jasonmolenda added a subscriber: jasonmolenda. jasonmolenda added a comment.
Maybe I'm an outlier here -- but I don't think we should show assembly code, unless specifically requested by the user, when we can't find a source file (but we have source-level debug information). lldb used to behave the other way - when a source file could not be found, we would show assembly before the prompt. There were many complaints from users about this behavior; users who are thinking at a source level and using source commands like "step" and "next", did not want to see assembly, it was meaningless to them. If they stepi into an assembly routine or put a breakpoint on it - where we have no debug information, then presenting the assembly makes sense. IIRC, gdb behaves the same way that lldb does today. e.g. clang -g a.c rm a.c lldb a.out (lldb) br s -n main (lldb) r - thread #1: tid = 0x5e8d52, 0x0000000100000ed7 a.out`main(argc=1, argv=0x00007fff5fbff958) + 39 at a.c:10, queue = 'com.apple.main-thread', stop reason = breakpoint 1.1 frame #0: 0x0000000100000ed7 a.out`main(argc=1, argv=0x00007fff5fbff958) + 39 at a.c:10 (lldb) n Process 38732 stopped - thread #1: tid = 0x5e8d52, 0x0000000100000eea a.out`main(argc=1, argv=0x00007fff5fbff958) + 58 at a.c:11, queue = 'com.apple.main-thread', stop reason = step over frame #0: 0x0000000100000eea a.out`main(argc=1, argv=0x00007fff5fbff958) + 58 at a.c:11 (lldb) I feel pretty strongly that this behavior should not be changed. Seeing mixed source and assembly is meaningful to me and I may personally like to see that -- but the vast majority of our users do not interact with their programs at that level. Repository: rL LLVM http://reviews.llvm.org/D12877 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
