Open bedroge opened 3 years ago
In this case, you were building IPython on top of an existing CVMFS stack and I think that IPython is querying the available python modules. CVMFS needs to first pull them into the cache, this is probably taking too long for IPythons liking and it thinks that something is up. That's why I think it works the second time around since the Python installation is already cached then.
Yes, that probably makes sense.
I think the actual issue was here: https://github.com/ipython/ipython/blob/master/IPython/core/completerlib.py#L141
And this timeout value is defined here: https://github.com/ipython/ipython/blob/master/IPython/core/completerlib.py#L52
I'm not sure if you can easily change it. There is a timeout scale factor for the tests:
https://github.com/ipython/ipython/blob/master/IPython/testing/__init__.py#L44
But, although they say they scale all timeouts via this factor, I don't see anywhere in the code that this also applies to the TIMEOUT_GIVEUP
. Could be worth trying, though it's a bit hard to test this...
I've seen the following error a few times during the test step of installing IPython:
This is the actual test:
Apparently, something is taking too long (slow I/O?) and causing a timeout. In all cases I was able to "solve" it by just redoing the installation.