Add a Fix-It to the warning about unnecessary @preconcurrency conformance
#73742
Conversation
…rmance When we diagnose an unnecessary `@preconcurrency` on a conformance, also provide a Fix-It to remove the `@preconcurrency`.
DougGregor
requested review from
ktoso,
AnthonyLatsis,
hborla,
slavapestov and
xedin
as code owners
|
@swift-ci please smoke test |
| : RootProtocolConformance(ProtocolConformanceKind::Normal, | ||
| conformingType), | ||
| Protocol(protocol), Loc(loc), Context(dc) { | ||
| Protocol(protocol), Loc(loc), PreconcurrencyLoc(preconcurrencyLoc), | ||
| Context(dc) { |
There was a problem hiding this comment.
Could you please add an assert to make sure that preconcurrencyLoc is never valid when isPreconcurrency is set to false?
| @@ -147,7 +147,7 @@ SourceLoc TypeRepr::findAttrLoc(TypeAttrKind kind) const { | |||
| for (auto attr : attrTypeRepr->getAttrs()) { | |||
| if (auto typeAttr = attr.dyn_cast<TypeAttribute*>()) | |||
| if (typeAttr->getKind() == kind) | |||
| return typeAttr->getAttrLoc(); | |||
| return typeAttr->getStartLoc(); | |||
There was a problem hiding this comment.
I was looking for what this change is going to affect and found https://github.com/apple/swift/blob/main/lib/Sema/TypeCheckDeclPrimary.cpp#L1803-L1806 which instead of 1 should also use strlen("@retroactive ")? We have a few tests for that diagnostic but none of them check fix-it :/
There was a problem hiding this comment.
Sure, I'll do that. I feel like there should be a better way than that strlen, but I wasn't up for adding more infrastructure right now.
There was a problem hiding this comment.
What devilry is this? The .getAdvancedLoc(1) thing is actually working for @retroactive, and I don't understand why
When we diagnose an unnecessary
@preconcurrencyon a conformance, also provide a Fix-It to remove the@preconcurrency.