Part of our continuous integration tests is an automatic check of all hyperlinks in our docs and docstrings. This is great in principle; however it seems to fail frequently for reasons beyond our control (notably, timeouts and -rate-limited cases. When it fails, it's also difficult to identify where the failure came from without reading every line of the output.
We need to find a way to make the output of this test more useful and to prevent it from failing for reasons beyond our control. I don't know much about the configuration options, but ideally, we should treat timeouts and -rate-limited cases as warnings not errors (in other words, ignore those issues).
Part of our continuous integration tests is an automatic check of all hyperlinks in our docs and docstrings. This is great in principle; however it seems to fail frequently for reasons beyond our control (notably, timeouts and
-rate-limitedcases. When it fails, it's also difficult to identify where the failure came from without reading every line of the output.We need to find a way to make the output of this test more useful and to prevent it from failing for reasons beyond our control. I don't know much about the configuration options, but ideally, we should treat timeouts and -rate-limited cases as warnings not errors (in other words, ignore those issues).