Conversation

Notices

1. Embed this notice
   Clément Robert :python: (neutrinoceros@fosstodon.org)'s status on Monday, 26-Feb-2024 19:51:15 JST Clément Robert :python:

   in reply to

   In fact one of my supported cases was never tested (and, of course, broken), but I would never have figured it out with the old code: all lines were being hit when the dictionary was initialised !

   Initially I was motivated to switch to newer syntax because it supposedly will help catching missing cases in the future (combining enums and `typing.assert_never`), but I didn't except to find an existing bug !

   • Embed this notice
     powersoffour (powersoffour@mastodon.social)'s status on Monday, 26-Feb-2024 19:51:15 JST powersoffour

     in reply to

     @neutrinoceros interesting. Thanks for posting this!

   • Embed this notice
     Clément Robert :python: (neutrinoceros@fosstodon.org)'s status on Monday, 26-Feb-2024 19:51:16 JST Clément Robert :python:

     Unexpected benefit from using #Python 3.10’s match/case syntax:
     I replaced code that was essentially equivalent to
     ```
     axes = {
     CARTESIAN: ("x", "y", "z"),
     CYLINDRICAL: ("r", "theta", "phi"),
     }[geometry]
     ```
     with
     ```
     match geometry:
     case CARTESIAN:
     axes = ("x", "y", "z")
     case CYLINDRICAL:
     axes = ("r", "theta", "phi")
     ```
     Which is equivalent, right ?
     But the *coverage* is not necessarily so !