[Lldb-commits] [lldb] Fix a thinko in the CallSite handling code: (PR #114896)

[via lldb-commits]

jimingham wrote:

This patch seems to fail on x86_64 Linux, but I don't understand how this patch could cause the reported failure.  The failure is in a lldb-dap test, here:

File "/var/lib/buildkite-agent/builds/linux-56-59b8f5d88-9dkps-1/llvm-project/github-pull-requests/lldb/test/API/tools/lldb-dap/evaluate/TestDAP_evaluate.py", line 104, in run_test_evaluate_expressions

Comparing the log to a successful one, all the breakpoints got set on the same lines.  Then we ran and are stopped at the first breakpoint - which parallels a successful test.   We also passed a bunch of self.assertEvaluate's to print various variables.  Then we go to run the "var" command in "repl" mode:

```
        if context == "repl":
            # In the repl context expressions may be interpreted as lldb
            # commands since no variables have the same name as the command.
            self.assertEvaluate("var", r"\(lldb\) var\n.*")

```

and get no reply.  Then we try to kill the DAP process, and that also seems to fail.

I can't reproduce this locally, and I can't see how this failure could result from this patch...

I'm going to apply this patch and keep an eye on the x86_64 bots to see if this was just a glitch or not.

https://github.com/llvm/llvm-project/pull/114896