resolve the declaration.
2016-04-25 16:57:19 -07:00
John McCall
38712ae277 A few more generic tests. 2016-04-22 21:34:51 -07:00
John McCall
5a735b5589 RemoteAST: type-check TypeReprs within a TopLevelCodeDecl
instead of just a ModuleDecl.

Fixes simple generic types.
2016-04-22 18:47:04 -07:00
John McCall
014bfa10f7 Handle MetadataKind::Class in readMetadata. This can only
come up as the exact value of the metadata kind for classes
where ObjC interop is disabled.
2016-04-21 23:23:47 -07:00
John McCall
c0566914ca Temporarily XFAIL this on Linux while I figure out what's
different there.
2016-04-21 22:15:16 -07:00
John McCall
d472f89c07 Flesh out nominal type testing a bit better. 2016-04-21 21:51:27 -07:00
John McCall
590706a8c6 Pass around address/buffer pairs in the MetadataReader instead of
just the address.  Use this to avoid repeatedly reading metadata so
often (even if it's cached, this is better).  Fix some bugs involving
nominal type parents.
2016-04-21 21:47:14 -07:00
John McCall
f8fcd0ce11 Implement the RemoteAST test harness and add some skeletal tests. 2016-04-21 16:25:19 -07:00