Runs a Snap monad action in the 'Iteratee IO' monad.

Snap is the Monad that user web handlers run in. Snap gives you:

1. stateful access to fetch or modify an HTTP Request
2. stateful access to fetch or modify an HTTP Response
3. failure / Alternative / MonadPlus semantics: a Snap handler can choose not to handle a given request, using empty or its synonym pass, and you can try alternative handlers with the <|> operator:

 a :: Snap String
 a = pass

 b :: Snap String
 b = return "foo"

 c :: Snap String
 c = a <|> b             -- try running a, if it fails then try b

1. convenience functions (writeBS, writeLBS, writeText, writeLazyText, addToOutput) for queueing output to be written to the Response:

 a :: (forall a . Enumerator a) -> Snap ()
 a someEnumerator = do
     writeBS "I'm a strict bytestring"
     writeLBS "I'm a lazy bytestring"
     writeText "I'm strict text"
     addToOutput someEnumerator

1. early termination: if you call finishWith:

 a :: Snap ()
 a = do
   modifyResponse $ setResponseStatus 500 "Internal Server Error"
   writeBS "500 error"
   r <- getResponse
   finishWith r

then any subsequent processing will be skipped and supplied Response value will be returned from runSnap as-is.

1. access to the IO monad through a MonadIO instance:

 a :: Snap ()
 a = liftIO fireTheMissiles

1. the ability to set a timeout which will kill the handler thread after N seconds of inactivity (the default is 20 seconds):

 a :: Snap ()
 a = setTimeout 30

1. throw and catch exceptions using a MonadCatchIO instance:

 foo :: Snap ()
 foo = bar `catch` \(e::SomeException) -> baz
   where
     bar = throw FooException

1. log a message to the error log:

 foo :: Snap ()
 foo = logError "grumble."

You may notice that most of the type signatures in this module contain a (MonadSnap m) => ... typeclass constraint. MonadSnap is a typeclass which, in essence, says "you can get back to the Snap monad from here". Using MonadSnap you can extend the Snap monad with additional functionality and still have access to most of the Snap functions without writing lift everywhere. Instances are already provided for most of the common monad transformers (ReaderT, WriterT, StateT, etc.).

MonadSnap is a type class, analogous to MonadIO for IO, that makes it easy to wrap Snap inside monad transformers.

This exception is thrown if the handler you supply to runSnap fails.

This is exception is thrown if the handler chooses to escape regular HTTP traffic.

This function brackets a Snap action in resource acquisition and release. This is provided because MonadCatchIO's bracket function doesn't work properly in the case of a short-circuit return from the action being bracketed.

In order to prevent confusion regarding the effects of the aquisition and release actions on the Snap state, this function doesn't accept Snap actions for the acquire or release actions.

This function will run the release action in all cases where the acquire action succeeded. This includes the following behaviors from the bracketed Snap action.

1. Normal completion
2. Short-circuit completion, either from calling fail or finishWith
3. An exception being thrown.

Short-circuits a Snap monad action early, storing the given Response value in its state.

Capture the flow of control in case a handler calls finishWith.

WARNING: in the event of a call to transformRequestBody it is possible to violate HTTP protocol safety when using this function. If you call catchFinishWith it is suggested that you do not modify the body of the Response which was passed to the finishWith call.

Fails out of a Snap monad action. This is used to indicate that you choose not to handle the given request within the given handler.

Terminate the HTTP session with the given exception.

Terminate the HTTP session and hand control to some external handler, escaping all further HTTP traffic.

The external handler takes two arguments: a function to tickle the timeout manager, and a write end to the socket.

Runs a Snap monad action only if the request's HTTP method matches the given method.

Runs a Snap monad action only if the request's HTTP method matches one of the given methods.

Runs a Snap monad action only for requests where rqPathInfo is exactly equal to the given string. If the path matches, locally sets rqContextPath to the old value of rqPathInfo, sets rqPathInfo="", and runs the given handler.

