-
Improvement
-
Resolution: Unresolved
-
Normal
-
None
-
2.2.3
-
None
At the moment, these functions cast to `int` somewhere in between, which fails for non-integer arguments (like the bitrate). It would be nice if the cast was to `float`.
- is duplicated by
-
PICARD-2257 Math, trim, replace functions not working with extracted numbers from strings
-
- Closed
-
[PICARD-1713] Allow functions operating on numbers (mul, div, ....) to be used with floating point numbers
> Alternative would be we play it save and keep those functions for operating with integers, but add "f" equivalents (mulf, divf).
Very likely safer.
The numeric operators are currently all handled by the _compute_int function in picard.script. As the name implies this does the int conversion.
Not sure, for most operations it would probably be fine if we just change this to be a _compute_float instead for all numeric operators, or what do you think? There still could be a few cases where people rely on the int conversion, though.
The exception is probably div, this is currently intentionally using operator.floordiv (the // operator basically).
Alternative would be we play it save and keep those functions for operating with integers, but add "f" equivalents (mulf, divf).
As noted adding a new set of float funcs would be better. It would not break compatibility with existing scripts in any way while also allowing a transition to float easily for anyone that needs it. And for sure float would be really appreciated (it opens a tons of possibility for editing AcousticBrainz high level data without using substring everywhere as I did).