Re: Introduction

From: Serassio Guido <>
Date: Sat, 30 Apr 2005 17:12:20 +0200

Hi Henrik,

At 16.41 30/04/2005, Henrik Nordstrom wrote:

>On Sat, 30 Apr 2005, Serassio Guido wrote:
>>I don't wrote that your solution is wrong, or that C++ classes are bad. I
>>wrote that your solution is slower.
>I doubt it actually is.

The Evgeny's specific implementation is it. Its read() checks for the fd
type at every I/O operation, when this could be done only one time when
creating the fd.

>And in addition one of the general design goals of Squid-3 is to get
>things object oriented. Macros is lagely the opposite.
>On my todo there is an item to get rid of the fd_table entirely, turning
>"file descriptor" into a class allowing native I/O implementations.

I agree totally: on the current Windows implementation this will solve some

>Even on UNIX there is subtle differences beetween the different classes of
>I/O handles, even more so as time passes and new interesting I/O or event
>mechanisms gets available.

Not only on UNIX, this should allow a simpler implementation of native
socket I/O on Windows.



Guido Serassio
Acme Consulting S.r.l. - Microsoft Certified Partner
Via Lucia Savarino, 1 10098 - Rivoli (TO) - ITALY
Tel. : +39.011.9530135 Fax. : +39.011.9781115
Received on Sat Apr 30 2005 - 09:12:55 MDT

This archive was generated by hypermail pre-2.1.9 : Sun May 01 2005 - 12:00:06 MDT