argentite added inline comments.
================ Comment at: clang/test/Interpreter/dynamic-library.cpp:1 +// RUN: head -n 7 %s | %clang -xc++ -o %T/libdynamic-library-test.so -fPIC -shared - +int ultimate_answer = 0; ---------------- v.g.vassilev wrote: > sgraenitz wrote: > > The use of `head` and `tail` here is a creative solution, but I wonder how > > well it scales for similar cases in the future in case this becomes a role > > model. We have this situation a lot in LLDB: compile an input file and use > > the output in a RUN line. It typically uses separate input files for such > > purposes (note that "Inputs" folders must be excluded from test discovery > > somewhere in lit config), e.g.: > > https://github.com/llvm/llvm-project/blob/release/16.x/lldb/test/Shell/Breakpoint/jit-loader_jitlink_elf.test > > > > What do you think? > I agree with that. I've seen that some clang tests surround the code in > question with `#ifdef LIB` and then in the invocation they add `-DLIB`. The `#ifdef` approach does not seem to work with clang-repl yet. It does not like unterminated `#ifdef`. So I am going with separate input file. "Inputs" folder seems to be excluded globally in `clang/test/lit.cfg.py`. ================ Comment at: clang/test/Interpreter/dynamic-library.cpp:9 + +// REQUIRES: host-supports-jit, system-linux +// RUN: tail -n 16 %s | env LD_LIBRARY_PATH=%T:$LD_LIBRARY_PATH clang-repl | FileCheck %s ---------------- sgraenitz wrote: > This requirement applies to the entire file right? The arrangement here may > imply the opposite. Moved it to the top of the file. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D141824/new/ https://reviews.llvm.org/D141824 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
