Revert host toolchain build change, merge master #1080
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…s correctly. Previously we would just forward the cleanup and create a normal "destroy" cleanup resulting in early deallocation and use after frees along error paths. As part of this I also had to tweak how DI recognizes self init writebacks to not use SILLocations. This is approach is more robust (since we aren't relying on SourceLocs/have verifiers to make sure we don't violate SIL) and also avoids issues from the write back store not necessarily have the same SILLocation. <rdar://problem/59830255>
LLVM's isel does not like integer types beyond a certain size (llvm::IntegerType::MAX_INT_BITS). rdar://63189452
…s with type param. assertions
The module name is not specified but the test explicitly overfits to an expected name. Adjust the test invocation to accommodate the expectation.
for call arguments. Consider variadic arguments 'skippable'. Also, don't treat 'VarargExpansionExpr' as a "context providing" expression. rdar://problem/62479469
static members cannot be used on protocol metatypes. rdar://problem/58886969
A lot of attributes are essentially default target configuration, and we should only differ when there's a good reason to. For the attributes we were already setting: - the ptrauth and target CPU/feature attributes are taken care of by Clang - I've updated the optsize/minsize attributes to the apparent intent - I've left the frame-pointer override in place for now Fixes rdar://63289339, which was caused by Swift's ptrauth IR attributes getting out of sync with Clang's.
…ubmap Serialization: Use checked substitution map deserialization for type nodes
…ess_huge_types IRGen: Fix enumPayload value witness emission for huge types
The ClangImporter currently calls into `ObjCSelector`'s `lookupDirect` in a couple of places, stashing the selector in a DenseMap to try and avoid re-entrancy problems. However this will become a problem once `ObjCSelector`'s `lookupDirect` is both requestified and starts pulling in members from the main module, so migrate the ClangImporter off calling it. Fortunately most of its uses only care about decls with associated Clang nodes. For those cases, we can use the existing member table, making sure to populate it with any method we import. In one case, the ClangImporter needs to check to see if there's a deserialized Swift method with a matching selector. Instead of calling through to `lookupDirect`, let's just query the Swift module loaders directly.
differentiation_activity_analysis.swift -> activity_analysis.swift
…when the declaration is synthesised (swiftlang#31915) * [Typechecker] Emit a specialized diagnostic for redeclaration errors when the declaration is synthesized * [Test] Update existing tests * [Typechecker] Diagnose the original wrapped property instead of the nearest non-implicit decl context * [Test] Update existing tests * [Typechecker] Do not diagnose redeclarations when both declarations are implicit * [Test] Update a AutoDiff test
…479469 [CodeCompletion] Handle variadic parameter in expr context analysis
…886969 [CodeCompletion] Stop suggesting static member on a protoco meta type
Use Clang's logic for adding the default IR attributes to a function
test: enable WebAssembly/WASI in lit.cfg
…1e6af034157dc9e8c1b31cf [silgen] Teach CleanupCloner how to handle OwnedValueWritebackCleanups correctly.
…in certain cases.
Don't column align PBD entries if any entry spans from the same line as
the var/let to another line. E.g.
```
// Previous behavior:
let foo = someItem
.getValue(), // Column-alignment looks ok here, but...
bar = otherItem
.getValue()
getAThing()
.andDoStuffWithIt()
let foo = someItem
.getValue() // looks over-indented here, which is more common.
getOtherThing()
.andDoStuffWithIt()
// New behavior
getAThing()
.andDoStuffWithIt()
let foo = someItem
.getValue() // No column alignment in this case...
doOtherThing()
let foo = someItem
.getValue(), // Or in this case (unfortunate, but less common)...
bar = otherItem
.getValue()
let foo = someItem.getValue(),
bar = otherItem.getValue() // but still column-aligned in this case.
```
Resolves rdar://problem/63309288
…am-3 GenericSignatureImpl, swiftlang#31712: Plug remaining relevant methods with type param. assertions
…-gyb Fix and reenable NewArray.swift.gyb
…r being bound to hole has a key path root representatee
rdar://20933449
…ts emitted from dependency JSON
runtime: repair the android build
Today, the driver does not propagate the flag to the frontend but also does not emit an error (silently consuming the flag instead). I'll open a similar PR for apple/swift-driver once this one is merged, in case there are any issues that need to be worked out first. Fixes SR-12834.
…ntil it can be fixed
…ly_definedin_backward_compatibility Disable attr_originally_definedin_backward_compatibility.swift test until it can be fixed.
[NFC][SwiftSyntax] Fix 'Traling' typo
…t instead from Swift Running shell commands using Swift in simulator tests is hard. We change the mechanism so that BuildModulesFromGraph.swift prints command line arguments and a simple python script picks these arguments and actually runs the command.
…-lone-get-issue [IDE][InterfaceGen] Always print the top-level decl in header file generated interfaces.
Further reduce the number of parallel jobs when testiong LLDB under ASAN
`argument_matching.swift` in test. New filename indicates that is also contains a test case for matching of unlabeled arguments.
from `diagnostics.swift` to `argument_matching.swift` which has more specific interest.
Propagate `-vfsoverlay` from driver to frontend.
build: hide symbols using CMake
…nter_non_leaf_only IRGen: Always eliminate frame pointers of leaf functions
…file [ConstraintSystem][NFC] Make test file named `argument_matching.swift`
DependenciesScanner: report command-line arguments for building pcms explicitly
…f-the-chain Delete ReferencedNameTracker
…-build # Conflicts: # test/lit.cfg
kateinoigakukun
approved these changes
May 24, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
#1065 is failing due to the lack of disk space on build runners, I assume that is caused by additional build artifacts coming from the host toolchain build. Let's revert #1024 and see if that helps. I have an idea how to get fresh host toolchains in a separate build run to save disk space, but that will come later as a separate PR.