[Lldb-commits] [lldb] [lldb-dap] Fix raciness in launch and attach tests (PR #137920)

[John Harrison via lldb-commits]

ashgti wrote:

To clarify the DAP flow a little, see 'Launch Sequencing' in https://microsoft.github.io/debug-adapter-protocol/overview

Once the DAP server send the response to the `initialize` request the following happen in parallel:

* The `initialized` event triggers the client sending `setBreakpoints`, etc. with the `configurationDone` request at the end of the chain.
* `launch` or `attach` requests are sent and we send back a `process` event.

The flow from VSCode at least often ends up with:

* `initialize seq=1`
* `launch seq=2`
* `setBreakpoints seq=3`
* `configurationDone seq=4`

But these do not strictly have to be in that order.

The DAP server can respond to these events asynchronously but today the lldb-dap is single threaded for handling each of these responses.

Looking at the failing test https://lab.llvm.org/buildbot/#/builders/18/builds/15174/steps/6/logs/stdio, we had the following sequence:

```
{"command":"initialize", ...}
{"command":"attach", ...}
{"command":"setBreakpoints", ...}
{"command":"configurationDone", ...}
{"event":"stopped", ...}
```

But that stopped event here is wrong. This attach did not set `"stopOnEntry": true`, so that should have continued the process at point.

https://github.com/llvm/llvm-project/pull/137920