Toby Inkster

@NacreData, you wrote: > looking into Kavorka to isolate out what we are actually using and what actually needs 5.14 to see if we can use a fork of some...

There is a very large overlap in syntax. (This is intentional.) Buddy Burden (one of the maintainers of Method::Signatures) and I did discuss the possibility of some kind of Any::Signatures...

Additional information: if my external monitor is plugged in and switched on *before I log in* then everything works fine.

I guess I should update this issue. I've upgraded to Ubuntu 20.04 and things work okay now. ``` $ lsb_release -a No LSB modules are available. Distributor ID: Ubuntu Description:...

Hmmm, interesting. **StrMatch** is supposed to be a child of **Str**, which does do a definedness check. It looks like **StrMatch**'s inlining code is not being strict as its parent...

I actually can't reproduce this error, as `eval_closure` normally compiles code without `use warnings`. However, I still think `StrMatch` ought to be checking definedness, so I'll add that check.

Ah, I see. You're using Eval::Closure, which *does* switch on warnings. Eval::TypeTiny does not.

These two lines are exactly equivalent in functionality: ```perl !ref( $foo ) and $foo =~ m/.../ defined( $foo ) and !ref( $foo ) and $foo =~ m/.../ ``` The only...

Oh yeah, in normal code, I'd keep the warnings on for sure and would check definedness. Inline-generated code, I tend to value performance over maintainability and readability though. I'll keep...

While I agree that compile error is definitely not what should happen, I also don't think it's clear what the "correct" behaviour would be either. A slurpy parameter is intended...