On Tue, 14 Apr 1998, I wrote:
> I have made a note to try to improve this. Given that circumstances such
> as yours exist, I will have to think of a way of making the fudge factor
> dependent on the size of the message, say 10% of the size, or something.
I have found a way of greatly improving things for the next release. It
turned out to be quite easy to keep a count of the number of lines in
the message's body and original headers. In simple cases this means that
a pretty accurate SIZE estimate can be given. To allow for
configurations where headers get added, I've invented an option for the
smtp transport which contains a value to add to the original message
size. I have set the default value at 1024, which will allow modest
additions without the administrator having to learn about this.
I suppose, if people really want it, I could also provide a way of
turning off the use of SIZE altogether for those that don't believe in
it. Is there any demand for this?
This exercise also turned up a buglet in the size values given in the S=
entries in log lines, which are currently a bit bigger than they should
be (it's including the entire -H file length, instead of just the
headers).
I'm probably going to make a 1.91 release with various bug fixes and
tidies in a couple of weeks, since there have been a number of minor
things that have cropped up in 1.90. (I am away again next week.)
--
Philip Hazel University Computing Service,
ph10@??? New Museums Site, Cambridge CB2 3QG,
P.Hazel@??? England. Phone: +44 1223 334714
--
*** Exim information can be found at
http://www.exim.org/ ***
