Wrong render of the regular expression string in the Basics tutorial

[s-and-witch]

Here https://hackage.haskell.org/package/regex-tdfa-1.3.2/docs/Text-Regex-TDFA.html in Basics section first regular expression rendered by haddock as

>>> let emailRegex = "[a-zA-Z0-9+._-]+\[-a-zA-Z]+\.[a-z]+"

But that's not a valid Haskell string, it should be

let emailRegex = "[a-zA-Z0-9+._-]+\\@[-a-zA-Z]+\\.[a-z]+"

Perhaps you need to add more slashes or something

[andreasabel]

The source gives a correct Haskell string:

regex-tdfa/lib/Text/Regex/TDFA.hs

Line 35 in 693a7f8

>>> let emailRegex = "[a-zA-Z0-9+._-]+\\@[-a-zA-Z]+\\.[a-z]+"

The Haddock documentation contains an example with a escape sequence in a string, and there is no extra escaping: https://haskell-haddock.readthedocs.io/en/latest/markup.html#examples

-- >>> putStrLn "foo\nbar"
-- foo
-- bar

I think this should be considered a haddock bug rather than a bug here, do you agree @s-and-witch ?

[andreasabel]

Ok, I played a bit with the problem. The crux seems to be the explicit code block annotations @...@:

regex-tdfa/lib/Text/Regex/TDFA.hs

Lines 34 to 53 in 693a7f8

	@
	>>> let emailRegex = "[a-zA-Z0-9+._-]+\\@[-a-zA-Z]+\\.[a-z]+"
	>>> "my email is [email protected]" =~ emailRegex :: Bool
	True
	>>> "invalid@mail@com" =~ emailRegex :: Bool
	False
	>>> "[email protected]" =~ emailRegex :: Bool
	False
	>>> "#@invalid.com" =~ emailRegex :: Bool
	False
	/-- non-monadic/
	λ> \<to-match-against\> '=~' \<regex\>
	/-- monadic, uses 'fail' on lack of match/
	λ> \<to-match-against\> '=~~' \<regex\>
	@

It seems that removing them would fix the formatting. After all, >>> is already indicating code!

Blame for introducing @...@ goes to e3c15fb, but maybe at that time it was an overall improvement.

[andreasabel]

@s-and-witch : Please check the new rendering at https://hackage.haskell.org/package/regex-tdfa-1.3.2.1/candidate/docs/Text-Regex-TDFA.html

[s-and-witch]

Cool, thanks for such a quick fix!

[andreasabel]

Ok, then I'll release the fix!