April 2011 - web-devel

ANN: Hakyll 3.1
by Jasper Van der Jeugt 06 Apr '11

06 Apr '11

Hello all, I've just uploaded Hakyll 3.1.0.0 [1] to Hackage. It contains some backwards-incompatible changes (but they are really for the greater good). The most important change is the matching done in the rules DSL. What used to be written as: route "posts/*" someRoute compile "posts/*" someCompiler Is now written as: match "posts/*" $ do route someRoute compile someCompiler This new style is way more expressive, as you can define custom predicates such as: match (predicate (\i -> matches "foo/*" i && not (matches "foo/bar" i))) $ do ... For more information, see the Pattern documentation [2]. Another slight change is that the `Hakyll.Web.Tags` [3] module now uses an association list instead of a `Data.Map` -- in case you want to change the order in which tags are displayed. If you have any questions or trouble migrating, feel free to drop a mail to me or here [4] or visit us in #hakyll on Freenode. [1]: http://hackage.haskell.org/package/hakyll [2]: http://jaspervdj.be/hakyll/reference/Hakyll-Core-Identifier-Pattern.html [3]: http://jaspervdj.be/hakyll/reference/Hakyll-Web-Tags.html [4]: https://groups.google.com/forum/#!forum/hakyll Cheers, Jasper

1 0

Quitting Yesod
by Aur Saraf 06 Apr '11

06 Apr '11

Hello, I'm trying to profile a yesod app, and it seems like Haskell's profiler only writes to file when the program exists. Gracefully. So Ctrl+C twice is out of the question. What I'm doing: $ ~/Downloads/spawn-fcgi-1.6.3/src/spawn-fcgi -n -s /tmp/f.socket -- ~/projects/f/dist/build/fastcgi/fastcgi +RTS -p Then I enter a page that I want to profile. Then my system starts swapping because the page takes 100% RAM (I'm suspecting a space leak). Then I hit Ctrl+C Ctrl+C to quit the yesod program. Then I look at fastcgi.prof, and find it empty. How can I get profiling info to be written to file? Am I wrong in something I'm doing? Thanks, Aur

3 3

Re: [web-devel] [Haskell-beginners] happstack versus yesod
by Michael Snoyman 06 Apr '11

06 Apr '11

This probably isn't the right forum for expressing this thought, but here goes anyway: What exactly is the freedom that Yesod takes away from you? I know this idea gets passed around a lot: Yesod isn't "really" Haskell, Yesod is its own language, etc. But all of that is really just talking about some sugar that Yesod puts in place for you: under the surface, it's just functions, data types and type classes. You can go ahead and write a complete Yesod application without using Template Haskell of QuasiQuotation. So what's the restriction that Yesod has that Happstack/Snap don't? Yesod requires you to provide some datatypes. That's it. What's the purpose of the datatypes? Mostly, it's about type-safe URLs. I'm not exactly sure why the term "restriction" or "freedom" even applies here: Yesod has a *feature* that Happstack/Snap do not include out of the box. Let me phrase this in a different way: "But I don't like Haskell, because the idea of Python in more interesting for me. Python gives you some freedom." I really do see this as the same argument. You can go ahead and right some code without static type checking in Python, and you might even write it faster than you would have in Haskell. But go ahead and start making changes, and unless you have some very good unit tests, you're in for a world of hurt. In Haskell, often times the compiler will save you from yourself. The same is true in Yesod. I've refactored incredibly large codebases, with hundreds of different resource patterns, and *never* had to worry about 404s. Maybe some people consider that giving up freedom. I consider it giving up the freedom to shoot yourself in the foot. Oh, and if you are absolutely dependent on the idea of getting rid of this data-type requirements, and want all the freedom of dynamic routes, you can try out a 44 line module[1] that I mentioned recently[2]. OK, enough of this rant. I'm not trying to attack you, I'm just tired of this notion that Yesod is a ball-and-chain around your ankles. People can have legitimate reasons to prefer Happstack or Snap to Yesod, I've never claimed that it was a perfect framework for everyone. I just don't think this is one of those legitimate reasons. Michael [1] https://gist.github.com/884802 <https://gist.github.com/884802>[2] http://osdir.com/ml/general/2011-04/msg02160.html On Tue, Apr 5, 2011 at 8:35 PM, Alexey G <kreed131(a)gmail.com> wrote: > Hello. It's very interesting theme for me and I want to express my > opinion. > Yesod - very fast web-framework( > http://www.yesodweb.com/blog/preliminary-warp-cross-language-benchmarks) > But I don't like Yesod, because idea of Happstack more interesting for me. > Happstack gives to you some freedom. > > I trying to use Snap Framework now. And it's very good for me. Try it > http://snapframework.com/! > > Sorry for my english. > > 2011/4/5 Ramy Abdel-Azim <ramy.abdelazim(a)gmail.com> > >> I already posted about this a little bit but i think i should let the old >> thread get back to it's regularly scheduled conversaion. on to my question: >> >> So I know there are a million easier ways to createa blog (wordpress, >> django, drupal, etc. etc.) but which would you recommend I use to build a >> beginner website that is simliar to a blog: happstack or yesod? >> _Ramy >> >> _______________________________________________ >> Beginners mailing list >> Beginners(a)haskell.org >> http://www.haskell.org/mailman/listinfo/beginners >> >> > > _______________________________________________ > Beginners mailing list > Beginners(a)haskell.org > http://www.haskell.org/mailman/listinfo/beginners > >

