Vasil Rangelov

Ughhh... fine... I guess a wildcard (as in the "*" and "?" meta characters, with their Windows wildcard semantics) wouldn't be difficult to implement consistently across IDEs. This means the...

@lividgreen Regardless, how do you expect IDEs to behave when a class has such a "template" defined? Offer to complete the method/property name up to the first "*"?

To add to @bkuhl's point about accuracy... does Doctrine _generate_ its ORM methods, or are they determined at runtime by a common class? If they're supposed to be generated, IMHO,...

... and in addition to that, there's no single regex language. There are multiple dialects, which differ on some more exotic features. Having that said > Which regex dialect to...

But that's just it - _in principle_, if you have an API design that demands the use of such defined-only-at-runtime properties, your API design is bad. A better one is...

Whoa. Well... _that_ certainly changes the whole discussion. I was actually thinking yesterday about the moving cursor part you were talking about above (and started leaning "pro" due to that),...

Note that the only reason this workaround works is because "GIT: 0.1" is not actually a valid "GIT" version vector, and is therefore treated as part of a plain text...

The two forms shown in the docs are the two supported ones - a number without any sort of prefix, and a VCS vector enclosed with "$". As I noted...

IMHO, it's a better practice to define (class) constants, and then explicitly clarify in the description that acceptable values are one of said constants. But disregarding that bit... Consider scenarios...

There are already array specifiers, right? So there's no need to worry about that bit. It's only a matter of specifiers for scalar types (which should extend to those that...