SE - Nullable: Add Null constraint for empty nullable constructor
#6853
Conversation
| @@ -67,10 +67,15 @@ public static IEnumerable<SymbolicContext> Process(SymbolicContext context) | |||
| { | |||
| if (Simple.TryGetValue(context.Operation.Instance.Kind, out var simple)) // Operations that return single state | |||
| { | |||
| context = context.WithState(simple.Process(context)); | |||
| return new[] { context.WithState(simple.Process(context)) }; | |||
There was a problem hiding this comment.
Unrelated improvement. If an operation is registered in the 1st list, there's no reason look in the 2nd one too
There was a problem hiding this comment.
Do we have some kind of test or check, so that we don't end up with one operation kind in both lists?
There was a problem hiding this comment.
No, we don't. It should be easy to catch, as the change goes for "simple" first, so whatever complex logic we'll have should just not work. And should be easy to catch
github-actions
bot
assigned martin-strecker-sonarsource and unassigned pavel-mikula-sonarsource
| @@ -67,10 +67,15 @@ public static IEnumerable<SymbolicContext> Process(SymbolicContext context) | |||
| { | |||
| if (Simple.TryGetValue(context.Operation.Instance.Kind, out var simple)) // Operations that return single state | |||
| { | |||
| context = context.WithState(simple.Process(context)); | |||
| return new[] { context.WithState(simple.Process(context)) }; | |||
There was a problem hiding this comment.
Do we have some kind of test or check, so that we don't end up with one operation kind in both lists?
| { | ||
| const string code = """ | ||
| int? explicitType = new Nullable<int>(); | ||
| int? targetTyped = new(); |
There was a problem hiding this comment.
Propably later:
void M<T>() where T: struct
{
T? a = new T(); // NotNull (this should come from me)
T? b = new T?(); // Null (this should come from you)
}There was a problem hiding this comment.
This already works - so I've added those cases
| """; | ||
| var validator = SETestContext.CreateCS(code).Validator; | ||
| validator.ValidateTag("ExplicitType", x => x.HasConstraint(ObjectConstraint.Null).Should().BeTrue()); | ||
| validator.ValidateTag("TargetTyped", x => x.HasConstraint(ObjectConstraint.NotNull).Should().BeTrue("new() of int produces value 0")); |
There was a problem hiding this comment.
😮 Interesting behavior.
There was a problem hiding this comment.
Indeed. It took me a debug session and a manual test to verify. definitely deserves a special message :)
github-actions
bot
assigned pavel-mikula-sonarsource and unassigned martin-strecker-sonarsource
pavel-mikula-sonarsource
force-pushed
the
Pavel/SE/03-NewNull
branch
from
07e9718 to
4e747eb
Compare
|
Kudos, SonarCloud Quality Gate passed!
|
|
Kudos, SonarCloud Quality Gate passed! |
Part of #6812
new Nullable<int>()should haveNullconstraint