I've been starting to take advantage of Python code coverage tools lately, which work by hooking into the Python interpreter (much like profiling tools) and record which lines in your code are actually exercised when running the test suite. This is usually summarised as a percentage score (fraction of lines of code tested, comments not counted).
It would be nice to have something similar for Galaxy tools assessing how much of the tool XML file is actually exercised in the tests. The Cheetah syntax in the <command> tag is probably the hardest and would need a similarly invasive approach. Likewise the <configfile> tag.
For now I think focusing on static analysis via planemo lint seems most practical.
hexylena
commented
7 years ago
I've been starting to take advantage of Python code coverage tools lately, which work by hooking into the Python interpreter (much like profiling tools) and record which lines in your code are actually exercised when running the test suite. This is usually summarised as a percentage score (fraction of lines of code tested, comments not counted).
It would be nice to have something similar for Galaxy tools assessing how much of the tool XML file is actually exercised in the tests. The Cheetah syntax in the
<command>tag is probably the hardest and would need a similarly invasive approach. Likewise the<configfile>tag.For now I think focusing on static analysis via
planemo lintseems most practical.