enormous .xsession-errors.old

[marmarta]

Qubes OS version
R4.1

Affected component(s) or functionality
everything

Brief summary
icon-receiver gets stuck in a loop of errors and produces 121GB .xsession-errors.old file. (or bigger. That's just my pool size).

Traceback (most recent call last):
  File "/usr/lib64/python3.8/asyncio/events.py", line 81, in _run
    self._context.run(self._callback, *self._args)
  File "/usr/lib/qubes/icon-receiver", line 259, in handle_pending_events
    for ev in iter(self.conn.poll_for_event, None):
  File "/usr/lib/python3.8/site-packages/xcffib/__init__.py", line 569, in wrapper
    self.invalid()
  File "/usr/lib/python3.8/site-packages/xcffib/__init__.py", line 559, in invalid
    raise ConnectionException(err)
xcffib.ConnectionException: xcb connection errors because of socket, pipe and other stream errors.

[rmol]

I'm seeing this too, but only with XFCE. It doesn't happen for me with i3. I've put together a PR that helps, by having icon-receiver retry the xcb connection, so eventually (after 6-7 seconds on my system, usually) the situation resolves and the log spam stops. The fact that it seems to be XFCE-specific suggests that there might be more that could be done to correct it -- adjusting ordering/timing of autostart tasks, maybe, or introducing a delay either when icon-receiver starts, or when errors are encountered? I'm just not familiar enough with the pieces in play here; I ran into it tangentially when testing other things under 4.1.

[github-actions]

This issue is being closed because:

• This issue is believed to affect only Qubes OS 4.1 (and possibly earlier).
• Qubes OS 4.1 has reached end-of-life (EOL).

If anyone believes that this issue should be reopened, please leave a comment saying so.
(For example, if a bug still affects Qubes OS 4.2, then the comment "Affects 4.2" will suffice.)