[prev in list] [next in list] [prev in thread] [next in thread] 

List:       vnc-list
Subject:    Re: Mac VNC stability
From:       Charlie McLachlan <cim () uk ! research ! att ! com>
Date:       1999-08-31 7:59:18
[Download RAW message or body]

On Tue, 31 Aug 1999, Jeffrey M. Vinocur wrote:

> instability.  Any recommendation on what's likely the problem?  I don't use
> it enough to be able to try the different combinations and compare, but I
> thought you might have an opinion.

There are bits of the current server that use functions that access the
memory manager. Which all fine and dandy as long as these bits are not run
at inturrupt level. The strategy that the server uses to catch updates
means that it cannot be sure which execution level it is running at. Which
menas once in a while a memory manager call will be made at inturrupt
level which will cause the whole machine to crash.

The next release of the server isolates all memory manager calls and does
them all in the main event loop (which is always run at user
level) hence, the next release should be a lot more stable (fingers
crossed).

That said, it may be something totally different. I think there is a small
bug in the hextile encoder (hence the option to turn it off) that will
also crash the machine if it is given the "wrong" picture to encode. 

The server obviously adds a latency to graphics calls which might confuse
some programs into crashing (although this is pretty unlikely).

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Charlie McLachlan - Responsible for VNC on the Mac
	AT&T Research Lab. Cambridge.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=


---------------------------------------------------------------------
The VNC mailing list - see http://www.uk.research.att.com/vnc/intouch.html
---------------------------------------------------------------------

[prev in list] [next in list] [prev in thread] [next in thread]