You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To avoid infecting the dependency graph with potentially unwanted regex features. If the user wants extra regex features, they should enable them in their project (e.g. just add a dependency on "regex" with default features enabled).
I wonder about perf-* features vs unicode-* features though: while a missing unicode feature will probably show up during testing and the devs will realize that they need to opt in, a missing perf feature might just silently make the binary or the unittests slower. Since they don't cause as much binary bloat (just longer compile times), maybe it makes sense to leave those in for predicate-rs.
This is a subtle breaking change (runtime behavior, rather than APIs)
I'd rather not people have to depend on regex to get the features they want but I also don't want to re-exporting
In a lot of cases, I can see someone wanting things to "just work" and adding extra hoops could be frustrating.
I think I'd be more inclined to make the regex feature off by default instead. That could also possibly be done in mockall, I'm not aware of its design to say whether it needs it or not (which is how starship is pulling in predicates, I presume)
To avoid infecting the dependency graph with potentially unwanted regex features. If the user wants extra regex features, they should enable them in their project (e.g. just add a dependency on "regex" with default features enabled).
PR #143
The text was updated successfully, but these errors were encountered: