averello/swift-mirror
1
0
Fork 0
mirror of https://github.com/apple/swift.git synced 2025-12-14 20:36:38 +01:00
Code Issues Packages Projects Releases Wiki Activity
53,276 Commits 622 Branches 2,942 Tags
f15cad360901b4cad3955b8da53d24a882e0e46f
Commit Graph

931 Commits

Author SHA1 Message Date
John McCall
338825e73d Fix the emission of r-value pointer conversions to delay the 
conversions and extend lifetimes over the call.

Apply this logic to string-to-pointer conversions as well as
array-to-pointer conversions.

Fix the AST verifier to not blow up on optional pointer conversions,
and make sure we SILGen them correctly.  There's still an AST bug
here, but I'll fix that in a follow-up patch.
 2017-04-26 14:15:44 -04:00
John McCall
f7e73ccc48 Delay the formal accesses associated with inout-to-pointer and 
array-to-pointer argument conversions until just before the call.

rdar://31781386
 2017-04-25 03:54:49 -04:00
John McCall
e6d52d8808 LValueToPointerExpr is dead code. RIP LValueToPointerExpr. 2017-04-25 03:01:16 -04:00
Joe Groff
d5cdf658da KeyPaths: Generate _kvcKeyPathString for ObjC-compatible keypaths. 2017-04-21 16:56:17 -07:00
Joe Groff
517c45aa3d Followups from merging master 2017-04-19 20:58:52 -07:00
Joe Groff
595e0e4ede Merge branch 'master' into keypaths 2017-04-19 18:38:24 -07:00
Joe Groff
4522cd09aa SILGen: Tweak key path computed property lowering to produce a consistent runtime-callable ABI. 2017-04-18 11:44:39 -07:00
practicalswift
7eb7d5b109 [gardening] Fix 100 typos. 2017-04-18 17:01:42 +02:00
Joe Groff
ee66ff35c5 SILGen: Lower computed property key path components. 2017-04-14 17:13:38 -07:00
Joe Groff
d6ced9d9e4 SIL: Rework KeyPathInst to accommodate computed components. 
- Separate out a uniquable KeyPathPattern that describes the context-free shape of the key path, with generic parameters and (eventually) subscript index slots factored out.
- Add component kinds for gettable and settable properties.
 2017-04-12 20:09:17 -07:00
Joe Groff
39a0849362 SILGen: Codegen for key path applications. 2017-04-09 16:38:34 -07:00
Joe Groff
964dc0e174 Sema: (wip) Overload resolution for keypath subscripts. 
TODO: Some diagnostic regressions:
test-macosx-x86_64/Compatibility/tuple_arguments.swift
test-macosx-x86_64/Constraints/diagnostics.swift
test-macosx-x86_64/Constraints/tuple_arguments.swift
test-macosx-x86_64/expr/unary/keypath/keypath.swift
test-macosx-x86_64/expr/unary/selector/selector.swift
 2017-04-09 16:38:02 -07:00
Joe Groff
d42f2049f7 KeyPaths: Implement in-place instantiation of invariant key paths. 
For key paths without generic or subscript parameterization, we can turn the compiler-generated key path pattern into a global object in-place.
 2017-04-05 08:46:45 -07:00
Joe Groff
f929c29bdf IRGen: Lower keypath instructions to patterns for the runtime to instantiate. 2017-04-04 11:31:15 -07:00
Joe Groff
18a6ca271a SILGen: Lower KeyPathExprs to KeyPathInsts. 2017-04-04 11:31:15 -07:00
Joe Groff
a5ca6ccd61 Restructure KeyPathExpr to allow more kinds of components. 
Expand the representation in anticipation of more kinds of components. NFC intended yet.
 2017-04-04 11:31:15 -07:00
Joe Groff
eb5d006e40 Rename ObjCKeyPathExpr -> KeyPathExpr. 
We can use the same general shape of expression for native key paths.
 2017-04-04 11:31:15 -07:00
Slava Pestov
f5b6f22ed6 Revert "[sil] Add a new API SILFunction::getTypeLowering(SILType)." 
The new API is broken. Popping a generic context frees all
dependent type lowerings, so this function returns a pointer
to freed memory.

