On Fri, 8 Jun 2018 at 18:48, Leonard Mosescu <mose...@google.com> wrote: > > So what's your take on this particular case? As far as I can see (please > correct me if I'm wrong), the LIT-only tests: > > 1. Don't cover the case where a function is split into disjoint regions, > right?
For this particular case, I don't think having a split function is particularly interesting. Neither the code under test or the test itself do anything special with the functions, the just play around with sequences of PCs and line numbers. I definitely agree a discontinuous function is an interesting case, but not in the context of this test/bug. > 2. Also don't cover the cross-targeting case (ex. the native PDB reader > hosted on Linux) This is not correct. clang and lld are perfectly capable of producing exes and pdbs on linux. The only reason we don't run these tests on other platforms is because our pdb reader works only on windows (this would be great to fix, btw). > 3. A bug in LLD might inadvertently make the tests pass w/o testing what they > are supposed to > (let's say that it incorrectly ingores the hardcoded order and lays > everything contiguously) I would hope lld has tests for that. Taking a step back, I think the place we disagree is on the purpose of these tests. I view these tests as regression tests, and I don't expect this test to validate that lldb works with a particular toolchain. I expect it to check that it works for a particular kind of input. In this case, that input was a file with discontinuous line tables. To me, a linker order file (of any linker) sounds like a good abstraction level for generating that kind of input (on linux, I might have preferred a .s file with hardcoded .loc directives, but that doesn't seem to be a thing on windows). For an integration-level test that checks our integration with say MSVC 2017, I'd recommend getting dotest-style tests running with that toolchain. _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
