clayborg added a comment. In D110827#3042820 <https://reviews.llvm.org/D110827#3042820>, @kpdev42 wrote:
> In D110827#3034767 <https://reviews.llvm.org/D110827#3034767>, @clayborg > wrote: > >> Should we be testing if these directories exist before trying to use them? >> Might be nice to avoid compiler warnings if the compiler will emit warnings >> or errors if these directories don't exist. > > I think testing if these directories exists may be a little bit redundant > because clang ignores non-existent include paths passed with -I no worries then >> LLDB also tests with compilers that were built, like when LLDB builds clang >> and uses that clang and clang++ that it built to run the test suite. If we >> had settings in LLDB that users could set, then the test suite would be able >> to use the include files for the compiler that is being used instead of >> always defaulting to the system headers. > > Could you please clarify: "LLDB builds clang" - here you mean clang which was > build with LLDB? And I would like to mention that starting from > https://reviews.llvm.org/D89013 libcxx puts __config_site header to target > specific folder We often build the full clang compiler during LLDB builds and then use the clang we built as the compiler when running our test suite. So anything we can do to make sure what ever clang we use for building the test suite binaries has all of the headers that it was built with along with any support library headers (libcxx, etc) that would be great. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D110827/new/ https://reviews.llvm.org/D110827 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
