- web-devel - Haskell.org

Looking for criticism and comments on web-mongrel2.
by Clint Moore 21 Jan '11

21 Jan '11

After sufficient hand-wringing I finally uploaded web-mongrel2 up to hackage and I'm interested in criticism of my approach. The module itself is terribly simple as it's meant to provide a lightweight layer between an application or framework and Mongrel2, but if nothing else, it's a start. Anyway, the hackage page is at http://hackage.haskell.org/package/web-mongrel2-0.0.2.2 with a link to the github repo.

2 5

Re: [web-devel] [Haskell-cafe] Looking for criticism and comments on web-mongrel2.
by Clint Moore 21 Jan '11

21 Jan '11

On Fri, Jan 21, 2011 at 5:30 AM, Vo Minh Thu <noteed(a)gmail.com> wrote: > The link given on hackage to git is broken, it seems to be > web-mongrel2 instead of haskell-mongrel2. If you click on the latest version the link should be correct. And yes, it's web-mongrel2 - I caught that SECONDS after uploading it to hackage.

1 0

Re: [web-devel] [Haskell-cafe] Looking for criticism and comments on web-mongrel2.
by Michael Snoyman 21 Jan '11

21 Jan '11

On Fri, Jan 21, 2011 at 3:30 PM, Vo Minh Thu <noteed(a)gmail.com> wrote: > 2011/1/21 Clint Moore <cmoore(a)wamboli.com>: >> After sufficient hand-wringing I finally uploaded web-mongrel2 up to >> hackage and I'm interested in criticism of my approach. The module >> itself is terribly simple as it's meant to provide a lightweight layer >> between an application or framework and Mongrel2, but if nothing else, >> it's a start. >> >> Anyway, the hackage page is at >> http://hackage.haskell.org/package/web-mongrel2-0.0.2.2 with a link to >> the github repo. > > Hi, > > The link given on hackage to git is broken, it seems to be > web-mongrel2 instead of haskell-mongrel2. It worked fine for me...

1 0

web-routes-quasi gotcha
by Nubis 21 Jan '11

21 Jan '11

Hi guys, I was just investigating an issue in my yesod app that prevented me from running the devel-server. It was a warning about route patterns overlapping S4M/Server/Controller.hs:33:0: Warning: Pattern match(es) are overlapped In the definition of `parse[awnQ]': parse[awnQ] : "con" : "f" [] = ... parse[awnQ] : "con" : "h" [] = ... parse[awnQ] : "con" : "m" : "list" [] = ... parse[awnQ] : "con" : "m" : "c" [] = ... The issue seems to come up when part of the route is a single character (in this case, f, h and m) like /con/f/ /con/h/ /con/m/list/ I ended up just changing my routes to be more 'expressive' which is fine by me. Haven't looked at the source of the issue yet, I can reproduce it on 0.6.2 but haven't tried 0.6.3 so I apologize in advance if it is already fixed :) Just wanted to get it out here so I don't forget to look into it later (I'm heading home now) Cheers! ----nubis :)

2 1

Directory listings
by Michael Snoyman 19 Jan '11

19 Jan '11

Hi everyone, I'm working on a static file serving WAI app[1], which can either be used from a library, or combined with Warp as a standalone executable. One of its features will be directory listings. Is any interested in designing a directory listing page? Extra points if you can style it using some Haskell colors. Thanks, Michael [1] https://github.com/snoyberg/wai-app-static