Runs a Snap monad action only when the first path component is successfully parsed as the argument to the supplied handler function.

Runs a Snap monad action only when the rqPathInfo of the request starts with the given path. For example,

 dir "foo" handler

Will fail if rqPathInfo is not "/foo" or "/foo/...", and will add "foo/" to the handler's local rqContextPath.

Runs a Snap monad action only when rqPathInfo is empty.

A web handler which, given a mapping from URL entry points to web handlers, efficiently routes requests to the correct handler.

The URL entry points are given as relative paths, for example:

 route [ ("foo/bar/quux", fooBarQuux) ]

If the URI of the incoming request is

 /foo/bar/quux

or

 /foo/bar/quux/...anything...

then the request will be routed to "fooBarQuux", with rqContextPath set to "/foo/bar/quux/" and rqPathInfo set to "...anything...".

A path component within an URL entry point beginning with a colon (":") is treated as a variable capture; the corresponding path component within the request URI will be entered into the rqParams parameters mapping with the given name. For instance, if the routes were:

 route [ ("foo/:bar/baz", fooBazHandler) ]

Then a request for "/foo/saskatchewan/baz" would be routed to fooBazHandler with a mapping for:

 "bar" => "saskatchewan"

in its parameters table.

Longer paths are matched first, and specific routes are matched before captures. That is, if given routes:

 [ ("a", h1), ("a/b", h2), ("a/:x", h3) ]

a request for "/a/b" will go to h2, "/a/s" for any s will go to h3, and "/a" will go to h1.

The following example matches "/article" to an article index, "/login" to a login, and "/article/..." to an article renderer.

 route [ ("article",     renderIndex)
       , ("article/:id", renderArticle)
       , ("login",       method POST doLogin) ]

URL decoding

A short note about URL decoding: path matching and variable capture are done on decoded URLs, but the contents of rqContextPath and rqPathInfo will contain the original encoded URL, i.e. what the user entered. For example, in the following scenario:

 route [ ("a b c d/", foo ) ]

A request for "/a+b+c+d" will be sent to foo with rqContextPath set to "a+b+c+d".

This behaviour changed as of Snap 0.6.1; previous versions had unspecified (and buggy!) semantics here.

The routeLocal function is the same as route', except it doesn't change the request's context path. This is useful if you want to route to a particular handler but you want that handler to receive the rqPathInfo as it is.

Grabs the Request object out of the Snap monad.

Grabs something out of the Request object, using the given projection function. See gets.

Grabs the Response object out of the Snap monad.

Grabs something out of the Response object, using the given projection function. See gets.

Puts a new Request object into the Snap monad.

Puts a new Response object into the Snap monad.

Modifies the Request object stored in a Snap monad.

Modifes the Response object stored in a Snap monad.

Runs a Snap action with a locally-modified Request state object. The Request object in the Snap monad state after the call to localRequest will be unchanged.

Fetches the Request from state and hands it to the given action.

Fetches the Response from state and hands it to the given action.

Log an error message in the Snap monad

Sends the request body through an iteratee (data consumer) and returns the result.

If the iteratee you pass in here throws an exception, Snap will attempt to clear the rest of the unread request body before rethrowing the exception. If your iteratee used terminateConnection, however, Snap will give up and immediately close the socket.

Returns the request body as a lazy bytestring.

This function is deprecated as of 0.6; it places no limits on the size of the request being read, and as such, if used, can result in a denial-of-service attack on your server. Please use readRequestBody instead.

Returns the request body as a lazy bytestring. New in 0.6.

Normally Snap is careful to ensure that the request body is fully consumed after your web handler runs, but before the Response enumerator is streamed out the socket. If you want to transform the request body into some output in O(1) space, you should use this function.

Note that upon calling this function, response processing finishes early as if you called finishWith. Make sure you set any content types, headers, cookies, etc. before you call this function.

Contains all of the information about an incoming HTTP request.

Represents an HTTP response.

A typeclass for datatypes which contain HTTP headers.

