Hi folks. I'm looking at a test process that takes about 16 minutes for a full run.
Naturally, I'd like to speed it up. We've already parallelized it - mostly. It seems like the next thing to look at is setting up a local pypi, and building some of the packages that're compiled from C/C++ every time we do a full test run. (We're using docker and building dependencies for each full test run) Also, we could conceivably set up a web proxy...? Does having a local pypi obviate the web proxy? And what local pypi servers do folks recommend for speed? We need support mostly for CPython 3.x, but we still have a little CPython 2.x we require, and it's possible we'll need the 2.x for a while. Thanks! -- https://mail.python.org/mailman/listinfo/python-list
