-
Notifications
You must be signed in to change notification settings - Fork 4.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[release/9.0-staging] Ignore modopts/modreqs for UnsafeAccessor
field targets
#109709
[release/9.0-staging] Ignore modopts/modreqs for UnsafeAccessor
field targets
#109709
Conversation
The generated IL remains correct and doesn't require a .volatile prefix for the generated accessor. This is based off of what Roslyn currently emits if a getter was available.
Tagging subscribers to this area: @dotnet/area-system-runtime-compilerservices |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm. we will take for consideration in 9.0.x
If it helps support this, this issue is also blocking ComputeSharp on .NET 9 when using primary constructors. I noticed it there because my CI runs started failing when I tried to bump the TFM of my samples to .NET 9 (see here). |
/ba-g "WASM timeouts not detected" |
4b94961
into
release/9.0-staging
Backport of #109694 to release/9.0-staging
/cc @AaronRobinsonMSFT
Customer Impact
Customer reported issue: #109665. This was originally found in a MSBuild BinLog reader scenario. The specific issue is fields that are marked as
volatile
in C# aren't discoverable when usingUnsafeAccessor
s and as such if someone writes the following, it is unable to be accessed.This can't be discovered because the
volatile
in C# is embedded via a modreq and there is no place to describe that using theUnsafeAccessorAttribute
.The general issue is that when discovering the target field member, we currently require a complete match for fields. One could imagine we add some support for this, but given they are generally considered compiler directives, there is little value in considering them in this lookup scenario and so they should be ignored on fields. This is already done in the method case, but the relaxation was missed on field discovery.
Regression
Early in the .NET 9 release, the logic for signature comparrison was rewritten and regressed some unsupported issues and has uncovered a regression in a supported scenario.
Testing
Tests for the particular issue were added. I reached out to the Roslyn compiler team and verified the added test case is the only scenario where this occurs.
Risk
Low This is a permissive change on the way fields are looked-up and their types compared. Since C#/VB.NET don't provide other modopts/modreqs it is unlikely this will impact any other scenario.