A type alias for the HTTP parameters mapping. Each parameter key maps to a list of ByteString values; if a parameter is specified multiple times (e.g.: "GET /foo?param=bar1&param=bar2"), looking up "param" in the mapping will give you ["bar1", "bar2"].

Enumerates the HTTP method values (see http://tools.ietf.org/html/rfc2068.html#section-5.1.1).

A datatype representing an HTTP cookie.

Adds a header key-value-pair to the HasHeaders datatype. If a header with the same name already exists, the new value is appended to the headers list.

Sets a header key-value-pair in a HasHeaders datatype. If a header with the same name already exists, it is overwritten with the new value.

Gets a header value out of a HasHeaders datatype. If many headers came in with the same name, they will be catenated together.

Gets all of the values for a given header.

Lists all the headers out of a HasHeaders datatype. If many headers came in with the same name, they will be catenated together.

Clears a header value from a HasHeaders datatype.

Modifies the Request in the state to set the rqRemoteAddr field to the value in the X-Forwarded-For header. If the header is not present, this action has no effect.

This action should be used only when working behind a reverse http proxy that sets the X-Forwarded-For header. This is the only way to ensure the value in the X-Forwarded-For header can be trusted.

This is provided as a filter so actions that require the remote address can get it in a uniform manner. It has specifically limited functionality to ensure that its transformation can be trusted, when used correctly.

Modifies the Request in the state to set the rqRemoteAddr field to the value from the header specified. If the header specified is not present, this action has no effect.

This action should be used only when working behind a reverse http proxy that sets the header being looked at. This is the only way to ensure the value in the header can be trusted.

This is provided as a filter so actions that require the remote address can get it in a uniform manner. It has specifically limited functionality to ensure that its transformation can be trusted, when used correctly.

The server name of the request, as it came in from the request's Host: header.

Returns the port number the HTTP server is listening on.

The remote IP address.

The remote TCP port number.

The local IP address for this request.

Returns the HTTP server's idea of its local hostname.

Returns True if this is an HTTPS session.

Returns the Content-Length of the HTTP request body.

Returns the HTTP request method.

Returns the HTTP version used by the client.

Returns a list of the cookies that came in from the HTTP request headers.

Handlers can (will be; --ed) be hung on a URI "entry point"; this is called the "context path". If a handler is hung on the context path "/foo/", and you request "/foo/bar", the value of rqPathInfo will be "bar".

The "context path" of the request; catenating rqContextPath, and rqPathInfo should get you back to the original rqURI (ignoring query strings). The rqContextPath always begins and ends with a slash ("/") character, and represents the path (relative to your component/snaplet) you took to get to your handler.

Returns the URI requested by the client.

Returns the HTTP query string for this Request.

Returns the Params mapping for this Request. "Parameters" are automatically decoded from the query string and POST body and entered into this mapping.

Looks up the value(s) for the given named parameter. Parameters initially come from the request's query string and any decoded POST body (if the request's Content-Type is application/x-www-form-urlencoded). Parameter values can be modified within handlers using rqModifyParams.

See rqParam. Looks up a value for the given named parameter in the Request. If more than one value was entered for the given parameter name, getParam gloms the values together with:

    intercalate " "

See rqParams. Convenience function to return Params from the Request inside of a MonadSnap instance.

Modifies the parameters mapping (which is a Map ByteString ByteString) in a Request using the given function.

Writes a key-value pair to the parameters mapping within the given request.

An empty Response.

Sets the HTTP response code.

Sets the HTTP response status. Note: normally you would use setResponseCode unless you needed a custom response explanation.

Returns the HTTP status code.

Returns the HTTP status explanation string.

Sets the Content-Type in the Response headers.

Adds an HTTP Cookie to Response headers.

Gets an HTTP Cookie with the given name from Response headers.

Returns a list of Cookies present in Response

Deletes an HTTP Cookie from the Response headers. Please note this does not necessarily erase the cookie from the client browser.