3 3

Yesod CRUD Help
by Francisco Jose CHAVES ALONSO 05 Apr '11

05 Apr '11

Hi In order to know how the CRUD works, I was trying the blog example of the Screencast, updated to yesod 7 but the mkToForm gives me the following error: No instance for (PersistEntity [Database.Persist.Base.EntityDef]) arising from a use of `mkToForm' Possible fix: add an instance declaration for (PersistEntity [Database.Persist.Base.EntityDef]) In the expression: mkToForm In the first argument of `share', namely `[mkToForm, mkPersist, mkMigrate "migrateAll"]' In the expression: share [mkToForm, mkPersist, mkMigrate "migrateAll"] [Database.Persist.Base.EntityDef "Entry" [] [("title", "String", []), ("day", "Day", ["Desc"]), ....] [] []] Any suggestion? Thanks in advance Francisco The code is: {-# LANGUAGE TypeFamilies, QuasiQuotes, GeneralizedNewtypeDeriving, TemplateHaskell, MultiParamTypeClasses #-} import Yesod import Yesod.Helpers.Crud import Database.Persist import Database.Persist.Sqlite import Database.Persist.TH import Data.Time (Day) import Network.Wai.Middleware.Debug (debug) import Network.Wai.Handler.Warp (run) share [mkToForm, mkPersist, mkMigrate "migrateAll"] [persist| Entry title String day Day Desc content Html deriving |] -- Necesary for CRUD instance Item Entry where itemTitle = entryTitle data Blog = Blog { pool :: ConnectionPool } type EntryCrud = Crud Blog Entry mkYesod "Blog" [parseRoutes| / RootR GET /entry/#EntryId EntryR GET -- Subsite /admin AdminR EntryCrud defaultCrud |] instance Yesod Blog where approot _ = "http://localhost:3000" instance YesodPersist Blog where type YesodDB Blog = SqlPersist runDB db = liftIOHandler $ fmap pool getYesod >>= runSqlPool db getRootR = do entries <- runDB $ selectList [] [EntryDayDesc] 0 0 defaultLayout $ do setTitle $ string "Yesod Blog Tutorial Homepage" [hamlet| <h1> Archive <ul> $forall entry <- entries <li> <a href="@{EntryR (fst entry)}">#{entryTitle (snd entry)} <p> <a href="@{AdminR CrudListR}">Admin |] getEntryR entryid = do entry <- runDB $ get404 entryid defaultLayout $ do setTitle $ string $ entryTitle entry [hamlet| <h1> #{entryTitle entry} <h2> #{show (entryDay entry)} #{entryContent entry} |] withBlog :: (Application -> IO a) -> IO a withBlog f = withSqlitePool "blog.db3" 8 $ \pool -> do flip runSqlPool pool $ do (runMigration migrateAll) insert $ Entry "First Entry" (read "2011-04-04") $ preEscapedString "<h3>First Entry</h3>" let h = Blog pool toWaiApp h >>= f main = withBlog $ run 3000 . debug

2 1

Auth plugin usage questions
by Ian Duncan 05 Apr '11

05 Apr '11

I currently have a function of type Widget (), and I need to find get the user's authorization (or not) in order to render the appropriate confidential information. How exactly do I get the authorization information? There's currently not really any documentation for this stuff, so I thought I'd ask and see what the best way to do this is.

