You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

286 lines
11 KiB

{-# OPTIONS_HADDOCK hide #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE OverloadedStrings #-}
module Network.Xmpp.Connection_ where
import Control.Applicative((<$>))
import Control.Concurrent (forkIO, threadDelay)
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
import System.IO.Error (tryIOError)
import Control.Monad
import Control.Monad.IO.Class
import Control.Monad.Trans.Class
--import Control.Monad.Trans.Resource
import qualified Control.Exception.Lifted as Ex
import qualified GHC.IO.Exception as GIE
import Control.Monad.State.Strict
import Data.ByteString as BS
import Data.ByteString.Char8 as BSC8
import Data.Conduit
import Data.Conduit.Binary as CB
import Data.Conduit.Internal as DCI
import qualified Data.Conduit.List as CL
import Data.IORef
import Data.Text(Text)
import qualified Data.Text as T
import Data.XML.Pickle
import Data.XML.Types
import Network
import Network.Xmpp.Types
import Network.Xmpp.Marshal
import Network.Xmpp.Pickle
import System.IO
import Text.Xml.Stream.Elements
import Text.XML.Stream.Parse as XP
import Text.XML.Unresolved(InvalidEventStream(..))
import System.Log.Logger
import Data.ByteString.Base64
import Control.Concurrent.STM.TMVar
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
import Control.Monad.Error
-- Enable/disable debug output
-- This will dump all incoming and outgoing network taffic to the console,
-- prefixed with "in: " and "out: " respectively
debug :: Bool
debug = False
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-- TODO: Can the TLS send/recv functions throw something other than an IO error?
wrapIOException :: IO a -> StateT Connection IO (Either XmppFailure a)
wrapIOException action = do
r <- liftIO $ tryIOError action
case r of
Right b -> return $ Right b
Left e -> return $ Left $ XmppIOException e
pushElement :: Element -> StateT Connection IO (Either XmppFailure Bool)
pushElement x = do
send <- gets (cSend . cHandle)
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
wrapIOException $ send $ renderElement x
-- | Encode and send stanza
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
pushStanza :: Stanza -> TMVar Connection -> IO (Either XmppFailure Bool)
pushStanza s = withConnection' . pushElement $ pickleElem xpStanza s
-- XML documents and XMPP streams SHOULD be preceeded by an XML declaration.
-- UTF-8 is the only supported XMPP encoding. The standalone document
-- declaration (matching "SDDecl" in the XML standard) MUST NOT be included in
-- XMPP streams. RFC 6120 defines XMPP only in terms of XML 1.0.
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
pushXmlDecl :: StateT Connection IO (Either XmppFailure Bool)
pushXmlDecl = do
con <- gets cHandle
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
wrapIOException $ (cSend con) "<?xml version=\"1.0\" encoding=\"UTF-8\" ?>"
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
pushOpenElement :: Element -> StateT Connection IO (Either XmppFailure Bool)
pushOpenElement e = do
sink <- gets (cSend . cHandle)
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
wrapIOException $ sink $ renderOpenElement e
-- `Connect-and-resumes' the given sink to the connection source, and pulls a
-- `b' value.
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
runEventsSink :: Sink Event IO b -> StateT Connection IO (Either XmppFailure b)
runEventsSink snk = do -- TODO: Wrap exceptions?
source <- gets cEventSource
(src', r) <- lift $ source $$++ snk
modify (\s -> s{cEventSource = src'})
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
return $ Right r
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
pullElement :: StateT Connection IO (Either XmppFailure Element)
pullElement = do
Ex.catches (do
e <- runEventsSink (elements =$ await)
case e of
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
Left f -> return $ Left f
Right Nothing -> return $ Left XmppOtherFailure -- TODO
Right (Just r) -> return $ Right r
)
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
[ Ex.Handler (\StreamEnd -> return $ Left StreamEndFailure)
, Ex.Handler (\(InvalidXmppXml s) -- Invalid XML `Event' encountered, or missing element close tag
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-> return $ Left XmppOtherFailure) -- TODO: Log: s
, Ex.Handler $ \(e :: InvalidEventStream) -- xml-conduit exception
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-> return $ Left XmppOtherFailure -- TODO: Log: (show e)
]
-- Pulls an element and unpickles it.
