- web-devel - Haskell.org

wai 0.4.0 uprade for warp 0.8.0 breaks hoogle
by Mark Wright 25 Apr '11

25 Apr '11

Hi, I try to bump warp to 0.8.0 in the gentoo haskell overlay, which requires wai 0.4.0, this breaks hoogle. This is my incomplete and failed attempt to bump wai to 0.4.0 in hoogle: --- hoogle-4.2.1-orig/hoogle.cabal 2011-01-26 22:16:52.000000000 +1100 +++ hoogle-4.2.1/hoogle.cabal 2011-04-20 23:59:18.021501043 +1000 @@ -35,7 +35,7 @@ array, bytestring, containers, directory, filepath, process, random, safe, binary, - parsec == 2.1.*, + parsec >= 2.1, transformers == 0.2.*, uniplate == 1.6.*, haskell-src-exts >= 1.9 && < 1.11 @@ -107,9 +107,11 @@ cmdargs == 0.6.*, tagsoup >= 0.11 && < 0.13, enumerator == 0.4.*, - blaze-builder == 0.2.*, - wai == 0.3.0, - warp == 0.3.0, + blaze-builder >= 0.2 && < 0.4, + http-types >= 0.6 && < 0.7, + case-insensitive >= 0.2 && < 0.3, + wai >= 0.4.0 && < 0.5.0, + warp >= 0.4.0 && < 0.5.0, Cabal >= 1.8 && < 1.11 other-modules: --- hoogle-4.2.1-orig/src/General/Web.hs 2011-01-26 22:16:52.000000000 +1100 +++ hoogle-4.2.1/src/General/Web.hs 2011-04-21 00:03:49.840705869 +1000 @@ -16,6 +16,8 @@ import General.System import General.Base import Network.Wai +import Network.HTTP.Types +import Data.CaseInsensitive(original) import Blaze.ByteString.Builder(toLazyByteString) import Data.Enumerator.List(consume) import qualified Data.ByteString.Lazy.Char8 as LBS @@ -23,13 +25,15 @@ type Args = [(String, String)] +type ResponseHeader = Header --------------------------------------------------------------------- -- WAI STUFF -statusOK = status200 -hdrContentType = fromString "Content-Type" :: ResponseHeader -hdrCacheControl = fromString "Cache-Control" :: ResponseHeader +hdrContentType :: Ascii -> Header +hdrContentType = headerContentType +hdrCacheControl :: Ascii -> Header +hdrCacheControl = headerCacheControl responseOK = responseLBS statusOK responseBadRequest x = responseLBS status400 [] $ fromString $ "Bad request: " ++ x @@ -125,7 +129,7 @@ cgiResponse r = do (status,headers,body) <- responseFlatten r LBS.putStr $ LBS.unlines $ - [LBS.fromChunks [ciOriginal a, fromString ": ", b] | (a,b) <- headers] ++ + [LBS.fromChunks [original a, fromString ": ", b] | (a,b) <- headers] ++ [fromString "",body] --- hoogle-4.2.1-orig/src/Web/Response.hs 2011-01-26 22:16:52.000000000 +1100 +++ hoogle-4.2.1/src/Web/Response.hs 2011-04-20 23:06:21.128254027 +1000 @@ -15,6 +15,7 @@ import Data.Time.Format import System.Locale import Network.Wai +import Network.HTTP.Types(headerContentType) import System.IO.Unsafe(unsafeInterleaveIO) @@ -24,7 +25,7 @@ response :: FilePath -> CmdLine -> IO Response response resources q = do logMessage q - let response x ys z = responseOK ((hdrContentType,fromString x) : ys) (fromString z) + let response x ys z = responseOK ((headerContentType $ fromString x) : ys) (fromString z) dbs <- unsafeInterleaveIO $ case queryParsed q of Left _ -> return mempty --- hoogle-4.2.1-orig/src/Web/Server.hs 2011-01-26 22:16:52.000000000 +1100 +++ hoogle-4.2.1/src/Web/Server.hs 2011-04-20 23:55:34.234388414 +1000 @@ -15,19 +15,18 @@ import qualified Data.ByteString.Lazy.Char8 as LBS import qualified Data.ByteString.Char8 as BS - server :: CmdLine -> IO () server q@Server{..} = do v <- newMVar () putStrLn $ "Starting Hoogle Server on port " ++ show port run port $ \r -> liftIO $ do - withMVar v $ const $ putStrLn $ bsUnpack (pathInfo r) ++ bsUnpack (queryString r) + withMVar v $ const $ putStrLn $ bsUnpack (rawPathInfo r) ++ bsUnpack (rawQueryString r) talk q r -- FIXME: Avoid all the conversions to/from LBS talk :: CmdLine -> Request -> IO Response -talk Server{..} Request{pathInfo=path_, queryString=query_} +talk Server{..} Request{rawPathInfo=path_, rawQueryString=query_} | path `elem` ["/","/hoogle"] = do let args = parseHttpQueryArgs $ drop 1 query cmd <- cmdLineWeb args The compiler error is: [61 of 72] Compiling Web.Server ( src/Web/Server.hs, dist/build/hoogle/hoogle-tmp/Web/Server.o ) src/Web/Server.hs:46:14: Couldn't match expected type `Response' with actual type `Maybe FilePart -> Response' In the return type of a call of `ResponseFile' In the expression: ResponseFile statusOK hdr file In the second argument of `($)', namely `if not b then responseNotFound file else ResponseFile statusOK hdr file' I was wondering how to convert from Maybe FilePart to Response, this looks tricky. Thanks, Mark PS: I have built all of yesod 0.8.0 for gentoo, but we would need to somehow fix hoogle in order for me to apply these changes to the gentoo haskell overlay.

