Sunil,
On Wed, Jun 05, 2002 at 11:19:20AM +0530, Sunil Shetye wrote:
Quoting from Eric S. Raymond's mail on Tue, Jun 04, 2002 at 04:42:30PM -0400:
* Cygwin port patch from Jason Tishler.
The above patch is not used if fetchmail has been compiled with ssl
and is polling from a host without ssl.
Note that I have already posted "take 2" of the above patch to handle
the --with-ssl case too:
http://cygwin.com/ml/cygwin/2002-05/msg01694.html
This patch should fix this.
Nevertheless, thanks for your patch. Yours is "better" because it removes
the code duplicate between the SSL and non-SSL cases. My second version
is "better" because it minimizes perturbing the existing code base.
Maybe a combination of the two would be best?
o elimination of code duplication (from yours)
o use of cygwin_read (via #def fm_read) to hide the Cygwin Winsock
workaround (from mine)
Eric,
Do you have any opinions on what is the best approach? Also, would
you consider one of these patches for inclusion in 6.0.0 (or 5.9.13)?
Or, should I resubmit after 6.0.0?
Thanks,
Jason