Mathics3 / mathics-scanner

Tokenizer, and character tables, operator precedence, and conversion routines for the Wolfram Language.
GNU General Public License v3.0
17 stars 3 forks source link

Amstext+wl unicode convert #51

Closed rocky closed 2 years ago

rocky commented 2 years ago

Go over character tables

Because of this, wl_to_unicode() needs to use "ascii" when there is no unicode-equivalent, but "has-unicode-inverse" is set.

rocky commented 2 years ago

As suggested in https://github.com/Mathics3/mathics-scanner/pull/43#issuecomment-1250410661 it is possbile that ascii-operator-to-name is not what is needed in Mathics core, but rather ascii-operator-to-unicode (the standard unicode not the WMA unicode since most non-WMA front-ends don't have WMA fonts installed). Of course there could also be a ascii-operator-to-wl-unicode but let's add that only when it becomes necessary.

That said, let's merge this now and feel free to drop or add tables as it becomes necessary based on how Mathics core handles ASCII to unicode conversion of operators.