Kill ArgumentShuffleExpr #23672
Merged
Kill ArgumentShuffleExpr #23672
Conversation
This file contains 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
slavapestov
force-pushed
the
kill-argument-shuffle-expr
branch
2 times, most recently
from
91ea9a8 to
70b23aa
Compare
beccadax
reviewed
Mar 30, 2019
| @@ -5428,7 +5339,7 @@ inline bool ApplyExpr::validateArg(Expr *e) const { | |||
| else if (isa<BinaryExpr>(this)) | |||
| return isa<TupleExpr>(e); | |||
| else | |||
| return isa<ParenExpr>(e) || isa<TupleExpr>(e) || isa<ArgumentShuffleExpr>(e); | |||
| return isa<ParenExpr>(e) || isa<TupleExpr>(e); | |||
There was a problem hiding this comment.
I've been waiting for this day since I first added ApplyExpr::validateArg(). 🎉
Instead of building ArgumentShuffleExprs, lets just build a TupleExpr, with explicit representation of collected varargs and default arguments. This isn't quite as elegant as it should be, because when re-typechecking, SanitizeExpr needs to restore the 'old' parameter list by stripping out the nodes inserted by type checking. However that hackery is all isolated in one place and will go away soon. Note that there's a minor change the generated SIL. Caller default arguments (#file, #line, etc) are no longer delayed and are instead evaluated in their usual argument position. I don't believe this actually results in an observable change in behavior, but if it turns out to be a problem, we can pretty easily change it back to the old behavior with a bit of extra work.
slavapestov
force-pushed
the
kill-argument-shuffle-expr
branch
from
70b23aa to
478c133
Compare
|
@swift-ci Please smoke test |
|
@swift-ci Please test source compatibility |
|
Nice! |
|
@rjmccall FWIW, I realized I might want to bring back the ArgSpecialDest stuff... if would help tidy up enum constructor call emission with payloads. |
|
I was wondering about that. |
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.
This completes a refactoring that simplifies our call argument expressions.
In Swift 5, we used a rather complex TupleShuffleExpr to represent tuple conversions as well as call arguments containing varargs and default arguments. This included array literals, which were a degenerate case of varargs.
In #23591 I split up TupleShuffleExpr into DestructureTupleExpr (for tuple conversions) and ArgumentShuffleExpr (for argument lists containing default arguments and varargs).
However ArgumentShuffleExpr was still far too general. Ever since https://github.com/apple/swift-evolution/blob/master/proposals/0060-defaulted-parameter-order.md was accepted for Swift 3.0, we always evaluate arguments in the order they're written in the source, so the "shuffling" part was never taken advantage of.
Recently @pschuh implemented a way to form arrays directly in SIL by lowering an ArrayExpr, without using an ArgumentShuffleExpr. This work landed in #23618.
This meant that Sema could now collect varargs into a single ArrayExpr, without using an ArgumentShuffleExpr. The only remaining case where an ArgumentShuffleExpr was needed was argument lists with default arguments.
I did some preliminary refactoring in #23650, to ensure that all argument lists are split up into individual argument expressions when building the PreparedArguments object.
This PR builds on top of that by introducing a DefaultArgumentExpr and CallerDefaultArgumentExpr. Now Sema knows how to build argument lists containing ArrayExpr for varargs, and DefaultArgumentExpr and CallerDefaultArgumentExpr for default arguments. SILGen's ArgEmitter now knows how to directly emit delayed arguments, without going through the ArgumentShuffleExpr code path.
Note that after this change, all CallExprs now have a ParenExpr or TupleExpr as their argument. So the next step is 'flattening' the representation so that a CallExpr can hold multiple argument expressions.