Re: Debug of Test 15

[Christopher Faylor <cgf at redhat dot com>]

On Sat, Apr 07, 2001 at 04:28:55PM -0400, Harold Hunt wrote:
>However, select is a "blocking function"; that is, when select is called it
>blocks, taking no clock cycles, until one of the events being waited for has
>occurred.  So, select, if it was working correctly, would take no clock
>cycles when it doesn't return.
>
>This leads me to believe that select may not be working correctly.
>
>Chris, any ideas?  I believe Pierre is using that latest cygwin1.dll
>snapshot.  Could you confirm that to the list Pierre?

It is pretty hard to believe that select isn't working correctly but I
guess anything is possible.

Some (most) select operations in cygwin actually need to resort to
polling so they do consume CPU.  The easiest way to find out what is
happening is to attach a gdb to the running process.

Also, I have to point out that I don't read every message in this
mailing list.  If you want to be sure to get my attention, change the
subject line to something like "cgf input needed".

cgf