3 6

Persistent Value. convert from
by Anton Cheshkov 25 Apr '11

25 Apr '11

Hello Michael. Hello All ! I want to ask you about the best way to convert from Int to Persistent Key. by now I see the single way *fromSinglePiece . Text.pack . show* **but it seems not so pretty Are there the alternative way ? Thanks. -- Best regards, Cheshkov Anton Phone: +7 909 005 18 82 Skype: cheshkov_anton

2 1

utf8-light fails to install
by vagif.verdi＠gmail.com 23 Apr '11

23 Apr '11

arch linux 64 bit, ghc 7.0.2. Deleted .cabal and .ghc folders. cabal update cabal install yesod. Ends with this error: Registering yesod-auth-0.4.0... Installing library in /home/vagif/.cabal/lib/yesod-auth-0.4.0/ghc-7.0.2 Registering yesod-auth-0.4.0... cabal: Error: some packages failed to install: hjsmin-0.0.14 depends on utf8-light-0.4 which failed to install. language-javascript-0.4.4 depends on utf8-light-0.4 which failed to install. utf8-light-0.4 failed during the building phase. The exception was: ExitFailure 1 yesod-0.8.0 depends on utf8-light-0.4 which failed to install. Tried 2 times. Same error. Regards, Vagif Verdi

2 4

Authorization/Permissions for Persistent
by Max Cantor 23 Apr '11

23 Apr '11

We have what is becoming a rather large webapp using yesod/persistent but have been doing authorization in a rather ad-hoc way. I'm wondering if anyone else has dealt with this problem and has a smarter way to do it. Here's a summary of the issue. In persistent we have a User table and we pull UserIds from the maybeAuth or requireAuth functions in yesod.auth. We then have some other tables which might reference the userId directly or reference something else. Based on the userId, the user either should have no access, read access, or write access to that row. doesn't need to be a perfect or foolproof solution but something better than our current, completely ad-hoc approach would be an improvement. max