2 1

Re: [web-devel] [Haskell-beginners] happstack versus yesod
by Michael Snoyman 05 Apr '11

05 Apr '11

I'm not sure where you got that implication from. Happstack is a good framework, and using it will definitely help you learn Haskell. But the same is true of Yesod or Snap. I guess you might say that if your goal is to learn Haskell, and not to create websites, then you shouldn't be using a framework at all, and should be writing code that reads/sends ByteStrings over sockets. You'll definitely learn more Haskell that way... Michael On Tue, Apr 5, 2011 at 9:50 PM, Ramy Abdel-Azim <ramy.abdelazim(a)gmail.com>wrote: > well, i'll add my $.02 here (hopefully without throwing fuel on the fire) > and say that - without knowing much about either, I asked the question for > pete's sake! - is that from what you're saying, if I'dlike to use one of > these frameworks to not only build my website but to also solidify what > knowledge of haskell i've acquired, I should go with Happstack. > > > On Tue, Apr 5, 2011 at 2:32 PM, Michael Snoyman <michael(a)snoyman.com>wrote: > >> This probably isn't the right forum for expressing this thought, but here >> goes anyway: >> >> What exactly is the freedom that Yesod takes away from you? I know this >> idea gets passed around a lot: Yesod isn't "really" Haskell, Yesod is its >> own language, etc. But all of that is really just talking about some sugar >> that Yesod puts in place for you: under the surface, it's just functions, >> data types and type classes. You can go ahead and write a complete Yesod >> application without using Template Haskell of QuasiQuotation. >> >> So what's the restriction that Yesod has that Happstack/Snap don't? Yesod >> requires you to provide some datatypes. That's it. What's the purpose of the >> datatypes? Mostly, it's about type-safe URLs. I'm not exactly sure why the >> term "restriction" or "freedom" even applies here: Yesod has a *feature* >> that Happstack/Snap do not include out of the box. >> >> Let me phrase this in a different way: "But I don't like Haskell, because >> the idea of Python in more interesting for me. Python gives you some >> freedom." I really do see this as the same argument. You can go ahead and >> right some code without static type checking in Python, and you might even >> write it faster than you would have in Haskell. But go ahead and start >> making changes, and unless you have some very good unit tests, you're in for >> a world of hurt. In Haskell, often times the compiler will save you from >> yourself. >> >> The same is true in Yesod. I've refactored incredibly large codebases, >> with hundreds of different resource patterns, and *never* had to worry about >> 404s. Maybe some people consider that giving up freedom. I consider it >> giving up the freedom to shoot yourself in the foot. >> >> Oh, and if you are absolutely dependent on the idea of getting rid of this >> data-type requirements, and want all the freedom of dynamic routes, you can >> try out a 44 line module[1] that I mentioned recently[2]. >> >> OK, enough of this rant. I'm not trying to attack you, I'm just tired of >> this notion that Yesod is a ball-and-chain around your ankles. People can >> have legitimate reasons to prefer Happstack or Snap to Yesod, I've never >> claimed that it was a perfect framework for everyone. I just don't think >> this is one of those legitimate reasons. >> >> Michael >> >> [1] https://gist.github.com/884802 >> <https://gist.github.com/884802>[2] >> http://osdir.com/ml/general/2011-04/msg02160.html >> >> >> On Tue, Apr 5, 2011 at 8:35 PM, Alexey G <kreed131(a)gmail.com> wrote: >> >>> Hello. It's very interesting theme for me and I want to express my >>> opinion. >>> Yesod - very fast web-framework( >>> http://www.yesodweb.com/blog/preliminary-warp-cross-language-benchmarks >>> ). >>> But I don't like Yesod, because idea of Happstack more interesting for >>> me. Happstack gives to you some freedom. >>> >>> I trying to use Snap Framework now. And it's very good for me. Try it >>> http://snapframework.com/! >>> >>> Sorry for my english. >>> >>> 2011/4/5 Ramy Abdel-Azim <ramy.abdelazim(a)gmail.com> >>> >>>> I already posted about this a little bit but i think i should let the >>>> old thread get back to it's regularly scheduled conversaion. on to my >>>> question: >>>> >>>> So I know there are a million easier ways to createa blog (wordpress, >>>> django, drupal, etc. etc.) but which would you recommend I use to build a >>>> beginner website that is simliar to a blog: happstack or yesod? >>>> _Ramy >>>> >>>> _______________________________________________ >>>> Beginners mailing list >>>> Beginners(a)haskell.org >>>> http://www.haskell.org/mailman/listinfo/beginners >>>> >>>> >>> >>> _______________________________________________ >>> Beginners mailing list >>> Beginners(a)haskell.org >>> http://www.haskell.org/mailman/listinfo/beginners >>> >>> >> >> _______________________________________________ >> Beginners mailing list >> Beginners(a)haskell.org >> http://www.haskell.org/mailman/listinfo/beginners >> >> >

