The examples are being build on travis (as late as 6 hours agohttps://travis-ci.org/matplotlib/matplotlib/builds/158107836) so I think this is a temporary outage.

I would suggest just disabling them from running as parts of the docs build as suggested in#7049

@jenshnielsen I think we should remove them:

1. This are in git history. They are not lost.
2. The finance module is already deprecated.
3. That code isterrible in so many way. We could fix it, by adding a timeout, but then we would need to disable the fact that warning raises errors in the build of our documentation. I personally think it is not worth it considering the code is already deprecated. It also means that people can't work offline (!!).
4. We also need to stop keeping code stacked, that no one uses, as it leads to situation were we have code that doesn't run, without anyone noticing for years and years and years.

1. Also any change in the yahoo API means the documentation is broken.

It may be that the finance module is already deprecated but we can't remove it unless we want lots of complaints from users that are still depending on it. I am just suggesting that we:

1. Disable all the finance examples from the docs now.
2. Move the finance module and all it's examples to it's own python package and delete it from matplotlib

I think the data used in the examples can be saved to mpl repository and used to 'proxify' urllib requests. It can also be used in the tests for financial module

Oh sorry for closing. This was a mistake.

Why not generate totally fake data instead, would this make a difference for the example ?

I agree that it's not a good use of time to do any work on these examples to fix the data loading then I would rather delete them, I just suggest that they should be deleted when they are moved with the finance code to their own repository. Then if anyone should feel like improving the code and or the examples they can do so in that repository independent of matplotlib.

Why not generate totally fake data instead, would this make a difference for the example ?

There isfetch_historical_yahoo I think examples should be rewritten to use it and in that waycachename parameter can be used for this

I will try to get around to creating a mpl_finance package asap that way we can delete the examples right away and perhaps even the finace module in time for 2.0

We probably should shim out mpl.finance to import from the new package and raise an exception with instructions of what to install to fix it.

To keep things moving for@NelleV and company I am going to merge this and then immediantly open a PR to revert the merge. I think this unblock current progress and make sure we do not (will make it less likely) that these fall through the cracks.

Just waiting for the docs to finish on travis to make sure we do not have an explicit reference to any of these.

There seems to be a hard-coded list someplace

/home/travis/build/matplotlib/matplotlib/venv/lib/python3.5/site-packages/numpy/ma/core.py:4144: UserWarning: Warning: converting a masked element to nan.  warnings.warn("Warning: converting a masked element to nan.")Exception occurred:  File "/home/travis/build/matplotlib/matplotlib/venv/lib/python3.5/site-packages/sphinx/util/parallel.py", line 99, in _join_one    raise SphinxParallelError(*result)sphinx.errors.SphinxParallelError: FileNotFoundError: [Errno 2] No such file or directory: '/home/travis/build/matplotlib/matplotlib/doc/mpl_examples/pylab_examples/finance_work2.py'The full traceback has been saved in /tmp/sphinx-err-jsfn96ab.log, if you want to report the issue to the developers.

Minimal mpl_finance package up athttps://github.com/matplotlib/mpl_finance

backported to v2.x asb152ff9

\o/ Thanks!