closes #73.

In Dash if you try to do this you get an output like:

Address already in use
Port 8080 is in use by another program. Either identify and stop that program, or start the server with a different port.

So we mimic that here.

Hi @jayceslesar!

Thanks for contributing, looks promising! 👏🏼 I did not immediately find the dash implementation; could you provide a link to which snippet you based this implementation on?

TODO:

• [ ] Review and compare with dash implementation
• [ ] Add tests

Codecov Report

Merging #79 (ea4c3aa) into main (f25cc4e) will decrease coverage by 0.38%. The diff coverage is 57.14%.

@@            Coverage Diff             @@
##             main      #79      +/-   ##
==========================================
- Coverage   98.11%   97.73%   -0.39%     
==========================================
  Files          10       10              
  Lines         743      750       +7     
==========================================
+ Hits          729      733       +4     
- Misses         14       17       +3     

Impacted Files	Coverage Δ
...tly_resampler/figure_resampler/figure_resampler.py	86.56% <57.14%> (-3.44%)	:arrow_down:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update f25cc4e...ea4c3aa. Read the comment docs.

also looks like this is the second time the failing test has failed on different python versions while the other two passed haha. Not a stranger to those kinds of tests

Hi @jayceslesar

I did some further digging today and found out why it takes (quite) some time before jupyter-dash throws an OSError when the address is already in use. In the run method multiple times they check if an alive message from the address can be received (see code below)

https://github.com/plotly/jupyter-dash/blob/b6dad2689af96b70b796ae3d766a9c56d306c47d/jupyter_dash/jupyter_app.py#L323-L342

As this apparently intended behavior of jupyter-dash, I'm not really keen on overwriting this in this library (as it is not at all our responsibility).

What do you think of this? Or were you targeting another bug / issue with this PR (that I am missing at the moment)?

Cheers, Jeroen

I think taking no action is a fine way to deal with this -- I completely agree that it is not this libraries job to better handle how things are treated upstream. Feel free to close

Good to close this and #73 ?

I experienced (plotly-resampler 0.8.0) that this dash app error might keep "hanging" indefinitely, seems to be caused by a dash-werkzeug combination. Will need to further look into this (fyi: @jvdd @jayceslesar )

https://github.com/plotly/jupyter-dash/issues/103#issue-1385617801

Will create a new issue for this later this day.