1 0

yesod-dynamic
by Michael Snoyman 04 Apr '11

04 Apr '11

Hi all, There's an inherent difference in how Yesod handles routing, by default, and how Happstack/Snap handle routing, by default. (Apologies if I am mistaken in my understanding of the latter.) Yesod is built in with the concept of type-safe URLs. This forms the basis of routing in Yesod: the list of path segments is first converted to a type-safe URL value, and then the appropriate handler is chosen from that. On the other hand, Happstack/Snap both allow you to perform routing directly on the path segments. What many people don't realize is that it's entirely possible to do segment-based routing in Yesod. I don't recommend the practice in general, since it defeats a lot of the type safety Yesod gives you. However, in some situations it's a necessity. For example, yesod-auth allows arbitrary authentication plugins, which would not mesh up well with type-safe URLs. The trick to bypassing the type safety is to use multi-pieces. Instead of describing how this is possible, I will merely point to an example: https://gist.github.com/884802 The Yesod.Dynamic module defined there is really a general way to allow Yesod routing based on path segments. Since the question has come up a number of times in the past, I'm curious if people would want this available on hackage. I'm sure a nice set of routing operators would be useful here as well. Michael

4 4

update of the screencast Blog Example error
by Francisco Jose CHAVES ALONSO 04 Apr '11

04 Apr '11

I was trying to run the screencast blog 2 example in Yesod 0.7, but I get the following error: blog2.hs:11:8: No instance for (PersistEntity [Database.Persist.Base.EntityDef]) arising from a use of `mkToForm' Possible fix: add an instance declaration for (PersistEntity [Database.Persist.Base.EntityDef]) In the expression: mkToForm In the first argument of `share', namely `[mkToForm, mkPersist, mkMigrate "migrateAll"]' In the expression: share [mkToForm, mkPersist, mkMigrate "migrateAll"] [Database.Persist.Base.EntityDef "Entry" [] [("title", "String", []), ("day", "Day", ["Desc"]), ....] [] []] Following is the code: {-# LANGUAGE TypeFamilies, QuasiQuotes, GeneralizedNewtypeDeriving, TemplateHaskell, MultiParamTypeClasses #-} import Yesod import Yesod.Helpers.Crud import Database.Persist import Database.Persist.Sqlite import Database.Persist.TH import Data.Time (Day) import Network.Wai.Middleware.Debug (debug) import Network.Wai.Handler.Warp (run) share [mkToForm, mkPersist, mkMigrate "migrateAll"] [persist| Entry title String day Day Desc content Html deriving |] -- Necesary for CRUD instance Item Entry where itemTitle = entryTitle data Blog = Blog { pool :: ConnectionPool } type EntryCrud = Crud Blog Entry mkYesod "Blog" [parseRoutes| / RootR GET /entry/#EntryId EntryR GET -- Subsite /admin AdminR EntryCrud defaultCrud |] instance Yesod Blog where approot _ = "http://localhost:3000" instance YesodPersist Blog where type YesodDB Blog = SqlPersist runDB db = liftIOHandler $ fmap pool getYesod >>= runSqlPool db getRootR = do entries <- runDB $ selectList [] [EntryDayDesc] 0 0 defaultLayout $ do setTitle $ string "Yesod Blog Tutorial Homepage" [hamlet| <h1> Archive <ul> $forall entry <- entries <li> <a href="@{EntryR (fst entry)}">#{entryTitle (snd entry)} <p> <a href="@{AdminR CrudListR}">Admin |] getEntryR entryid = do entry <- runDB $ get404 entryid defaultLayout $ do setTitle $ string $ entryTitle entry [hamlet| <h1> #{entryTitle entry} <h2> #{show (entryDay entry)} #{entryContent entry} |] withBlog :: (Application -> IO a) -> IO a withBlog f = withSqlitePool "blog.db3" 8 $ \pool -> do flip runSqlPool pool $ do (runMigration migrateAll) insert $ Entry "First Entry" (read "2011-04-04") $ preEscapedString "<h3>First Entry</h3>" let h = Blog pool toWaiApp h >>= f main = withBlog $ run 3000 . debug

