Possible race condition notes

[Christopher Faylor]

On Sun, Apr 08, 2001 at 09:08:31PM -0400, Harold Hunt wrote:
>> I'm not sure who else is going to be doing debugging but I
>> just want to
>> make it clear that I'm not going to be debugging the X
>> server.  This is
>> the first I've heard of any race conditions in select (which
>> was written
>> very carefully to be race-free <which means absolutely nothing of
>> course>).
>
>I sincerely doubt that the problem is in Cygwin's select.  I suspect that
>the race is due more the hack job on the timeout... I mean, it doesn't seem
>correct that XFree86 on Cygwin uses a static timeout of 10 milliseconds,
>when all other platforms use the same calculation to determine how long
>select should wait.  XFree86 has been around for awhile and I'm sure we'd
>see other hack jobs on the timeout value if it was the correct thing to do.
>
>I believe that the file descriptors being passed to select are network
>sockets only.  User input/all windows messages, are handled in a seperate
>function that does not use select, nor are the windows messages ever added
>to select's list of file descriptors.

Hmm.  Are these unix domain sockets or actual network sockets?  Do you see
different behavior between setting DISPLAY to :0 versus localhost:0?

cgf