4 3

yesod devel
by Justin Greene 23 Apr '11

23 Apr '11

I'm running into an issue with the new development server on Windows 7 x64. I've installed the Haskell Platform 2011.2.0.1 run cabal update, and then cabal installed yesod. I then ran yesod init and cabal install. Finally I ran yesod devel on the newly created site. The site came up the first time, but when I went to edit and save a hamlet template the first time and refreshed the page it immediately showed me "System.PosixCompat.Files.setFileTimes: not supported: illegal operation". Does anyone have any ideas how to get around this? Thanks,

3 2

Workaround for "ghc: panic! (the 'impossible' happened)" when building yesod 0.8 on ghc 7.0.2 on arch (bug 5004)
by Max Cantor 23 Apr '11

23 Apr '11

Has anyone run into this bug (http://hackage.haskell.org/trac/ghc/ticket/5004) that prevents yesod from building on arch? if so, whats the work around? max

2 3

base-unicode-symbols not compiling
by Michael Snoyman 22 Apr '11

22 Apr '11

Hi Bas, I'm getting sporadic bug reports about base-unicode-symbols not compiling, which is preventing Yesod from compiling. I actually got the same error on GHC 6.12.1, and hacked around it by manually modifying the file to avoid the problematic CPP code. You can see a report in a Yesod book comment[1]. I think I remember you saying this is a problem with the CPP in GHC 6.12.1. Do you know if there's any workaround? Thanks, Michael [1] http://www.yesodweb.com/book/migrate7to8#iy1

3 5

Yesod 0.8 migration guide (release candidate?)
by Michael Snoyman 21 Apr '11

21 Apr '11

Hi all, I've put up the Yesod 0.8 migration guide[1]. I'm planning on releasing this version of Yesod tomorrow night (Saturday, UTC+3). Now's your chance to try and find those last few bugs before release. I have one idea that I wanted to bounce around. The biggest thing in this release is the transition to Text. Overall, this is a good move, but it's just a little bit annoying to lose out on convenience functions like concat and ++. As a convenience, I'm considering exporting the following from the Yesod module: mconcat (from Data.Monoid) <> = mappend (from Data.Monoid) Text (the strict version, from Data.Text) I know there was some talk a while ago of providing a <> operator in base, but I haven't heard anything about it since. I think this could be a very nice convenience, and would appreciate feedback. Michael [1] http://www.yesodweb.com/book/migrate7to8

3 5

Re: [web-devel] base-unicode-symbols not compiling
by Chris Smith 21 Apr '11

21 Apr '11

I know in one application of mine, I had to add an explicit dependency for version 0.2.1.1 in order to get it to compile on GHC 6.12.3. That might be worth a shot. I'd heard that there was a fix in a later version for the issue I had, but if you're still getting failures, perhaps adding a version bound would help. Then again, this may be related to people not running cabal update since the fix was released... so maybe the first answer is to suggest a cabal update. On Apr 20, 2011 10:00 PM, "Michael Snoyman" <michael(a)snoyman.com> wrote: > Hi Bas, > > I'm getting sporadic bug reports about base-unicode-symbols not > compiling, which is preventing Yesod from compiling. I actually got > the same error on GHC 6.12.1, and hacked around it by manually > modifying the file to avoid the problematic CPP code. You can see a > report in a Yesod book comment[1]. > > I think I remember you saying this is a problem with the CPP in GHC > 6.12.1. Do you know if there's any workaround? > > Thanks, > Michael > > [1] http://www.yesodweb.com/book/migrate7to8#iy1 > > _______________________________________________ > web-devel mailing list > web-devel(a)haskell.org > http://www.haskell.org/mailman/listinfo/web-devel

2 1

preEscapedString :: String -> Html in hamlet 0.4.0
by Max Cantor 20 Apr '11

20 Apr '11

seems like preEscapedString didn't get the message about the migration to Data.Text. is this intentional or will it be fixed in 0.4.1? chag sameach max

2 1