Add convenience method for exception reporting

[ahopkins]

#2724 adds a new signal to trigger on all exceptions: "server.lifecycle.exception"

This PR adds a convenience for setting up a task for capturing this signal to do something with it. The obvious use case is for error reporting that does not need to be in the way of returning a response.

@app.report_exception
async def catch_any_exception(app: Sanic, exception: Exception):
    print("Caught exception:", exception)

Also, since it is still an unreleased signal, this PR renames it to "server.exception.report", which is I think is a less confusing name since it is not strictly tied to server lifecycle.

Lastly, in exploring this earlier PR, I realized that there was a bug (double dispatch on some built-in signals that happened inline in a request), and I also realized that we were not capturing exceptions in background tasks. This PR handles these scenarios as well.

[codecov]

Codecov Report

Patch coverage: 84.615% and project coverage change: -0.116 ⚠️

Comparison is base (31d7ba8) 88.911% compared to head (8b83306) 88.795%.

Additional details and impacted files

@@              Coverage Diff              @@
##              main     #2792       +/-   ##
=============================================
- Coverage   88.911%   88.795%   -0.116%     
=============================================
  Files           92        92               
  Lines         7034      7051       +17     
  Branches      1195      1197        +2     
=============================================
+ Hits          6254      6261        +7     
- Misses         534       540        +6     
- Partials       246       250        +4     

Impacted Files	Coverage Δ
sanic/signals.py	91.275% <66.666%> (+0.058%)	⬆️
sanic/app.py	88.870% <86.956%> (-0.865%)	⬇️

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.