rorth wrote: I guess I have to spell things out once again. Here's what I've done: Ran 4 builds each on Solaris/amd64 and Linux/x86_64, one pair with the `projects` layout, another with the `runtimes` layout, both with and without this patch. Here's what I find (listing the target triple used, the `liborc_rt` location, run ``` ClangReplInterpreterTests --gtest_filter=OutOfProcessInterpreterTest.FindRuntimeInterface ``` files searched and test output, test result:
Solaris/amd64: - target `amd64-pc-solaris2.11` (this is what the buildbot is running), `projects` layout, `sunos/liborc_rt-x86_64.a` unpatched: `FAIL` ``` lib/clang/23/lib/amd64-pc-solaris2.11/libclang_rt.orc_rt.a lib/clang/23/lib/amd64-pc-solaris2.11/liborc_rt.a ``` patched: `PASS` ``` Interpreter/OutOfProcessInterpreterTests.cpp:187: Skipped Host does not support out-of-process JIT ``` - target `x86_64-pc-solaris2.11`, `runtimes` layout, `x86_64-pc-solaris2.11/liborc_rt.a` both patched an unpatched: `PASS` ``` [==========] 0 tests from 0 test suites ran. (0 ms total) ``` - target `x86_64-pc-linux-gnu`, `projects` layout, `linux/liborc_rt-x86_64.a` unpatched: `FAIL` ``` lib/clang/23/lib/x86_64-unknown-linux-gnu/libclang_rt.orc_rt.a lib/clang/23/lib/x86_64-unknown-linux-gnu/liborc_rt.a ``` patched: `PASS` ``` Interpreter/OutOfProcessInterpreterTests.cpp:187: Skipped Host does not support out-of-process JIT ``` - target `x86_64-pc-linux-gnu`, `runtimes` layout, `x86_64-unknown-linux-gnu/liborc_rt.a` unpatched and patched: `PASS` ``` [==========] 0 tests from 0 test suites ran. (0 ms total) ``` As you can see, there's absolute no difference in behaviour between Linux and Solaris, as I've been saying all this time. However, I'm beyond caring now: this patch at least makes the failure go away! https://github.com/llvm/llvm-project/pull/176198 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
