process group killing

This seems to work pretty well.

Handled the process groups like this:

- git-annex processes started by the assistant for transfers are run in their
  own process groups.
- otherwise, rely on the shell to allocate a process group for git-annex

There is potentially a problem if some other program runs git-annex
directly (not using sh -c) The program and git-annex would then be in
the same process group. If that git-annex starts a transfer and it's
canceled, the program would also get killed. May or may not be a desired
result.

Also, the new updateTransferInfo probably closes a race where it was
possible for the thread id to not be recorded in the transfer info, if
the transfer info file from the transfer process is read first.

1 files changed, 8 insertions, 3 deletions

diff --git a/Assistant/WebApp/DashBoard.hs b/Assistant/WebApp/DashBoard.hs
index a04861ed6..e2245bf6c 100644
--- a/Assistant/WebApp/DashBoard.hs
+++ b/Assistant/WebApp/DashBoard.hs
@@ -29,7 +29,8 @@ import Yesod
 import Text.Hamlet
 import qualified Data.Map as M
 import Control.Concurrent
-import System.Posix.Signals (signalProcess, sigTERM, sigKILL)
+import System.Posix.Signals (signalProcessGroup, sigTERM, sigKILL)
+import System.Posix.Process (getProcessGroupIDOf)
 
 {- A display of currently running and queued transfers.
  -
@@ -180,7 +181,11 @@ cancelTransfer t = do
 			maybe noop killThread $ transferTid info
 			maybe noop killproc $ transferPid info
 			removeTransfer (daemonStatus webapp) t
+		{- In order to stop helper processes like rsync,
+		 - kill the whole process group of the process running the 
+		 - transfer. -}
 		killproc pid = do
-			void $ tryIO $ signalProcess sigTERM pid
+			g <- getProcessGroupIDOf pid
+			void $ tryIO $ signalProcessGroup sigTERM g
 			threadDelay 100000 -- 0.1 second grace period
-			void $ tryIO $ signalProcess sigKILL pid
+			void $ tryIO $ signalProcessGroup sigKILL g