walter-erquinigo wrote:

I have an idea for using lldb's process list logic, although it's a bit of code.

1. Don't use `"${command:pickProcess}"` at all. We need to bypass the regular 
command logic. You could use a wildcard like `${pickProcess}`.
2. In the debug config resolution logic, which is invoked before launching the 
debugger, you look for this wildcard. If you see it, then you programmatically 
invoke lldb-dap (you know its path at this point), and then invoke it with a 
new flag `--list-processes` that will just dump the list of processes in JSON 
format to stdout. With this information, you can show a custom made picker and 
await the result.
3. With the result, you can then modify the debug configuration and resume the 
debug process.

We shouldn't invoke lldb directly because we just don't know where it is, and 
it might be error-prone to try to find it. Instead, it might be better to 
simply add a flag to lldb-dap for this, which seems to be kind of appropriate 
for this kind of flow.

Let me know what you think.

https://github.com/llvm/llvm-project/pull/128943
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits

Reply via email to