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
pullUnpickle :: PU [Node] a -> StateT Connection IO (Either XmppFailure a)
pullUnpickle p = do
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
elem <- pullElement
case elem of
Left e -> return $ Left e
Right elem' -> do
let res = unpickleElem p elem'
case res of
Left e -> return $ Left XmppOtherFailure -- TODO: Log
Right r -> return $ Right r
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-- | Pulls a stanza (or stream error) from the stream.
pullStanza :: TMVar Connection -> IO (Either XmppFailure Stanza)
pullStanza = withConnection' $ do
res <- pullUnpickle xpStreamStanza
case res of
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
Left e -> return $ Left e
Right (Left e) -> return $ Left $ StreamErrorFailure e
Right (Right r) -> return $ Right r
-- Performs the given IO operation, catches any errors and re-throws everything
-- except 'ResourceVanished' and IllegalOperation, in which case it will return False instead
catchPush :: IO () -> IO Bool
catchPush p = Ex.catch
(p >> return True)
(\e -> case GIE.ioe_type e of
GIE.ResourceVanished -> return False
GIE.IllegalOperation -> return False
_ -> Ex.throwIO e
)
-- Connection state used when there is no connection.
xmppNoConnection :: Connection
xmppNoConnection = Connection
{ cHandle = ConnectionHandle { cSend = \_ -> return False
, cRecv = \_ -> Ex.throwIO
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
XmppOtherFailure
, cFlush = return ()
, cClose = return ()
}
, cEventSource = DCI.ResumableSource zeroSource (return ())
, cFeatures = SF Nothing [] []
, cState = ConnectionClosed
, cHostName = Nothing
, cJid = Nothing
, cStreamLang = Nothing
, cStreamId = Nothing
, cPreferredLang = Nothing
, cToJid = Nothing
, cJidWhenPlain = False
, cFrom = Nothing
}
where
zeroSource :: Source IO output
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
zeroSource = liftIO . Ex.throwIO $ XmppOtherFailure
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
connectTcp :: HostName -> PortID -> Text -> IO (Either XmppFailure (TMVar Connection))
connectTcp host port hostname = do
let PortNumber portNumber = port
debugM "Pontarius.Xmpp" $ "Connecting to " ++ host ++ " on port " ++
(show portNumber) ++ " through the realm " ++ (T.unpack hostname) ++ "."
h <- connectTo host port
debugM "Pontarius.Xmpp" "Setting NoBuffering mode on handle."
hSetBuffering h NoBuffering
let eSource = DCI.ResumableSource
((sourceHandle h $= logConduit) $= XP.parseBytes def)
(return ())
let hand = ConnectionHandle { cSend = \d -> do
let d64 = encode d
debugM "Pontarius.Xmpp" $
"Sending TCP data: " ++ (BSC8.unpack d64)
++ "."
catchPush $ BS.hPut h d
, cRecv = \n -> do
d <- BS.hGetSome h n
let d64 = encode d
debugM "Pontarius.Xmpp" $
"Received TCP data: " ++
(BSC8.unpack d64) ++ "."
return d
, cFlush = hFlush h
, cClose = hClose h
}
let con = Connection
{ cHandle = hand
, cEventSource = eSource
, cFeatures = (SF Nothing [] [])
, cState = ConnectionPlain
, cHostName = (Just hostname)
, cJid = Nothing
, cPreferredLang = Nothing -- TODO: Allow user to set
, cStreamLang = Nothing
, cStreamId = Nothing
, cToJid = Nothing -- TODO: Allow user to set
, cJidWhenPlain = False -- TODO: Allow user to set
, cFrom = Nothing
}
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
con' <- mkConnection con
return $ Right con'
where
logConduit :: Conduit ByteString IO ByteString
logConduit = CL.mapM $ \d -> do
let d64 = encode d
debugM "Pontarius.Xmpp" $ "Received TCP data: " ++ (BSC8.unpack d64) ++
"."
return d
-- Closes the connection and updates the XmppConMonad Connection state.
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-- killConnection :: TMVar Connection -> IO (Either Ex.SomeException ())
killConnection :: TMVar Connection -> IO (Either XmppFailure ())
killConnection = withConnection $ do
cc <- gets (cClose . cHandle)
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
err <- wrapIOException cc
-- (Ex.try cc :: IO (Either Ex.SomeException ()))
put xmppNoConnection
return err
-- Sends an IQ request and waits for the response. If the response ID does not
-- match the outgoing ID, an error is thrown.
pushIQ' :: StanzaID
-> Maybe Jid
-> IQRequestType
-> Maybe LangTag
-> Element
-> TMVar Connection
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-> IO (Either XmppFailure (Either IQError IQResult))
pushIQ' iqID to tp lang body con = do
pushStanza (IQRequestS $ IQRequest iqID Nothing to lang tp body) con
res <- pullStanza con
case res of
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
Left e -> return $ Left e
Right (IQErrorS e) -> return $ Right $ Left e
Right (IQResultS r) -> do
unless
(iqID == iqResultID r) . liftIO . Ex.throwIO $
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
XmppOtherFailure
-- TODO: Log: ("In sendIQ' IDs don't match: " ++ show iqID ++
-- " /= " ++ show (iqResultID r) ++ " .")
