colordiff could have special rules for IDs and numeric/date fields

[RollingStar]

Today the importer gave me a new MBid for an album. It's just a UUID like 39e73e7d-a25a-43b7-aa81-0035e9478950, unreadable. The string rules for the highlighter will still color same characters the same color.

Great for strings, not for IDs. I propose a simple
[color 1] -> [color 2] for IDs. Even when an "e" is in ID 1 and ID 2. Color them differently.

For numeric and date fields, we could have highlight rules for [new value is lower] and [new value is higher].

https://github.com/beetbox/beets/pull/3721/files#diff-9e6736f74918c0608860bf8ce0d4d9b3333adaa2025f7b7b4d28189cf64c1958R653

beets/beets/ui/__init__.py

Line 552 in e1a901d

def _colordiff(a, b, highlight='text_highlight',

[sampsyo]

That's a good idea! Highlighting for numbers and dates could get complicated, but at the very least not attempting a "string diff" on IDs seems like a good idea. 😃 Would you mind opening a feature request issue?

[RollingStar]

Highlighting for numbers and dates could get complicated

Do you have an example? In my mind it would be a 'whitelist' of non-string fields. Catalogs numbers/descriptors would probably stay normal strings, even if entirely numeric.

For date comparison, there are probably libraries that do it, but I know you like keeping the dependency list small. Non-MB-matched tags could have dates in any format, but we could skip those or apply a best guess (using a date library). Worst case a smaller date is actually printed in the "larger date" color, which isn't much different from present state.