From 3bd8fba2db932b7730ba497d60030db6ee6f6405 Mon Sep 17 00:00:00 2001 From: Joey Hess Date: Sun, 27 Jan 2013 09:30:53 +1100 Subject: addurl --fast: Use curl, rather than haskell HTTP library, to support https. --- Makefile | 1 - Utility/Url.hs | 87 ++++++++++------------------------- debian/changelog | 1 + debian/control | 1 - debian/rules | 4 +- doc/install/fromscratch.mdwn | 1 - doc/todo/wishlist:_addurl_https:.mdwn | 3 ++ 7 files changed, 31 insertions(+), 67 deletions(-) diff --git a/Makefile b/Makefile index e1f9724a3..41ebc7dc6 100644 --- a/Makefile +++ b/Makefile @@ -6,7 +6,6 @@ BASEFLAGS=-Wall -outputdir $(GIT_ANNEX_TMP_BUILD_DIR) -IUtility # you can turn off some of these features. # # If you're using an old version of yesod, enable -DWITH_OLD_YESOD -# Or with an old version of the uri library, enable -DWITH_OLD_URI FEATURES?=$(GIT_ANNEX_LOCAL_FEATURES) -DWITH_ASSISTANT -DWITH_S3 -DWITH_WEBDAV -DWITH_WEBAPP -DWITH_PAIRING -DWITH_XMPP -DWITH_DNS bins=git-annex diff --git a/Utility/Url.hs b/Utility/Url.hs index 67efdb558..8b924a699 100644 --- a/Utility/Url.hs +++ b/Utility/Url.hs @@ -16,8 +16,6 @@ module Utility.Url ( ) where import Common -import qualified Network.Browser as Browser -import Network.HTTP import Network.URI import Data.Either @@ -38,20 +36,34 @@ check url headers expected_size = handle <$> exists url headers - also returning its size if available. -} exists :: URLString -> Headers -> IO (Bool, Maybe Integer) exists url headers = case parseURI url of - Nothing -> return (False, Nothing) Just u | uriScheme u == "file:" -> do s <- catchMaybeIO $ getFileStatus (uriPath u) - return $ case s of - Nothing -> (False, Nothing) - Just stat -> (True, Just $ fromIntegral $ fileSize stat) + case s of + Just stat -> return (True, Just $ fromIntegral $ fileSize stat) + Nothing -> dne | otherwise -> do - r <- request u headers HEAD - case rspCode r of - (2,_,_) -> return (True, size r) - _ -> return (False, Nothing) + output <- readProcess "curl" curlparams + case lastMaybe (lines output) of + Just ('2':_:_) -> return (True, extractsize output) + _ -> dne + Nothing -> dne where - size = liftM Prelude.read . lookupHeader HdrContentLength . rspHeaders + dne = return (False, Nothing) + + curlparams = + [ "-s" + , "--head" + , "-L" + , url + , "-w", "%{http_code}" + ] ++ concatMap (\h -> ["-H", h]) headers + + extractsize s = case lastMaybe $ filter ("Content-Length:" `isPrefixOf`) (lines s) of + Just l -> case lastMaybe $ words l of + Just sz -> readish sz + _ -> Nothing + _ -> Nothing {- Used to download large files, such as the contents of keys. - @@ -80,54 +92,5 @@ download url headers options file {- Downloads a small file. -} get :: URLString -> Headers -> IO String -get url headers = - case parseURI url of - Nothing -> error "url parse error" - Just u -> do - r <- request u headers GET - case rspCode r of - (2,_,_) -> return $ rspBody r - _ -> error $ rspReason r - -{- Makes a http request of an url. For example, HEAD can be used to - - check if the url exists, or GET used to get the url content (best for - - small urls). - - - - This does its own redirect following because Browser's is buggy for HEAD - - requests. - -} -request :: URI -> Headers -> RequestMethod -> IO (Response String) -request url headers requesttype = go 5 url - where - go :: Int -> URI -> IO (Response String) - go 0 _ = error "Too many redirects " - go n u = do - rsp <- Browser.browse $ do - Browser.setErrHandler ignore - Browser.setOutHandler ignore - Browser.setAllowRedirects False - let req = mkRequest requesttype u :: Request_String - snd <$> Browser.request (addheaders req) - case rspCode rsp of - (3,0,x) | x /= 5 -> redir (n - 1) u rsp - _ -> return rsp - ignore = const noop - redir n u rsp = case retrieveHeaders HdrLocation rsp of - [] -> return rsp - (Header _ newu:_) -> - case parseURIReference newu of - Nothing -> return rsp - Just newURI -> go n newURI_abs - where -#if defined VERSION_network -#if ! MIN_VERSION_network(2,4,0) -#define WITH_OLD_URI -#endif -#endif -#ifdef WITH_OLD_URI - newURI_abs = fromMaybe newURI (newURI `relativeTo` u) -#else - newURI_abs = newURI `relativeTo` u -#endif - addheaders req = setHeaders req (rqHeaders req ++ userheaders) - userheaders = rights $ map parseHeader headers +get url headers = readProcess "curl" $ + ["-s", "-L", url] ++ concatMap (\h -> ["-H", h]) headers diff --git a/debian/changelog b/debian/changelog index 411f12a8c..df3021d69 100644 --- a/debian/changelog +++ b/debian/changelog @@ -3,6 +3,7 @@ git-annex (3.20130125) UNRELEASED; urgency=low * webapp: Now allows restarting any threads that crash. * Adjust debian package to only build-depend on DAV on architectures where it is available. + * addurl --fast: Use curl, rather than haskell HTTP library, to support https. -- Joey Hess Sat, 26 Jan 2013 15:48:40 +1100 diff --git a/debian/control b/debian/control index cf25d65df..2610ca53b 100644 --- a/debian/control +++ b/debian/control @@ -10,7 +10,6 @@ Build-Depends: libghc-pcre-light-dev, libghc-sha-dev, libghc-dataenc-dev, - libghc-http-dev, libghc-utf8-string-dev, libghc-hs3-dev (>= 0.5.6), libghc-dav-dev (>= 0.3) [amd64 i386 kfreebsd-amd64 kfreebsd-i386 sparc], diff --git a/debian/rules b/debian/rules index 8cd2b5aa5..1dd64872d 100755 --- a/debian/rules +++ b/debian/rules @@ -1,9 +1,9 @@ #!/usr/bin/make -f ifeq (install ok installed,$(shell dpkg-query -W -f '$${Status}' libghc-yesod-dev 2>/dev/null)) -export FEATURES=-DWITH_ASSISTANT -DWITH_S3 -DWITH_HOST -DWITH_OLD_URI -DWITH_PAIRING -DWITH_XMPP -DWITH_WEBAPP -DWITH_OLD_YESOD +export FEATURES=-DWITH_ASSISTANT -DWITH_S3 -DWITH_HOST -DWITH_PAIRING -DWITH_XMPP -DWITH_WEBAPP -DWITH_OLD_YESOD else -export FEATURES=-DWITH_ASSISTANT -DWITH_S3 -DWITH_HOST -DWITH_OLD_URI -DWITH_PAIRING -DWITH_XMPP +export FEATURES=-DWITH_ASSISTANT -DWITH_S3 -DWITH_HOST -DWITH_PAIRING -DWITH_XMPP endif ifeq (install ok installed,$(shell dpkg-query -W -f '$${Status}' libghc-dav-dev 2>/dev/null)) export FEATURES:=${FEATURES} -DWITH_WEBDAV diff --git a/doc/install/fromscratch.mdwn b/doc/install/fromscratch.mdwn index 91a0d7f61..8fdb8dcda 100644 --- a/doc/install/fromscratch.mdwn +++ b/doc/install/fromscratch.mdwn @@ -13,7 +13,6 @@ quite a lot. * [lifted-base](http://hackage.haskell.org/package/lifted-base) * [TestPack](http://hackage.haskell.org/cgi-bin/hackage-scripts/package/testpack) * [QuickCheck 2](http://hackage.haskell.org/package/QuickCheck) - * [HTTP](http://hackage.haskell.org/package/HTTP) * [json](http://hackage.haskell.org/package/json) * [IfElse](http://hackage.haskell.org/package/IfElse) * [bloomfilter](http://hackage.haskell.org/package/bloomfilter) diff --git a/doc/todo/wishlist:_addurl_https:.mdwn b/doc/todo/wishlist:_addurl_https:.mdwn index bfb1c9283..0a62eda6d 100644 --- a/doc/todo/wishlist:_addurl_https:.mdwn +++ b/doc/todo/wishlist:_addurl_https:.mdwn @@ -6,3 +6,6 @@ To give an example, here is a PDF file: If you switch the https: to http: it redirects you back to https:. As more sites provide https: for non-secret traffic, this becomes more of an issue. + +> I've gotten rid of the use of the HTTP library, now it just uses curl. +> [[done]] --[[Joey]] -- cgit v1.2.3