On Mon, 29 Apr 2013 18:04:47 +1200 "Richard A. O'Keefe"
<o...@cs.otago.ac.nz> wrote:

> so that there is no possibility of catching errors early;
> by definition in that processor there are no errors.

Haddock's markup isn't any better in that regard. I spent two hours on
my first day with haddock figuring out that I needed an empty comment
line before a code block. It didn't issue any warnings or errors either.

> To be perfectly honest, most of the time when looking at
> a Haddock page, I end up clicking on the Source button
> because there are things I need to know that are in the
> source but not the documentation.

Besides fixities, orphan instances, type family instances and partially
exported records? It would be beneficial of Haddock to list orphan
instances on top of the page. In red.

Iff adding markdown doesn't require a major restructuring of haddock,
then a GSOC might be better spent adding support for all of these
instead; someone else could add markdown later on their own after ML
bikeshedding came to some conclusion.

Haskell-Cafe mailing list

Reply via email to