gallopruss-sie wrote:
> I don't really know what the test was supposed to test, so I don't know if
> it's right or not. However, I don't really see the difference between not
> testing for the path prefix being there and overriding `-clang-resource-dir`,
> so it will always be there and nothing is really tested.
Yes, interesting point. The test was originally added in
de61aa3118b9bac85c468ea7ec40604a086744f5 which is about sysroot computation so
originally didn't care about this. I think it was later changed to check
resource paths in e0f22fe04a5c, so I think it does still need to test resource
paths.
An alternative would be to make the test less sensitive to the path (continuing
the direction from 1ab13f8cc3f7) (something like):
```
diff --git a/clang/test/Driver/riscv64-toolchain-extra.c
b/clang/test/Driver/riscv64-toolchain-extra.c
--- a/clang/test/Driver/riscv64-toolchain-extra.c
+++ b/clang/test/Driver/riscv64-toolchain-extra.c
@@ -28,7 +28,7 @@
// C-RV64-BAREMETAL-LP64-NOGCC: "-internal-isystem"
"{{.*}}/riscv64-nogcc/bin/../riscv64-unknown-elf/include"
// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-nogcc/bin/riscv64-unknown-elf-ld"
// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-nogcc/bin/../riscv64-unknown-elf/lib/crt0.o"
-// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-nogcc/{{.*}}/riscv64-unknown-unknown-elf/clang_rt.crtbegin.o"
+// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-unknown-unknown-elf/clang_rt.crtbegin.o"
// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-nogcc/bin/../riscv64-unknown-elf/lib"
-// C-RV64-BAREMETAL-LP64-NOGCC: "--start-group"
"{{.*}}/riscv64-nogcc/{{.*}}/riscv64-unknown-unknown-elf/libclang_rt.builtins.a"
"-lc" "-lgloss" "--end-group"
-// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-nogcc/{{.*}}/riscv64-unknown-unknown-elf/clang_rt.crtend.o"
+// C-RV64-BAREMETAL-LP64-NOGCC: "--start-group"
"{{.*}}/riscv64-unknown-unknown-elf/libclang_rt.builtins.a" "-lc" "-lgloss"
"--end-group"
+// C-RV64-BAREMETAL-LP64-NOGCC:
"{{.*}}/riscv64-unknown-unknown-elf/clang_rt.crtend.o"
```
Do you think that would be a better approach?
Thanks
Russ
https://github.com/llvm/llvm-project/pull/226952
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