This reverts commit 24dfae0716.
 2017-04-02 19:55:19 -07:00
John McCall
305c94fc5b Pare down the Initialization interface and allow implementations 
to insert code before performing an in-place initialization.

Intended to be NFC, but some of the clients needed more rewriting
than others.
 2017-03-31 18:55:09 -04:00
Slava Pestov
8fe8b89b0f SIL: Terminology change: [fragile] => [serialized] 
Also, add a third [serializable] state for functions whose bodies we
*can* serialize, but only do so if they're referenced from another
serialized function.

This will be used for bodies synthesized for imported definitions,
such as init(rawValue:), etc, and various thunks, but for now this
change is NFC.
 2017-03-29 16:47:28 -07:00
Michael Gottesman
4dacf68848 [silgen] Make sure that we do not disable ownership along failure paths when performing select enum checks. 
What is happening here is that we are performing a select_enum check and then
not emitting the cleanup on the path where we know that our optional value is
null. To express such an operation in semantic sil, one should use a switch_enum
instead. Thus we must insert the destroy_value so that we know that the optional
value is balanced along all paths.

rdar://31145255
 2017-03-26 10:47:12 -07:00
Joe Shajrawi
9790333d06 Support contextual init of opaque optional types 2017-03-22 17:30:49 -07:00
Joe Shajrawi
b9ee609369 Support casting of optional to optional for opaque values 2017-03-22 17:08:33 -07:00
Joe Shajrawi
a03f38c3c5 Add support for if statements for opaque value(s) under new mode 2017-03-22 12:48:01 -07:00
Michael Gottesman
b99c47c006 [silgen] Scope argument cleanups with a new scope called "ArgumentScope". 
Once this is in, I will be able to finish the SILGenApply part of Semantic SIL.

rdar://30955427
 2017-03-18 17:03:41 -07:00
Slava Pestov
64adf7495e SILGen: Change ArchetypeOpenings map to be keyed by ArchetypeType * and not CanType 2017-03-14 17:46:54 -07:00
John McCall
897f5ab7c5 Restore CanType-based micro-optimizations. 
This reverts commit 5036806e5a.
However, it preserves a pair of changes to the SIL optimizer
relating to walking through optional types.
 2017-03-14 11:38:11 -04:00
Michael Gottesman
213ddd753f [silgen] Compute the ResultPlan outside of SILGenApply so that we can untangle indirect result/normal argument lifetime scopes. 
rdar://30955427
 2017-03-14 00:39:53 -07:00
Michael Gottesman
4552355ff2 [silgen] Refactor SILGenFunction::apply to take a new struct called CalleeTypeInfo. 
This structure contains all of the type information that we use to build a
ResultPlanPtr. In a subsequent commit, I am going to move ResultPlanPtr creation
out of emitApply and place it before the creation of arguments. This is to
enable indirect result initializations to belong to the scope outside of any
argument based scopes.

Without this the lifetimes of the indirect result initializations and arguments
can not be separated without using hacks. There is no reason why we can't emit
the indirect result temporaries before we emit any arguments since they will
remain dormant until after the final apply/any future committed argument scopes
being popped.