2 2

Changes implemented in Hamlet
by Michael Snoyman 15 Jan '11

15 Jan '11

Hi everyone, I wanted to give you all an update on the changes in Hamlet 0.7[1]. So far, the following have been implemented: * Adding angle-bracket syntax * Remove period-delimited variables, and allow qualified functions. For example, $Data.List.sort myString$ * Change forall and maybe to use <-, eg $forall person <- people I haven't tried it on real life code yet, but I have converted the entire Hamlet test suite to use this, and I'm quite happy with the result. Overall, I think things are much more readable. Just for the record: the second item (period-delimited variables) affects Cassius and Julius as well. So now for the last few decisions: * Do we rename maybe to just? I am leaning towards no on this. * For Cassius, do we switch from a whitespace sensitive syntax to brackets? * Do we convert from $my variable$ to #{my variable}. * And if we do this, can we consider #{my (nested call)} and #{my $ nested call} the same? Please see the wiki page[1] for more information on all of these. As far as the I18N note at the bottom: I think we'll be able to release such a feature without introducing any breaking changes. Most likely, a "message" will be a specialization of standard variable interpolation, eg: #?{myI18NMessage} In other words, I'm in no rush to get that feature in the 0.7.0 release, it can wait until 0.7.1. I *do*, however, want to get this Hamlet released soon so that it doesn't hold up Yesod 0.7 any longer: since Yesod 0.6 will not work with WAI 0.3, it also won't work with Warp, and I think a lot of people will want to be able to use Warp soon. Michael [1] http://wiki.yesodweb.com/Hamlet%200.7%20changes

1 0

[Yesod] an API which return Root Relative URL.
by いとうかつとし 07 Jan '11

07 Jan '11

Hi I want an api which return Root Relative URL. I've been developing wiki clone. When a writer select and click an image from photo galary in order to insert a image, click event insert a wiki image link like as "![image_name]( http://localhost:3000/uploaded/1)". But I think this - wiki contents has self Absolute URL - is not good. Is there the API? I want the api, but i remove approot string from Absolute URL, now.

2 1

lazy RESTful Content
by Greg Weber 07 Jan '11

07 Jan '11

lazy and REST seem to go together :) Perhaps this will expose my ignorance of laziness/strictness and isn't at all Yesod specific. I have a handler that can serve html or json. When an html request comes in it still executes the json code. I assume this is because the json code uses liftIO. What is the strategy for avoiding executing the json code?

2 3

Yesod routes and Persistent entities in separate files
by Michael Snoyman 05 Jan '11

05 Jan '11

Hi all, The Yesod scaffolding tool currently does precisely one thing: produce a scaffolded site. It does *not* provide any facilities for adding new routes or entities after the fact, or creating basic CRUD interfaces the way Rails or Django do. Nubis has proposed adding this functionality. The current roadblock is that the Yesod routes and Persistent entities are defined with quasi-quotation within the code. I believe that technically speaking it should be possible to make this kind of feature work, but it would make more sense to do it as a separate file. I think the proposal would essentially change: mkYesod "MyApp" [$parseRoutes|...|] to mkYesod "MyApp" "MyApp.routes" with the ... being in MyApp.routes, and a similar change for the $persist quasi-quotation. What does everyone think of this? I see the following advantages: * Makes it more straight-forward to write these kinds of scaffolding tools. * Gets away from quasi-quotation for those who have an aversion to it. On the other hand, the downsides are: * One more file to deal with. * The standard issue of Haskell files not automatically recompiling when an external files are changed. This can be worked around with wai-handler-devel. Another possibility is offering *both* options, but as much as possible I'm trying to avoid creating multiple approaches. Michael

2 2

Re: [web-devel] Wrapping the widget generated by the authEmail backend.
by Nubis 04 Jan '11

04 Jan '11

Exactly! I was looking for addWidget. I'll give it a shot at providing my own html also, I'll let you know my final solution. Thank you all guys. cheers ----nubis :) On Tue, Jan 4, 2011 at 9:29 AM, Matt Brown <matt(a)softmechanics.net> wrote: > Regarding your error, I think you want addWidget instead of addHamlet. > This compiles for me: > > loginHandler = do > let tm = liftHandler getRouteToMaster > let email = tm >>= apLogin authEmail > defaultLayout $ do > setTitle $ string "Login" > addWidget $(hamletFile "login") > > -matt > >

1 0