Details
Diff Detail
- Repository
- rL LLVM
Event Timeline
clang/lib/Sema/SemaDeclCXX.cpp | ||
---|---|---|
2283 ↗ | (On Diff #208434) | This file seems untouched except for whitespace changes, let's revert them. |
clang/test/Sema/pragma-section.c | ||
57 ↗ | (On Diff #208434) | I agree it probably shouldn't, I imagine mingw users will want to be able to use this kind of code pattern to interop between GCC and Clang: __attribute__((section(".drective"))) const char LinkerFlags[] = "-export:foo -export:bar"; |
Thanks!
clang/test/Sema/pragma-section.c | ||
---|---|---|
57 ↗ | (On Diff #208434) | That's a nice argument, added as comment :) |
Thanks!
May I ask what inspired this? :)
I was looking through lld/COFF/Driver.cpp for something…ah, right, because D64156 touched it, and because I looked through all uses of getSpelling() for D64253, and noticed that LinkerDriver::parseDirectives() had an empty branch for OPT_natvis. I figured that was probably from before lld-link supported /natvis: and made a note to implement support for /natvis: in a .drectve (anticlimatic result of that sidequest was "link.exe doesn't allow /natvis in .drective", resulting in D64352), and I remembered #pragma data_seg(".drectve") before remembering #pragma comment(linker) and saw that cl.exe had this warning and I thought it was kind of useful.