Require bracketting when detecting floating point size in code notes. #971
Replies: 2 comments
-
This already occurs. "float" must appear as a singular word, with a non-alphabetic character (or end of string) on both sides. Here are tests cases illustrating: RAIntegration/tests/data/models/CodeNotesModel_Tests.cpp Lines 232 to 238 in 37ae6dd
That does not happen. "[8bit] Fish Up the Cat float" could be affected. Maybe something could be done to ignore extraneous "float" words if an explicit non-adjacent size exists in the string.
That would be inconsistent with non-float notations. There are many notes like "16-bit money value" already in the system. Making brackets required would break those. |
Beta Was this translation helpful? Give feedback.
-
That's probably worth doing. I think I was thinking of old behavior, thanks to a conversation about something, but preferring other noted sizes sounds worthwhile. |
Beta Was this translation helpful? Give feedback.
-
Code notes will very greedily take any instances of the letters FLOAT next to each other and treat the address as a 32-bit floating point (single precision) size. Those letters frequently are meant to be a word in the note's description rather than size, which can lead to oddness like "[8bit] Fish Up the Cat floater" being treated as a floating point. My suggestion would be to only treat those letters as a floating point size if they are the sole letters in a word, and also included in brackets of some kind (any kind).
Summary:
Beta Was this translation helpful? Give feedback.
All reactions