Silence EMPTY_LAYOUT warnings

[maximlt]

The EMPTY_LAYOUT warning is emitted when an empty layout (e.g. pn.Row()) is rendered. Panel use cases are different than Bokeh's ones, Panel is more used to serve interactive apps where it's perfectly valid to render an empty layout that is later populated with some content. In this case this warning is just noise.

The ability to silence specific Bokeh warnings was added to Bokeh a few years ago bokeh/bokeh#8739. This PR is using the silence function to silence this warning, for both panel serve and pn.serve() usages.

[codecov]

Codecov Report

Merging #4056 (f5b5e6c) into master (f083d88) will increase coverage by 0.00%.
The diff coverage is 92.30%.

@@           Coverage Diff           @@
##           master    #4056   +/-   ##
=======================================
  Coverage   83.54%   83.54%           
=======================================
  Files         223      223           
  Lines       33193    33219   +26     
=======================================
+ Hits        27730    27754   +24     
- Misses       5463     5465    +2     

Flag	Coverage Δ
ui-tests	37.57% <30.76%> (-0.02%)	⬇️
unitexamples-tests	74.56% <92.30%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
panel/tests/test_server.py	99.44% <88.88%> (-0.37%)	⬇️
panel/command/serve.py	35.74% <100.00%> (+1.39%)	⬆️
panel/io/server.py	75.09% <100.00%> (+0.14%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[mattpap]

I would prefer to remove that warning (even as soon as now, i.e. 3.0). I always found it annoying that we warn about a perfectly valid behavior, which is counter productive at best.

[maximlt]

I've seen a rather long discussion on Bokeh about this warning, with some being in favor of removing it, and some being in favor of keeping it, but that was some years ago so opinions may have changed!

[mattpap]

Here's bokeh/bokeh#8725 the discussion. I will submit a PR.

[philippjfr]

Still going to merge this but thankfully we can then remove it again on the 3.0 branch.