From 582442e6d0ffcea16667fa896a6e9a60debac990 Mon Sep 17 00:00:00 2001 From: Sam T Date: Sun, 7 Jul 2013 17:10:38 +0400 Subject: ~ Turn some comments into literate. --- src/Network/BitTorrent/Internal.lhs | 77 +++++++++++++++++-------------------- 1 file changed, 36 insertions(+), 41 deletions(-) (limited to 'src') diff --git a/src/Network/BitTorrent/Internal.lhs b/src/Network/BitTorrent/Internal.lhs index ed4fd1a0..f46ca244 100644 --- a/src/Network/BitTorrent/Internal.lhs +++ b/src/Network/BitTorrent/Internal.lhs @@ -9,12 +9,9 @@ > -- provides sessions needed by Network.BitTorrent and > -- Network.BitTorrent.Exchange and modules. To hide some internals > -- of this module we detach it from Exchange. +> -- NOTE: Expose only static data in data field lists, all dynamic +> -- data should be modified through standalone functions. > -- -> - -**NOTE**: Expose only static data in data field lists, all dynamic -data should be modified through standalone functions. - > > {-# LANGUAGE OverloadedStrings #-} > {-# LANGUAGE RecordWildCards #-} @@ -513,27 +510,26 @@ Peer session > -- | Extensions such that both peer and client support. > , enabledExtensions :: [Extension] -> -- | To dissconnect from died peers appropriately we should check -> -- if a peer do not sent the KA message within given interval. If -> -- yes, we should throw an exception in 'TimeoutCallback' and -> -- close session between peers. -> -- -> -- We should update timeout if we /receive/ any message within -> -- timeout interval to keep connection up. +To dissconnect from died peers appropriately we should check if a peer +do not sent the KA message within given interval. If yes, we should +throw an exception in 'TimeoutCallback' and close session between +peers. + +We should update timeout if we /receive/ any message within timeout +interval to keep connection up. + > , incomingTimeout :: !TimeoutKey -> -- | To send KA message appropriately we should know when was last -> -- time we sent a message to a peer. To do that we keep registered -> -- timeout in event manager and if we do not sent any message to -> -- the peer within given interval then we send KA message in -> -- 'TimeoutCallback'. -> -- -> -- We should update timeout if we /send/ any message within timeout -> -- to avoid reduntant KA messages. -> -- -> , outcomingTimeout :: !TimeoutKey +To send KA message appropriately we should know when was last time we +sent a message to a peer. To do that we keep registered timeout in +event manager and if we do not sent any message to the peer within +given interval then we send KA message in 'TimeoutCallback'. -> -- TODO use dupChan for broadcasting +We should update timeout if we /send/ any message within timeout to +avoid reduntant KA messages. + +> , outcomingTimeout :: !TimeoutKey +> > -- | Broadcast messages waiting to be sent to peer. > , pendingMessages :: !(TChan Message) @@ -623,19 +619,18 @@ Peer session Broadcasting: Have, Cancel, Bitfield, SuggestPiece ------------------------------------------------------------------------ -> -- here we should enqueue broadcast messages and keep in mind that: -> -- -> -- * We should enqueue broadcast events as they are appear. -> -- * We should yield broadcast messages as fast as we get them. -> -- -> -- these 2 phases might differ in time significantly +Here we should enqueue broadcast messages and keep in mind that: + * We should enqueue broadcast events as they are appear. + * We should yield broadcast messages as fast as we get them. -> -- TODO do this; but only when it'll be clean which other broadcast -> -- messages & events we should send +these 2 phases might differ in time significantly -> -- 1. Update client have bitfield --\____ in one transaction; -> -- 2. Update downloaded stats --/ -> -- 3. Signal to the all other peer about this. +**TODO**: do this; but only when it'll be clean which other broadcast +messages & events we should send. + +1. Update client have bitfield --\____ in one transaction; +2. Update downloaded stats --/ +3. Signal to the all other peer about this. > available :: Bitfield -> SwarmSession -> IO () > available bf se @ SwarmSession {..} = {-# SCC available #-} do @@ -646,16 +641,16 @@ Broadcasting: Have, Cancel, Bitfield, SuggestPiece > atomically $ do > modifyTVar' clientBitfield (BF.union bf) > modifyTVar' (currentProgress clientSession) (downloadedProgress bytes) - +> > broadcast = mapM_ (writeTChan broadcastMessages . Have) (BF.toList bf) -> -- TODO compute size of messages: if it's faster to send Bitfield -> -- instead many Have do that -> -- -> -- also if there is single Have message in queue then the -> -- corresponding piece is likely still in memory or disc cache, -> -- when we can send SuggestPiece +TODO compute size of messages: if it's faster to send Bitfield +instead many Have do that + +also if there is single Have message in queue then the +corresponding piece is likely still in memory or disc cache, +when we can send SuggestPiece > -- | Get pending messages queue appeared in result of asynchronously > -- changed client state. Resulting queue should be sent to a peer -- cgit v1.2.3