Export all of original parsec's combinators about megaparsec

[abooij]

Some of these combinators have different semantics than Parsec's. Code will not work as-is, even with a fake translation layer.

many1 is not really "missing", it's a misnomer that has been purged. some is the new many1. Either the translation is easy, and might as well be done manually in translating code from Parsec to Megaparsec, or it is not easy, and cannot be done by merely adding drop-in lookalikes.

[neongreen]

some is the new many1

It should really be mentioned in Text.Megaparsec.Combinator (in fact, it should be reexported properly, because people won't notice it if it's simply mentioned in module description). Alternative is a very little-known class.

[neongreen]

Hm, on the other hand they are all reexported in Text.Megaparsec.Prim, so perhaps merely adding a link in Text.Megaparsec.Combinator would be enough after all.

[mrkkrp]

@recursion-ninja, This library has different naming conventions. many1 is called some and spaces is called space. There is nothing you need to define manually, just read the documentation. Some of combinators are placed in different modules, but there are reasons for it, this is different library.

@neongreen, I don't think we should mention this. The library is a fork and it's similar to Parsec and it does all the things Parsec does, but of course a bit differently. The library should be self-sufficient and we should stop define it in terms of Parsec. I don't like this trend “please look like Parsec or you will be never widely used”. There will be not a single step in that direction, that's not what I wanted when I started to work on this.

Closing.

[recursion-ninja]

At the time I made this suggestion I didn't realize that Megaparsec was diverging from Parsec with breaking changes to be a better standalone parsing library, not an improvement to Parsec.

The new documentation does clear this up!