What kind of content are you looking for? From a user POV, I think there is thankfully not much to learn about automata-based regexes like RE2 and rust/regex.

What kind of content are you looking for? From a user POV, I think there is thankfully not much to learn about automata-based regexes like RE2 and rust/regex. The exception is if you're writing a performance-sensitive system entirely around regexes, like an intrusion detection system or search engine. A regex is a very small concept with a very large syntax :) It's "just" literals, alternation, repetition, and concatenation. So it's mostly being able to see through the syntax IMO. Nice animated diagrams in this recent post - https://news.ycombinator.com/item?id=35503089 https://news.ycombinator.com/item?id=35503089 (On the other hand, from the implementation POV, there is a wealth of material online, but you already know where to find that :) I still think someone should "cliff notes" the Cox RE2 articles, since the ideas are important but there is a lot of detail. There are also dozens of references in those articles for people to read; it was a comprehensive survey at the time.) --- BTW Friedl's book covers backtracking engines, which makes sense because PCRE, Perl itself, Python, Java, JavaScript, etc. all use that style But as far as I remember, most of the book is irrelevant for users of automata-based engines. It's a whole bunch of detail about optimizing backtracking, reordering your clauses, and lots of detail about engine-specific and even version-specific differences. Personally I just ignore all of that, and it doesn't cause me a problem. I use regexes all the time in every language, and stick to the features that an automata-based engine can handle. Once you go into the backtracking stuff, then you might want the Friedl book, and that's a huge waste of time IMO. It's way easier to do that kind of thing outside of a regex engine, with normal parsing techniques. And honestly once you go outside, you'll find you don't even need to backtrack. Perl-style regexes can be an astoundingly inefficient way to write a pattern recognition algorithm. --- Eggex is my attempt to separate the two kinds of things, since it's not obvious based on the syntax: https://www.oilshell.org/release/latest/doc/eggex.html https://www.oilshell.org/release/latest/doc/eggex.html It makes the syntax smaller, more familiar, and more composable. It's closer to the classic lex tool (e.g. GNU flex). It's funny that lex has a way better syntax for regular expressions for Perl, but it's used probably 10,000x less often because most people want an interpreter and not a compiler to C code. But lex has 90%-100% of what you need, and it's way simpler, with a nicer syntax. Literals are quoted so you don't have the literal/metachar confusion, and you can reuse patterns by name. The manual is worth skimming: https://westes.github.io/flex/manual/Simple-Examples.html#Simple-Examples https://westes.github.io/flex/manual/Simple-Examples.html#Si... I use re2c though, not flex, and its manual is also worth skimming: https://re2c.org/manual/manual_c.html https://re2c.org/manual/manual_c.html