Re: [Tigervnc-devel] race condition in readVersion?

2012-03-19 Thread Pierre Ossman
On Sun, 18 Mar 2012 18:18:43 -0400 Brian Hinz bph...@users.sourceforge.net wrote: However since this is the very first read that takes place in the handshake, there should not be any data in the input buffer, and because it's a checkNoWait a blocking read should cause the while loop to

[Tigervnc-devel] race condition in readVersion?

2012-03-18 Thread Brian Hinz
Hi, From time to time during testing I've seen an issue with the new Java client where it fails to read the version message from the server. Previously, I had only seen it when connecting to rfbproxy, never when connecting to a real server, so I had assumed that the problem was with rfbproxy.