Details
Diff Detail
Unit Tests
Time | Test | |
---|---|---|
360 ms | linux > HWAddressSanitizer-x86_64.TestCases::sizes.cpp |
Event Timeline
clang/lib/Sema/SemaDeclCXX.cpp | ||
---|---|---|
5887–5915 | I wonder if this code block could be simplified by calculating two booleans:
We want to mark the vast majority of C++ method decls referenced. The conditions I see here are:
I tried to come up with a simpler description of those conditions, and couldn't come up with one. Nevermind. I don't think my suggestion will make the code any simpler, but I'll leave this as a comment for your consideration. | |
5898 | Part of me wants to handle explicitly defaulted things separately from implicit special members, so we don't have to check for explicitly defaulted-ness twice. | |
clang/test/CodeGenCXX/dllexport.cpp | ||
929 | Let's also add a test case where the constructor is explicitly defaulted outside the body of the class. From reading the code, I know we will take the isUserProvided path instead, but I'd like to have it for completeness. |
clang/lib/Sema/SemaDeclCXX.cpp | ||
---|---|---|
5898 | Thanks, that makes it a bit nicer. |
clang/lib/Sema/SemaDeclCXX.cpp | ||
---|---|---|
5906–5907 | should this grow a assert(TSK != TSK_ExplicitInstantiationDefinition) for symmetry? |
clang/lib/Sema/SemaDeclCXX.cpp | ||
---|---|---|
5906–5907 | Hmm no, we can still have TSK==TSK_ExplicitInstantiationDefinition here, it's just for explicitly defaulted methods where we don't want to pass it to the consumer. |
Part of me wants to handle explicitly defaulted things separately from implicit special members, so we don't have to check for explicitly defaulted-ness twice.