I have a simple patch that allows Admins to control if the default_handler
should make ETags with the inode or not.
The patch allows client-side image-caching (aka server gives 304s) to take
place if the web farm has > 1 unit.
As you know, since rsync'ed files never share the same inode, ETags are
always different for each round robin requests to a web farm.
(and since the ETag is different, 304s are not sent out).
I've been waiting for 1.3.17 to be released on ftp.apache.org before I
release it.
Please consider this change before releasing 1.3.19.
Patch to come this weekend.
> -----Original Message-----
> From: Jim Jagielski [SMTP:[EMAIL PROTECTED]]
> Sent: Friday, February 23, 2001 10:49 AM
> To: [EMAIL PROTECTED]
> Subject: Regarding 1.3.19...
>
> We need to consider what to do for 1.3.19. There is one open issue,
> and that's the warnings about the use of tmpnam() in 2 of the
> support programs. We could change these to use mkstemp() if available
> (we do use it already in the main code, but only under systems
> with mmap using MAP_TMPFILE. So to do it right, we should check
> to see if mkstemp() is available during Configure.
>
> Comments?
> --
> ==========================================================================
> =
> Jim Jagielski [|] [EMAIL PROTECTED] [|] http://www.jaguNET.com/
> "Casanova will have many weapons; To beat him you will
> have to have more than forks and flatulence."
