How to debug with the new documentation?

[noklam]

https://pygls.readthedocs.io/en/latest/howto/use-the-pygls-playground.html

There are new instructions for differetion VSCode versions now:

1. VSCode v1.89+
2. VSCode v1.88 and older

For 1, It doesn't seem to work for me when I add debugger I see the server get restarted but the breakpoint is not triggered (LSP working normally).

For 2, am I suppose to launch a debugger separately? and how do I attach the debugger properly?

p.s. I have autosave enable so it's quite annoying everytime I type something in a file it restart my LSP.

[noklam]

I do remember that I used to borrow configuration from somewhere else, so it may not be the change of documentations. (unfortunately I lost my launch.json)

But I still struggle with debugging following https://pygls.readthedocs.io/en/latest/howto/use-the-pygls-playground.html, I have "pygls.server.debug": true

I am using the hover.py example, this is what I experienced.

[alcarney]

When it comes to debugging, there shouldn't be any difference between 1+2 (as it's the same extension), just the steps to get VSCode into a state where you can enable the playground itself are different.

Of course that's the theory, there maybe something I've overlooked 😅

The good(?) news is that I see the same behaviour as your gif, almost everything is working except that the breakpoints aren't working. Removing the following lines from .vscode/launch.json fixed the issue for me (I'm using VSCode v1.89)

  "pathMappings": [
      {
          "localRoot": "${workspaceFolder}",
          "remoteRoot": "."
      }
  ],

Does that work for you? (I really don't understand what this option does, or why we added it in the first place...)

p.s. I have autosave enable so it's quite annoying everytime I type something in a file it restart my LSP.

Yes, I can imagine! 😅, it should be easy enough to add an option to the playground to disable that.

[noklam]

Does that work for you? (I really don't understand what this option does, or why we added it in the first place...)

I suspect this is related but I am also very confused with how debugging work in VSCode in general.

I have two different errors:

1. Sometimes it's looking for the wrong interpreter even if I set it properly.

Exception has occurred: FileNotFoundError
[Errno 2] No such file or directory: '/Users/Nok_Lam_Chan/.local/lib/python3.10'

2. Sometimes it's a weird Blocking error

Exception has occurred: BlockingIOError
[Errno 35] Resource temporarily unavailable
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/asyncio/selector_events.py", line 115, in _read_from_self
    data = self._ssock.recv(4096)
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/asyncio/events.py", line 80, in _run
    self._context.run(self._callback, *self._args)
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/asyncio/base_events.py", line 1909, in _run_once
    handle._run()
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/asyncio/base_events.py", line 603, in run_forever
    self._run_once()
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/asyncio/base_events.py", line 636, in run_until_complete
    self.run_forever()
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/site-packages/pygls/server.py", line 258, in start_io
    self.loop.run_until_complete(
  File "/Users/Nok_Lam_Chan/dev/demo/pygls/examples/servers/hover.py", line 87, in <module>
    server.start_io()
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "/Users/Nok_Lam_Chan/miniconda3/envs/lsp/lib/python3.10/runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
BlockingIOError: [Errno 35] Resource temporarily unavailable

I cannot make it work yet, but this settings work for my extension, borrow from the Microsoft template.
https://github.com/kedro-org/vscode-kedro/blob/main/.vscode/launch.json

I notice they have this configuration

            "env": {
                "USE_DEBUGPY": "True"
            },

[alcarney]

Sometimes it's looking for the wrong interpreter even if I set it properly.

That's annoying, I know @tombh has had issues like that in the past, I wish I knew why... unfortunately/fortunately for me, these things tend to just work for me, so I don't have many ideas on what to try :/

So that we can focus on the debugging bit first, you can try setting the pygls.server.pythonPath option in .vscode/settings.json to skip the interpreter selection code for now.

Sometimes it's a weird Blocking error

That is weird... is that something you only see when trying to debug? Or can it happen in non-debug mode also?

I notice they have this configuration

"env": {
  "USE_DEBUGPY": "True"
},

Looking at the code, I think that is an option specific to the template, looks like they launch a dedicated debug script...