cygwin/x symantec antivirus conflict (fixed in snapshot?)

Christopher Faylor
Mon Oct 11 04:06:00 GMT 2004

On Sun, Oct 10, 2004 at 11:33:00PM -0400, Philip Gladstone wrote:
>Christopher Faylor wrote:
>>Ok.  I've changed the algorithm in select.  It only opens a DGRAM
>>socket now, one time per thread.  It uses this to terminate the socket
>>thread, if necessary.  This socket is never closed until the thread
>>It sounds like this would more or less fix the problem that you're
>>seeing.  Would you mind trying a new snapshot?
>This fixes it for me! Thanks immensely.

That's good news!  Thanks for the fast feedback.

I've cc'ed the cygwin mailing list since this has been a long standing

For those who haven't been following along at home, it looks like a
change I just made to select() may solve the dreaded "slows down to a
crawl with Symantec AntiVirus" problem.

This may also improve the performance of things that use sockets

So, I'd appreciate reports on the latest snapshot.  Does it fix any
problems?  Cause any problems?  No change?

In this one case, I'd like to hear "me toos" since the change was to a
fundamental part of cygwin and it is in socket code, which has proved to
be problematic.  So, I'd like to know if things still work on Windows 9x
and all flavors of NT.

"Things" would be, openssh, telnet, ftp, rsync, etc.  Anything which
uses sockets or communicates via TCP/IP.

I've reset the reply-to for this message to the cygwin mailing list, so
if there is further Cygwin/X discussion necessary, please make sure that
it goes to the cygwin-xfree mailing list.


More information about the Cygwin-xfree mailing list