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
return $ Right $ Right r
_ -> return $ Left XmppOtherFailure
-- TODO: Log: "sendIQ': unexpected stanza type "
-- | Send "</stream:stream>" and wait for the server to finish processing and to
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
-- close the connection. Any remaining elements from the server are returned.
-- Surpresses StreamEndFailure exceptions, but may throw a StreamCloseError.
closeStreams :: TMVar Connection -> IO (Either XmppFailure [Element])
closeStreams = withConnection $ do
send <- gets (cSend . cHandle)
cc <- gets (cClose . cHandle)
liftIO $ send "</stream:stream>"
void $ liftIO $ forkIO $ do
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
threadDelay 3000000 -- TODO: Configurable value
(Ex.try cc) :: IO (Either Ex.SomeException ())
return ()
collectElems []
where
-- Pulls elements from the stream until the stream ends, or an error is
-- raised.
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
collectElems :: [Element] -> StateT Connection IO (Either XmppFailure [Element])
collectElems es = do
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
result <- pullElement
case result of
Tweak failure approach I'm assuming and defining the following: 1. XMPP failures (which can occur at the TCP, TLS, and XML/XMPP layers (as a stream error or forbidden input)) are fatal; they will distrupt the XMPP session. 2. All fatal failures should be thrown (or similar) by `session', or any other function that might produce them. 3. Authentication failures that are not "XMPP failures" are not fatal. They do not necessarily terminate the stream. For example, the developer should be able to make another authentication attempt. The `Session' object returned by `session' might be useful even if the authentication fails. 4. We can (and should) use one single data type for fatal failures. (Previously, both StreamFailure and TlsFailure was used.) 5. We can catch and rethrow/wrap IO exceptions in the context of the Pontarius XMPP error system that we decide to use, making the error system more intuitive, Haskell-like, and more straight-forward to implement. Calling `error' may only be done in the case of a program error (a bug). 6. A logging system will remove the need for many of the error types. Only exceptions that seem likely to affect the flow of client applications should be defined. 7. The authentication functions are prone to fatal XMPP failures in addition to non-fatal authentication conditions. (Previously, `AuthStreamFailure' was used to wrap these errors.) I'm hereby suggesting (and implementing) the following: `StreamFailure' and `TlsFailure' should be joined into `XmppFailure'. `pullStanza' and the other Connection functions used to throw `IOException', `StreamFailure' and `TlsFailure' exceptions. With this patch, they have been converted to `StateT Connection IO (Either XmppFailure a)' computations. They also catch (some) IOException errors and wrap them in the new `XmppIOException' constructor. `newSession' is now `IO (Either XmppFailure Session)' as well (being capable of throwing IO exceptions). Whether or not to continue to a) wrap `XmppFailure' failures in an `AuthStreamFailure' equivalent, or, b) treat the authentication functions just like the other functions that may result in failure (Either XmppFailure a), depends on how Network.Xmpp.Connection.auth will be used. Since the latter will make `auth' more consistent, as well as remove the need for a wrapped (and special-case) "AuthFailure" type, I have decided to give the "b" approach a try. (The drawback being, of course, that authentication errors can not be accessed through the use of ErrorT. Whether or not this might be a problem, I don't really know at this point.) As the SASL code (and SaslM) depended on `AuthStreamFailure', it remains for internal use, at least for the time-being. `session' is now an ErrorT computation as well. Some functions have been updated as hacks, but this will be changed if we decide to move forward with this approach.
13 years ago
Left StreamEndFailure -> return $ Right es
Left e -> return $ Left $ StreamCloseError (es, e)
Right e -> collectElems (e:es)
debugConduit :: Pipe l ByteString ByteString u IO b
debugConduit = forever $ do
s' <- await
case s' of
Just s -> do
liftIO $ BS.putStrLn (BS.append "in: " s)
yield s
Nothing -> return ()