remove clean_parse_tables.py (to be moved to astropy-tools) - #4300

👍

Perhaps this should be accompanied with a few comments in the relevant source files as to where to get this script, and that it should be run when updating the parser/lexers.

@embray - good point - how about the README I just added?

@astrofrog - this seems really weird... coveralls said coverage decreased, but the only change here is removing code that definitely was not being tested. So the denominator should shrink and therefore coverage should have increased, right?

Not important for the PR itself, but possibly signifying a coveralls issue?

The readme isn't too helpful because who would know to go looking for it? It's not in a place where anyone would look when making updates to the parsers. Instead it would be good to just have a short comment directly in the code such as in this file: https://github.com/astropy/astropy/blob/master/astropy/units/format/cds.py#L27

Additionally, mixing our own files with vendored code is confusing.

@embray - so you're thinking everywhere that there's a lextab/parsetab?

I confess I don't really understand when/how those files are generated, so I could be confused. If you (or @mdboom) have a specific idea of where this is most likely to be recognized, a PR to my branch might be fastest.

Or maybe the best solution is for someone (probably shouldn't be me given the above...) to add a developer doc along the lines of "Upkeep of PLY-generated files in astropy"?

It's only used in a handful of files:

https://github.com/astropy/astropy/blob/c7099d60a1d38b6074801d528ad069f4d06a8202/astropy/units/format/cds.py

https://github.com/astropy/astropy/blob/c7099d60a1d38b6074801d528ad069f4d06a8202/astropy/units/format/generic.py

https://github.com/astropy/astropy/blob/c7099d60a1d38b6074801d528ad069f4d06a8202/astropy/units/format/ogip.py

https://github.com/astropy/astropy/blob/c7099d60a1d38b6074801d528ad069f4d06a8202/astropy/coordinates/angle_utilities.py

Just a copy/pasted comment near the top of all of those files should suffice.

@eteq - Re coveralls: we saw this before with other PRs, too (e.g. #3628). My best guess was that the issue is somewhat related that the coverage changed between the base of this PR and the top of master, and the change this PR introduces is negative compared to that.

If this right, rebasing this should make the fail go away (given that no change happens in master in the meantime).

OK, is this what you mean, @embray (the commit I just pushed up)? Again, I really don't understand what PLY/lex/yacc are doing and when the tables are getting generated, so I just sort of made this up...

This is fine.

Alright, I'll go ahead and merge.

backported to v1.1.x in c8f65a0