
Hi fellow devs, Do any of you use github for browsing the GHC code? I find it useful sometimes, especially when I'm collaborating with someone remotely and wish to send a link to some code, or when I don't have a particular branch to hand. As of October, github now renders all .lhs files as Markdown. It seems other Haskellers have requested this, but it wreaks havoc on GHC's code, which does not use Markdown. In particular, Markdown uses ~~~~ to delineate code segments, and GHC uses that to underline Notes. This has bothered me for months, but I finally rooted down the cause and opened a ticket. The ticket is here: https://github.com/github/markup/issues/196 If you have run into this problem and have an opinion about it, please share it there! Thanks, Richard

Opening a ticket is a good idea - this problem has been iritating me as well and in fact I almost gave up on browsing GHC code on github. As a side note it seems to me that writing literate Haskell is not really popular and I would be happy to see the source code of GHC converted to .hs one day (I wouldn't mind converting some of it myself). Janek

On Thu, Apr 25, 2013 at 2:21 PM, Jan Stolarek
As a side note it seems to me that writing literate Haskell is not really popular and I would be happy to see the source code of GHC converted to .hs one day (I wouldn't mind converting some of it myself).
Care to explain what this means? Surely you're not asking everyone to change their coding convention to suit your taste? You don't need permission to fork and rewrite code as you see fit. Forking on social coding sites is nice because you get a free open invitation to everyone similarly inclined to join in and help you out. -- Kim-Ee

Surely you're not asking everyone to change their coding convention to suit your taste? So, is literate Haskell everyone's coding convention? It seems to me that today people are more likely to document their code with Haddock and the code written in literate Haskell is remnant of a distant past. That's my impression at least. I'm not asking people to blindly do what I ask, no need to get upset here. But I think a bit of discussion on this subject wouldn't hurt. Right know GHC has some code written in literate Haskell and some code written in plain Haskell. Having a uniform way of doing things would not be a bad thing. Especially that using both literate Haskell and plain Haskell can use to problems such as this: http://hackage.haskell.org/trac/ghc/ticket/7605
Janek

On Thu, Apr 25, 2013 at 4:45 PM, Jan Stolarek
Vestigial file problems can get nasty; kudos on the catch! -- Kim-Ee

On Thu, Apr 25, 2013 at 12:21 AM, Jan Stolarek
Opening a ticket is a good idea - this problem has been iritating me as well and in fact I almost gave up on browsing GHC code on github.
+1 I use "raw" view for .lhs files as a workaround.
As a side note it seems to me that writing literate Haskell is not really popular and I would be happy to see the source code of GHC converted to .hs one day (I wouldn't mind converting some of it myself).
My experience has been that outside of papers, tutorials, and blog posts that lhs is more trouble than it's worth. The downside is that haddock is not a very good markup language. It's just fine for plain text comments, but if you want to say something mathematical in a haddock comment it may get the wrong idea (x /y + a/b, would render y + a in italics and not show the divisions). So I don't think we should convert them unless there is a stronger reason than github rendering. I think the right thing is to leave the .lhs files how they are for now and try to convince github to stop treating .lhs as markdown. Jason
participants (4)
-
Jan Stolarek
-
Jason Dagit
-
Kim-Ee Yeoh
-
Richard Eisenberg