On Mon, Mar 29, 2010 at 9:18 AM, Jeremy Shaw <jeremy@n-heptane.com> wrote:

On Mon, Mar 29, 2010 at 8:47 AM, Michael Snoyman <michael@snoyman.com> wrote:
Jeremy,

We'd been discussing the PathInfo class previously; I understand now what you're trying to achieve with it, but I think for a lot of use cases using a parser like that will be unnecesary. For those cases, I'd hate to introduce a parsec dependency, especially given the 2/3 split we're dealing with right now.

Well, parsec is in the haskell platform, so it's really a question of can be it implemented so that it works with both 2 and 3.

I know it's painful to split this up into even more packages, but we do you think of adding web-routes-parser?

I would consider it, but I don't understand how that would work...

So far you have been pretty unexcited about the parsing having a type similar to:

[String] -> (Either String a, [String])

because it is often 'unnecessary'. But the problem with, [String] -> (Either String a), is that, as far as I can tell, it won't work in the cases where it *is* necessary. And you can't simply switch to, [String] -> (Either String a, [String]), when that happens.. But perhaps I am missing something here..

The reason I'm unexcited is that I never would have dreamed of defining my routes that way. I don't feel like drawing out this point too much, because you clearly *would* define your routes that way. However, just to draw the distinction in how I would do things differently, I'll use an example of mine that you quoted earlier:

instance Yesod PB where

resources = [$mkResources|

GET: indexHandler

/entries/$entryId:

GET: entry

/entries/$entryId/$filename:

GET: media

/feed:

GET: feed

If I were to convert this to a datatype, it would be:

data PBRoutes = Home | Entry String | File String String | Feed

I simply wouldn't nest a datatype inside any of the constructors. I understand that you want to do this in some circumstances, but I would simply "duplicate" the parsing code for the Entry and File constructors, since I find that parsing code trivial. In particular:

parsePB ["entries", eid] = Entry eid

parsePB ["entries", eid, filename] = File eid filename

I don't see a need for providing a sophisticated parser.

Also, question about the Site datatype: Did we want the formatLink and parseLink functions to work on [String] instead of String? I think I see advantages either way. Also, I'm not sure if I sent this previously, but I think the defaultPage function is redundant; that should probably be specified by 'parseLink ""'.

I will have to think about [String] vs String more. I don't want using Site to mean you have to use PathInfo, but switching to [String] wouldn't have that effect, and would mean that people would not have to worry about those pesky encoding issues. The only real use case I have for not using PathInfo is if you wanted to create a function like: gfromPathSegments :: (Data url) => PathInfo -> Either String url. And that can be done most easily if PathInfo = [String]. So I think I'll make that change. In the old code base, where there was no url encoding done by the library, it did not really make a difference.

Regarding defaultPage, not all methods of generating formatLink / parseLink lend themselves to handling "" very easily, so it seemed nice to have a single place that can add that case in, rather than add similar code to all the parsers.. Also, I think that I sometimes changes the landing page of my site. But I think I will make 'defaultPage' a Maybe value, and bill it as a way to 'override' where "" resolves to.

That sounds good.

I'll try to get started on porting the Yesod mkResources code over to web-routes now.

awesome!

- jeremy