Teach SubstitutionMap::isIdentity about non-canonical generic params #25767
Conversation
|
Can you think of any other interesting test cases that I didn't include? @swift-ci Please test |
|
Build failed |
lib/AST/SubstitutionMap.cpp
Outdated
| return; | ||
|
|
||
| unsigned ordinal = sig->getGenericParamOrdinal(paramTy); | ||
| auto replacement = getReplacementTypes()[ordinal]->getCanonicalType(sig); |
There was a problem hiding this comment.
I think it would be clearer if you just did Type(paramTy).subst(this)
There was a problem hiding this comment.
That's not the same check, is it? That would also accept SubstitutionMaps that didn't have an entry for that parameter.
There was a problem hiding this comment.
SubstitutionMaps are always "complete". There's an entry for each parameter
lib/AST/SubstitutionMap.cpp
Outdated
| }); | ||
|
|
||
| return !hasNonIdentityReplacement && | ||
| countOfGenericParams == getReplacementTypes().size(); |
There was a problem hiding this comment.
Is it ever possible for countOfGenericParams to not equal getReplacementTypes().size()? Maybe it should be an assertion instead?
There was a problem hiding this comment.
I was wondering that, but I didn't know if it was safe. All valid SubstitutionMaps should behave like that, but what about a degenerate one?
|
Build failed |
jrose-apple
force-pushed
the
mistaken-identity
branch
from
d2d6631 to
5b8eeb7
Compare
|
@swift-ci Please test |
|
Build failed |
|
Build failed |
jrose-apple
force-pushed
the
mistaken-identity
branch
from
5b8eeb7 to
c1f08ec
Compare
|
@swift-ci Please smoke test |
Do a weaker check here that only looks at the canonical generic params and guarantees that *those* substitute to themselves. There may be replacement types for other generic params too, to canonicalize them, but that's not a problem. This fixes a crash trying to mangle decls with opaque result types that have generic signatures that canonicalize away a generic parameter. rdar://problem/51775857
jrose-apple
force-pushed
the
mistaken-identity
branch
from
c1f08ec to
d46aa9a
Compare
|
@swift-ci Please smoke test |
…wiftlang#25767) Do a weaker check here that only looks at the canonical generic params and guarantees that *those* substitute to themselves. There may be replacement types for other generic params too, to canonicalize them, but that's not a problem. This fixes a crash trying to mangle decls with opaque result types that have generic signatures that canonicalize away a generic parameter. rdar://problem/51775857 (cherry picked from commit 8157ccf)
Do a weaker check here that only looks at the canonical generic params and guarantees that those substitute to themselves. There may be replacement types for other generic params too, to canonicalize them, but that's not a problem.
This fixes a crash trying to mangle decls with opaque result types that have generic signatures that canonicalize away a generic parameter.
rdar://problem/51775857