rdar://30955427
 2017-03-14 00:39:41 -07:00
Slava Pestov
5036806e5a AST: Remove some unnecessary getCanonicalType() calls 2017-03-13 02:24:36 -07:00
Joe Groff
99ea154151 Merge pull request #7965 from jckarter/invariant-load 
Add a `loadInvariant` builtin.
 2017-03-09 06:39:42 -08:00
Joe Groff
39ecc53a25 Add a loadInvariant builtin. 
Lowers to an invariant load in LLVM; probably useful for SIL too at some point too, but NFC at that level yet.
 2017-03-08 21:02:03 -08:00
Michael Gottesman
a68975a21b [semantic-sil] Fix a place where we are forwarding the underlying cleanup vs creating a new one when performing upcast. 
rdar://29791263
 2017-03-08 16:45:17 -08:00
swift-ci
26ca022053 Merge pull request #7984 from gottesmm/tuple_init_fixes 2017-03-08 15:24:59 -08:00
Michael Gottesman
c643c7ba87 [silgen] Fix tuple initialization for semantic sil. 
rdar://29791263
 2017-03-08 15:06:05 -08:00
Slava Pestov
76eb5dd66d SILGen: Clean up ad-hoc SubstitutionList construction when calling intrinsics 
Change emitApplyOfLibraryIntrinsic() to take a SubstitutionMap,
and use the correct abstractions to build the map.

This gets rid of the last remaining uses of gatherAllSubstitutions()
in SIL.
 2017-03-08 13:54:29 -08:00
Michael Gottesman
87ca878870 [silgen] Fix another tuple explosion -> borrow + copy since we do not have the destructure operation yet. 
rdar://29791263
 2017-03-05 11:44:11 -08:00
Michael Gottesman
80e297d6aa [silgen] Go through and fix up places where during bringup of load_borrow I just put in load_borrow without any end_borrow cleanups. Now use real cleanups. 
rdar://29791263
 2017-03-02 19:53:33 -08:00
Joe Shajrawi
4b6cfaf533 Merge pull request #7796 from shajrawi/deinit_existential 
Support for deinit of opaque existentials: deinit_existential_opaque instruction + SILGen support
 2017-02-27 15:43:32 -08:00
Joe Shajrawi
16b6cb5e1d Support for deinit of opaque existentials: deinit_existential_opaque instruction + SILGen support 2017-02-27 14:46:43 -08:00
Mikio Takeuchi
488d531846 Enhance -assume-single-threaded option (SR-3945) 2017-02-27 12:17:53 +09:00
Michael Gottesman
784d5d16fc [silgen] Change emitClassConstructorInitializer to use ownership and make calling designated/chaining initializes use proper ownership. 
rdar://29791263
 2017-02-23 08:48:58 -08:00
Michael Gottesman
5920daa373 [silgen] Refactor getAddressForInPlaceInitialization onto SGFContext(). 2017-02-19 10:58:43 -08:00
Michael Gottesman
d051583378 [semantic-sil] When emitting a base for an accessor use a load_borrow when appropriate. 
This builds on 92bf345. I am very pleased that a couple of the cases I needed to
update, a retain/release was eliminated on the base.

rdar://29791263
 2017-02-18 00:55:36 -08:00
Michael Gottesman
78a514cdb1 [silgen] Use the name FormalAccess to refer to state related to accesses in a single formal evaluation context. 2017-02-17 19:19:35 -08:00
Arnold Schwaighofer
ae0b6b1e06 Merge pull request #7186 from aschwaighofer/cow_exist_silgen_support 
SIL/SILGen support for copy-on-write existentials
 2017-02-15 17:41:10 -08:00
Michael Gottesman
d4ae7a3f8a [semantic-sil] When calling emitRValueForDecl, borrow the value before deciding whether or not to copy the value. 
rdar://29791263
 2017-02-15 15:29:30 -08:00
Michael Gottesman
58c3959e90 [silgen] Remove ManagedBorrowedValue in favor of the usage of FormalEvaluationScopes. 
rdar://29791263
 2017-02-15 15:28:14 -08:00
Arnold Schwaighofer
ec802fe9cd SILGen: Thread AccessKind to openExistential() 
We will eventually create different open_existential_addr instructions depending on the kind of access to the opened value
 2017-02-15 14:22:18 -08:00
Slava Pestov
41eba98902 Gardening: Fix some unused variable warnings in no-assert builds 2017-02-15 12:57:35 -08:00