Toby Inkster

> Having the list not an exact multiple of the number of variables has documented behaviour, and in many circumstances, that will be reasonable behaviour. Isn't that also true of...

https://github.com/tobyink/p5-linq is still quite immature. I only started yesterday.

That may be it. Though other HTTPS-only sites (like github) don't suffer similarly. Could it be your choice of cyphers?

I don't think so - type constraints are for coercing and validating incoming data. Here we're wanting to change the class of outgoing data. Currently URI::Namespace is doing something like...

Data::UUID and UUID::Tiny pretty much do the same thing. Data::UUID might be a little faster because it's XS, but I really doubt UUID generation is a performance bottleneck for many...

I don't know whether you consider the coverage thing to be a problem. Using Devel::Hide, or maybe an environment variable toggle, it might be possible to test both the XS...

Could do something like: ```perl my $HAS_XS = !$ENV{PERL_CLASS_TINY_PP} && eval { require Class::XSAccessor; 1 }; ``` And write a few test cases which do this right at the top...

Do you think the tests I submitted are sufficient?

`BUILD` does get passed `$args` as a parameter. This only includes arguments passed to the constructor (technically it returns the hashref produced by `BUILDARGS`) and not defaults.

`use Data::Checks` shouldn't enable or disable strict or warnings. `no Data::Checks` shouldn't enable or disable strict or warnings. In answer to your example: ```perl no strict; no warnings; use Data::Checks;...