Partition implementations are special, they generate a CMI, but it
does not have an 'export' line, and we cannot export anything from the
it [that is it can only make decls available to other members of the
owning module, not to importers of that].
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
clang/test/Modules/cxx20-10-1-ex1.cpp | ||
---|---|---|
8–9 | In my implementation, I replace ':' to '-' in pcm filename. I do so since I remember this is the behavior of GCC. I think '-' is better than '_' since '_' is possible to occur in name. And I am surprised that the compiler wouldn't complain about that it couldn't find the corresponding pcm. Since I don't see corresponding code so far. |
address review comments, rebase
clang/test/Modules/cxx20-10-1-ex1.cpp | ||
---|---|---|
8–9 | The filename could/should be completely decoupled from the module name; changing the filenames to something randomly chosen should make no difference to the tests passing. I made a comment on why/how this works in 3/8 in reply a similar comment made there (essentially the pairing of module-name and filename is something known by the lmodule loader and/or the header search mechanisms - it is not something that Sema needs to know). | |
22 | we are all hostage to some style decisions made by others? |
llvm::unreachable ("how did we get a partition type set?");