Disable loading scalars as addresses when materializing an entity variable #1452
Conversation
|
@swift-ci test |
|
@augusto2112 Can you create the same pull request for master-next? And can you also make one for swift/release/5.3? I'd like to at least discuss the option of taking this for the 5.3 branch. |
|
@augusto2112 It would be great if you could add the test you showed me on IRC that explicitly showcases this. It was very simple, and we should have it in the test suite. |
augusto2112
force-pushed
the
class-constrained-protocol-eval
branch
from
d97db90
to
fdac8f9
Compare
|
@adrian-prantl could you take a look and see if the test is all right? |
lldb/test/API/lang/swift/expression/self_from_register/TestSelfFromRegister.py
Outdated
Show resolved
Hide resolved
lldb/test/API/lang/swift/expression/self_from_register/TestSelfFromRegister.py
Outdated
Show resolved
Hide resolved
lldb/test/API/lang/swift/expression/self_from_register/main.swift
Outdated
Show resolved
Hide resolved
lldb/test/API/lang/swift/expression/self_from_register/TestSelfFromRegister.py
Outdated
Show resolved
Hide resolved
augusto2112
force-pushed
the
class-constrained-protocol-eval
branch
from
fdac8f9
to
b8ee781
Compare
|
@swift-ci test |
|
@adrian-prantl It looks like the Linux test timed out: |
|
@swift-ci test |
|
If it happens again I'd be more suspicious. |
augusto2112
force-pushed
the
class-constrained-protocol-eval
branch
2 times, most recently
from
b8ee781
to
4599655
Compare
|
@adrian-prantl I thought we merged this 😮 |
augusto2112
force-pushed
the
class-constrained-protocol-eval
branch
from
4599655
to
45b9f0d
Compare
augusto2112
force-pushed
the
class-constrained-protocol-eval
branch
from
45b9f0d
to
b1cb186
Compare
|
@adrian-prantl Looks like the changes got to master in the 5.3 version. All that's left here is the removal of the obsolete |
|
@swift-ci test and merge |
|
Can you do master-rebranch and master-next, too? |
|
@swift-ci test |
This PR disables the incorrect loading of scalars as addresses in the context of the entity variable's materializer, which caused problems when materializing a variable from a register. This also enables a test which was affected by this bug on Linux.
@adrian-prantl @vedantk although this bug manifested on Linux, I believe it could affect other platforms as well, right? Also, the failing test happened to catch this behavior, but it wasn't really its intended purpose. Do you think it'd be worthwhile to craft a new test to check this behavior specifically?