This is the mail archive of the
cygwin-patches
mailing list for the Cygwin project.
[PATCH 05/14] Cygwin: FIFO: fix the error checking in raw_read
- From: Ken Brown <kbrown at cornell dot edu>
- To: "cygwin-patches at cygwin dot com" <cygwin-patches at cygwin dot com>
- Date: Sun, 14 Apr 2019 19:15:58 +0000
- Subject: [PATCH 05/14] Cygwin: FIFO: fix the error checking in raw_read
- References: <20190414191543.3218-1-kbrown@cornell.edu>
If the pipe is empty, we can get either ERROR_NO_DATA or
ERROR_PIPE_LISTENING.
---
winsup/cygwin/fhandler_fifo.cc | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/winsup/cygwin/fhandler_fifo.cc b/winsup/cygwin/fhandler_fifo.cc
index 764420ffd..2da579b95 100644
--- a/winsup/cygwin/fhandler_fifo.cc
+++ b/winsup/cygwin/fhandler_fifo.cc
@@ -749,19 +749,16 @@ fhandler_fifo::raw_read (void *in_ptr, size_t& len)
fifo_client_unlock ();
return;
}
- /* In the duplex case with no data, we seem to get nread
- == -1 with ERROR_PIPE_LISTENING on the first attempt to
- read from the duplex pipe (fc_handler[0]), and nread == 0
- on subsequent attempts. */
+ /* If the pipe is empty, we usually get nread == -1 with
+ ERROR_NO_DATA or ERROR_PIPE_LISTENING. An exception is
+ that in the duplex case we may get nread == 0 when we
+ attempt to read from the duplex pipe (fc_handler[0]). */
else if (nread < 0)
switch (GetLastError ())
{
case ERROR_NO_DATA:
- break;
case ERROR_PIPE_LISTENING:
- if (_duplexer && i == 0)
- break;
- /* Fall through. */
+ break;
default:
fifo_client_unlock ();
goto errout;
--
2.17.0