Note: the FIR: equivalent for this converts the i1 to a fir.logical<4>.
I expect this is a bug in HLFIR expression lowering, but a separate bug
to what is added by this patch.
Details
Details
- Reviewers
jeanPerier clementval
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Comment Actions
I am working on this TODO because it is also firing from some IO statement in SPEC and just uploaded a patch that handle the logical case that was indeed the reason I left this a TODO, the type need to be taken from the origin expression. https://reviews.llvm.org/D143421
I took the liberty of adding your test to the patch.