When building standalone, since llvm-strip is a symlink, it is created
using add_custom_command/add_custom_target which cannot be exported, and
thus cannot be depended on by lldb.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Comment Actions
@JDevlieghere has been touching similar things today. You should coordinate with him on this change.
lit/CMakeLists.txt | ||
---|---|---|
64 ↗ | (On Diff #223703) | why not if(TARGET llvm-strip)? I think that expresses the intent more cleanly (and conforms to the existing pattern). |
lit/CMakeLists.txt | ||
---|---|---|
64 ↗ | (On Diff #223703) | I actually ran into an issue with that today, we had the following code in the top-level CMake list: if(TARGET dsymutil) add_lldb_test_dependency(dsymutil) endif() Nevertheless, ninja lldb-test-deps didn't build dsymutil. $ /V/J/l/build-ra> ninja lldb-test-deps [1/1] Python script sym-linking LLDB Python API $ /V/J/l/build-ra> ninja dsymutil [1/1] Linking CXX executable bin/dsymutil |
Comment Actions
I think I'm guilty for adding llvm-strip to LLDB_TEST_DEPS I wasn't aware that it might cause problems.
lit/CMakeLists.txt | ||
---|---|---|
64 ↗ | (On Diff #223703) | My slight preference for this is to actually have a good way to identify that this can be simplified once unified builds are the only supported build style. I suppose a comment along with the if(TARGET) check would work too. |