Jeffrey Friedl's "Mastering Regular Expressions" was my gateway into Perl and server-side programming so you'll excuse me if I don't buy(!) one of your licences. Regular expressions used intelligently are amazing.
> After all, most people copy&paste complex regexps rather than author their own.
They do? Somehow I have not seen this happening. But if this is really how software development is done these days, then it is highly irresponsible. Simply copy-pasting complex regexps without understanding what they do is ripe for disaster. Not to mention that every regex flavor has subtle differences from every other flavor. So at minimum the person copying has to understand it and adjust it for differences in flavors.
I think that's pretty common, yes? After all, it's pretty hard to copy&paste without at least some tweaking, even if only to change the variable names appropriately, and that requires some level of understanding.
To explain what I mean, here's a regexp which is starting to get complex - match an IEE754 double.
I could and have done it myself. These days I'll search for then copy&paste something like:
Eyeballing looks right (I might need to add [+-]inf, NaN, and nan), and at the very least it's a good start, yes?
From personal experience, it's easy to forget .4e1 is valid (no digit before the period, lowercase "e", no sign for the exponent). I would rather start with someone's worked-out version than from scratch.
Regexps can of course be far more complex than that. OTOH, the most common use cases are not that hard to find, like "How do you validate a URL with a regular expression in Python?" at https://stackoverflow.com/questions/827557/how-do-you-valida... , with several good answers at different levels of completeness, and this entirely appropriate comment:
"I've needed to do this many times over the years and always end up copying someone else's regular expression who has thought about it way more than I want to think about it."
> And those who have licenses generally know it’s a better idea not to use them.
This is basically saying "don't use regex" which of course is silly. However you probably shouldn't use regex for user-facing input validation in production applications, at least not by default.