Modifies an HTTP Cookie with given name in Response headers. Nothing will happen if a matching Cookie can not be found in Response.

Expire the given Cookie in client's browser.

Gets the HTTP Cookie with the specified name.

Gets the HTTP Cookie with the specified name and decodes it. If the decoding fails, the handler calls pass.

A note here: if you want to set the Content-Length for the response, Snap forces you to do it with this function rather than by setting it in the headers; the Content-Length in the headers will be ignored.

The reason for this is that Snap needs to look up the value of Content-Length for each request, and looking the string value up in the headers and parsing the number out of the text will be too expensive.

If you don't set a content length in your response, HTTP keep-alive will be disabled for HTTP/1.0 clients, forcing a Connection: close. For HTTP/1.1 clients, Snap will switch to the chunked transfer encoding if Content-Length is not specified.

Removes any Content-Length set in the Response.

Performs a redirect by setting the Location header to the given target URL/path and the status code to 302 in the Response object stored in a Snap monad. Note that the target URL is not validated in any way. Consider using 'redirect\'' instead, which allows you to choose the correct status code.

Performs a redirect by setting the Location header to the given target URL/path and the status code (should be one of 301, 302, 303 or 307) in the Response object stored in a Snap monad. Note that the target URL is not validated in any way.

Sets an HTTP response body to the given Enumerator value.

Modifies a response body.

Adds the output from the given enumerator to the Response stored in the Snap monad state.

Adds the given Builder to the body of the Response stored in the | Snap monad state.

Adds the given strict ByteString to the body of the Response stored in the Snap monad state.

Warning: This function is intentionally non-strict. If any pure exceptions are raised by the expression creating the ByteString, the exception won't actually be raised within the Snap handler.

Adds the given lazy Text to the body of the Response stored in the Snap monad state.

Warning: This function is intentionally non-strict. If any pure exceptions are raised by the expression creating the ByteString, the exception won't actually be raised within the Snap handler.

Adds the given strict Text to the body of the Response stored in the Snap monad state.

Warning: This function is intentionally non-strict. If any pure exceptions are raised by the expression creating the ByteString, the exception won't actually be raised within the Snap handler.

Adds the given lazy ByteString to the body of the Response stored in the Snap monad state.

Warning: This function is intentionally non-strict. If any pure exceptions are raised by the expression creating the ByteString, the exception won't actually be raised within the Snap handler.

Sets the output to be the contents of the specified file.

Calling sendFile will overwrite any output queued to be sent in the Response. If the response body is not modified after the call to sendFile, Snap will use the efficient sendfile() system call on platforms that support it.

If the response body is modified (using modifyResponseBody), the file will be read using mmap().

Sets the output to be the contents of the specified file, within the given (start,end) range.

Calling sendFilePartial will overwrite any output queued to be sent in the Response. If the response body is not modified after the call to sendFilePartial, Snap will use the efficient sendfile() system call on platforms that support it.

If the response body is modified (using modifyResponseBody), the file will be read using mmap().

Causes the handler thread to be killed n seconds from now.

Returns an IO action which you can use to reset the handling thread's timeout value.

While Iteratees consume data, enumerators generate it. Since Iteratee is an alias for m (Step a m b), Enumerators can be considered step transformers of type Step a m b -> m (Step a m b).

Enumerators typically read from an external source (parser, handle, random generator, etc). They feed chunks into an Iteratee until the source runs out of data (triggering EOF) or the iteratee finishes processing (Yields a value).

An existential wrapper for the 'Enumerator ByteString IO a' type

Converts a CTime into an HTTP timestamp.

Converts an HTTP timestamp into a CTime.

Parses a string encoded in application/x-www-form-urlencoded format.

URL-escapes a string (see http://tools.ietf.org/html/rfc2396.html#section-2.4)

URL-escapes a string (see http://tools.ietf.org/html/rfc2396.html#section-2.4) into a Builder.

Decodes an URL-escaped string (see http://tools.ietf.org/html/rfc2396.html#section-2.4)