Some technical issues … that are related to the lack of ability to manage regular expressions are described on T176312, T214378, T236150 and T240884.

How would any of these be affected by a separate data type for regular expressions / patterns? The problem is the same: we still need to evaluate user-specified (untrusted) regular expressions safely; I don’t see what a separate data type would change here.

A specific data/property type for regular expressions or patterns would:

• ensure that the stored regular expressions or patterns are syntactically correct, in a similar way that quantity-type properties ensure that their statements are not paragraphs, something from which all implementations, tools and reusers would benefit;
• introduce a specific input (interface) control in Wikibase (not WikibaseQualityConstraints) that could make pattern editing more friendly: monospaced text, warnings, colored brackets, and all the amazing features that the developers want to implement. :-)

This does not solve the low-level security issues of each implementation (e.g., avoid running the regular expression "'; DROP TABLE important_things;").

Would it also allow for reusing a regex?
See https://w.wiki/RK2 for stats regarding multiple properties using the same regex:
[1-9]\d* is used 615 times
\d is used 542 times
[1-9][0-9]* is used 89 times

Also, having items for the regex patterns would mean that they would be better monitored by users and optimized ([1-9][0-9]* is the same as [1-9]\d*)

@DannyS712: I think I don't understand your question. What do you mean by reusing a regex? What would you expect could be done on the UI that is not possible now? On https://www.wikidata.org/wiki/Wikidata:2020_report_on_Property_constraints#format there are also some general stats about our regular expressions.