On Fri, Sep 6, 2024 at 12:13 PM Bart Schaefer wrote: > > Remaining question is whether stats should be printed on error cases, > e.g., "no such builtin": I've gone ahead with this, it's a bit ugly because of all the early "return;" but it works. Tests included, though they only check whether `time' returns something, not whether that something makes sense in terms of cpu usage etc. If anyone has any ideas how to make that predictable enough to pattern-match the output, feel free to update the tests.