Merge swift-DEVELOPMENT-SNAPSHOT-2019-10-13-a into tensorflow #27821
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
…space/time efficient than -g on that platform. This improves time spent to link products built with debug info quite a bit.
This gives us a one-stop shop for making changes to the override mechanism. rdar://problem/54752086
Vtable thunks were incorrectly not marked as thunks, which means (among other things) they're a false positive in the NonInlinableFunctionSkippingChecker.
`AutoDiffIndexSubset` is a fixed-size bit vector that is used for efficiently representing a subset of indices in automatic differentiation, specifically for representing a subset of parameters and results of a function to differentiate with respect to. It is uniqued in `ASTContext`. This patch adds definition and unit tests for `AutoDiffIndexSubset` along with new files `AutoDiff.h` and `AutoDiff.cpp` into the 'AST' target, with no changes to the compiler's behavior. More data structures used for AutoDiff will be added to these files. ---------------------------- This is part of the ongoing effort to merge the experimental [differentiable programming feature](https://forums.swift.org/t/differentiable-programming-mega-proposal/28547) (informally referred to as "AutoDiff") to the 'master' branch for code reviews and better maintenance. Upstreaming task: [TF-879](https://bugs.swift.org/browse/TF-879)
lit runs its tests with only the environment variables set up by its configuration, which is great for repeatability, but sometimes you want to allow a few controlled leaks. With this change, you can now specify “--param copy_env=FOO:BAR:BAZ” to cause the environment variables FOO, BAR, and BAZ to be visible in lit tests during that run.
Now that llvm::iterator_range has 'empty', there's not enough reason to keep our own version of it in the Swift repo. No functionality change.
…oduct description
If we're going to get rid of swift::IteratorRange, let's make llvm::iterator_range easy to use. No functionality change.
When an EnumElementDecl is parsed, we create the parameter list before creating the EnumElementDecl itself, so we have to re-parent those ParamDecls just like we do for functions and subscripts.
…parameters
Currently single parameter tuple splat fix/diagnostic supports only
cases where parameter is a concrete tuple type, let's enhance that to
support generic parameters as well e.g.:
```swift
func foo<T>(_: T) {}
foo(0, 1, 2) // `T` expects arguments to form a tuple e.g. `foo((0, 1, 2))`
```
In situations like this:
```swift
func foo(x: (Int, Int) {}
foo(x: 0, 0)
```
Left paren to form a missing tuple should be placed after
the label because belongs to the parameter and not the tuple.
Correctly set the parent DeclContext of ParamDecls in EnumElementDecls
Optimize for libSyntax parsing. The new 'parseListSyntax()' receives a vector storage, passes element builder to the callback for populating the elements, then automatically parses trailing commas. Also, it performs automatic recovery if the element has an error (e.g. Skip until ',' or ')').
(Implemented by Nathan Hawes @nathawes)
(implemented by Nathan Hawes @nathawes) Advance \p Loc to the last non-missing token of the specified or, if it doesn't contain any, the last non-missing token preceding it in the tree.
For non-migrated expression / decl attributes. Clean-up the slot after the use. Assert there's no existing element at the location.
In AST token range, string literal token must be treated as a single token.
rdar://54906120
[SILGen] Add the IsThunk bit to vtable thunks
…neric parameter Since it's hard to say whether tuple use is really intended, let's add `did you mean to pass a tuple?` note to the error message.
Fix property wrapper crasher
[PropertyWrappers] Look for constructors with default args when finding default init
Fixes rdar://problem/54352235
[tests] Add mechanism to override lit environment
…ps-6 Circular validation cleanups, part 6
Be more accepting of vendor patterns. Any name can be added by configuration, so loosen the test.
…kup. Swift 5.1's lookup for custom attributes skipped associated type members, which allowed code like the given example to compile. To maintain source compatibility, identify the narrow case that happens in practice---the property wrapper is at module scope but is now shadowed by an associated type---warn about it, and accept it. Fixes rdar://problem/56213175.
disallow in-tree builds
DebugInfo: loosen producer pattern
…urce-compat [Property wrappers] Fix source compatibility issue with attribute lookup
Remove time references, they do not add to the instructions.
Update WindowsBuild.md
…erge2 Tag build swift-DEVELOPMENT-SNAPSHOT-2019-10-13-a Simple merge including conflicts. Conflicts: include/swift/AST/ASTContext.h include/swift/AST/IndexSubset.h include/swift/Parse/Parser.h lib/AST/ASTContext.cpp lib/AST/CMakeLists.txt lib/AST/IndexSubset.cpp lib/Sema/TypeCheckDecl.cpp unittests/AST/CMakeLists.txt unittests/AST/IndexSubsetTests.cpp
Note: this was done as a distinct commit in order to make it easier to see what's actually going on.
dan-zheng
reviewed
Oct 21, 2019
| @@ -469,30 +469,30 @@ | |||
| "tensorflow": { | |||
| "aliases": ["tensorflow"], | |||
| "repos": { | |||
| "llvm": "swift-DEVELOPMENT-SNAPSHOT-2019-10-08-a", | |||
| "llvm": "swift-DEVELOPMENT-SNAPSHOT-2019-10-13-a", | |||
There was a problem hiding this comment.
I noticed that commit hashes for clang for lldb have not updated - I wonder if that is intentional (perhaps there was nothing new to merge)?
There was a problem hiding this comment.
Thanks for the catch @dan-zheng ! I tried to update them, but those repos are frozen due to the monorepo transition. So I'm going to leave them frozen for now. Does that sound okay?
There was a problem hiding this comment.
If the toolchain builds and passes tests without updating those repos, then we got lucky and that seems fine to me.
If we realize that we do need updates in those repos, options are:
- Merge past the monorepo transition, transition ourselves to the monorepo, and make the updates in the monorepo.
- Temporarily upload the updates to personal forks of the repos and point update-checkout-config there. After we transition to the monorepo, this temporary hack disappears.
|
@swift-ci please test tensorflow |
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.
Merging!