Description
This issue was originally filed by @cpaelzer in the MIR repo
Request
Our current template based approach is already much better than the old "read rules and write an answer from scratch".
But that does not mean it is perfect, in the MIR team discussion at the last sprint we identified in particular that some easier MIR cases often have us spend almost more time on text processing than on the inspection of the package.
The desire out of this is to consider converting the templates to tooling which you could run so that it automatically eliminates some of the sections. While some are hard to code and therefore out of reach, others like "python kg, not adding python2 dep" sound fairly easy.
So the idea here is to at least get the easy ones sorted out automatically before engineering time is spent.
Conceptually we do not want to make it strictly necessary to run a tool, but thought of prefixes in the template like "#IFGOLANG# and #IF!GOLANG# (or similar, this isn't meant to be a language but trivial) and if there is a checker for "GOLANG it could keep just the right related and drop the other lines. Once that structurally exists for one checker it is easy to be extended from there.
Replies in thread
(none)