[Lldb-commits] [lldb] [lldb] Change lldb's breakpoint handling behavior (PR #96260)

[Pavel Labath via lldb-commits]

labath wrote:

Thanks for the explanation, Jason. I believe I understand the problem now. Here, we're concerned with whether the physical breakpoint opcode got executed, and not about the higher level machinery (conditional breakpoints, ignore count, ...), which builds on top of that and can cause a breakpoint hit to "disappear". Since StopInfo tracks the "logical" stop reason of the process (after all of these filters are applied), using it to track the "physical" one would be tricky at best.

> Looking at the total requirements, the rule can be condensed to: If this thread stopped at a BreakpointSite which it did not execute, we should execute the breakpoint on Resume. In any other case, a thread sitting at a BreakpointSite should silently step past it and resume execution.

So, IIUC, there would just be one field (the "unexecuted breakpoint site") instead of two. I'll take that, I think having just one field to think about would make it easier to understand the overall logic.

https://github.com/llvm/llvm-project/pull/96260