1 0

Cannot control order of XML attributes
by Yitzchak Gale 04 Apr '11

04 Apr '11

After upgrading to xml-enumerator 0.2.1 and xml-types 0.2, I can no longer control the order of attributes in rendered XML. The XML specification specifically allows this - you can change the order of attributes with affecting the semantics of the document. I need the attributes to be in a sane order for reading by humans so that the XML will pass my customer's acceptance testing. Earlier versions of xml-types supported this, but 0.2 introduced an API change which forces attribute sets to be expressed as a Data.Map. Why not just use association lists? That seems like a perfect fit. Or just go back to the old API. Thanks, Yitz

3 6

Installing Yesod on OSX
by Tristan Sloughter 04 Apr '11

04 Apr '11

On OSX 10.6.6 I'm trying to install Yesod through Cabal but it fails because its unable to build cryptocipher. Has anyone seen this or know an answer? Thanks, Tristan cabal-install version 0.8.2 using version 1.8.0.6 of the Cabal library Resolving dependencies... Configuring cryptocipher-0.2.7... Preprocessing library cryptocipher-0.2.7... Preprocessing executables for cryptocipher-0.2.7... Building cryptocipher-0.2.7... [1 of 8] Compiling Number.Serialize ( Number/Serialize.hs, dist/build/Number/Serialize.o ) [2 of 8] Compiling Number.Generate ( Number/Generate.hs, dist/build/Number/Generate.o ) [3 of 8] Compiling Number.ModArithmetic ( Number/ModArithmetic.hs, dist/build/Number/ModArithmetic.o ) [4 of 8] Compiling Crypto.Cipher.DSA ( Crypto/Cipher/DSA.hs, dist/build/Crypto/Cipher/DSA.o ) [5 of 8] Compiling Crypto.Cipher.RSA ( Crypto/Cipher/RSA.hs, dist/build/Crypto/Cipher/RSA.o ) [6 of 8] Compiling Crypto.Cipher.Camellia ( Crypto/Cipher/Camellia.hs, dist/build/Crypto/Cipher/Camellia.o ) [7 of 8] Compiling Crypto.Cipher.AES ( Crypto/Cipher/AES.hs, dist/build/Crypto/Cipher/AES.o ) Crypto/Cipher/AES.hs:228:30: Not in scope: `V.concat' Crypto/Cipher/AES.hs:229:30: Not in scope: `V.concat' Crypto/Cipher/AES.hs:230:30: Not in scope: `V.concat' Configuring language-javascript-0.4.3... cabal: happy is required but it could not be found. cabal: Error: some packages failed to install: authenticate-0.8.2.2 depends on cryptocipher-0.2.7 which failed to install. cryptocipher-0.2.7 failed during the building phase. The exception was: ExitFailure 1 hjsmin-0.0.13 depends on language-javascript-0.4.3 which failed to install. http-enumerator-0.3.1 depends on cryptocipher-0.2.7 which failed to install. language-javascript-0.4.3 failed during the configure step. The exception was: ExitFailure 1 tls-0.3.3 depends on cryptocipher-0.2.7 which failed to install. yesod-0.7.3 depends on language-javascript-0.4.3 which failed to install. yesod-auth-0.3.2 depends on cryptocipher-0.2.7 which failed to install.

3 10