🚀 go-pugleaf

In article <900660267snz@isis.demon.co.uk>
           ian@isis.demon.co.uk "Ian Smith" wrote:

>On Wednesday, in article
>     <Snews.980715.185607.chris.10745@keris.demon.co.uk>
>     chris@keris.demon.co.uk "Chris Croughton" wrote:
>
>It doesn't look like source - try looking for pcelm21.zip, about 117k.
>It's for TC 1.5 and needed a bit of a massage to compile with BC 3.1 
>Last year I started to update it, but it hasn't got very far at all.

Thanks, I'll have a look.

>> It's a darn sight better than View, it doesn't bomb out at 300 messages in 
>> a mailbox.  What I'd prefer is real elm, but I still haven't managed to get
>> the configure script to run all through even under Win95 (it complains of
>> running out of virtual memory, which is ridiculous, it doesn't need 128Mb
>> to compile even on a DEC Alpha).
>
>Blimey !  I archive my mailboxes at about 20 messages.

Since I get at least one mailing list which comes in at over a hundred per
day, archiving at anything under 300 is not really an option (although I do
now have an AWK script which archives only unread messages, this helps).

>> BTW, I have modifications to POP3CLI and MAILCLI so that if you have lots
>> of POP3 mailboxes it doesn't try to kick them all at once.
>
>As I recall it, the POP3 client I based YAN's version on only ever 
>kicked the first one it found for the specified servername. YAN's 
>current version can kick (and drop) all "servers" for a servername or 
>just one selected by username.
>Which source did you base your client code on ?

Yours, YAN 1.9 (I haven't looked at 2.0 yet to see if there's any
change in that area).  The problem is that if you need to kick the 
mailboxes (and I have over 20) on a 'per user'basis, when do you do it?
If you kick them all at once the system runs out of memory, if you kick
them on a timer you can still run out of memory (if they take longer to
d/l than the timer) or can sit around wasting time (==money) if they
d/l faster.

My version kicks the first <n> servers (at the moment 4, this should be
configurable), then when one finishes it kicks the next unkicked one
and so forth (actually there's a timer in there to do the kicking as
well).  I've also upgraded the 'popclient list' output to display for
each mailbox whether it is running or kicked.

An alternate method would be to make each client (possibly as few as one)
look at the next thing to do when it has finished and if there are more
mailboxes do the next one.  This would save on process creation
and termination.

Another thing I've done is make it stop at <n> messages retrieved and
save the messages (and do QUIT to delete them from the server), because
if the line drops after 300 out of 500 messages it's really annoying to
lose the messages - in fact at one time I found it impossible to get
any of the messages because my line was prone to dropping (not one of
Demon's, I hasten to add, this was retrieving mail while in Germany).
At the moment <n> is again fixed (at 20), this too needs to be configurable.

This sort of thing is why I'm using fewer and fewer things where I don't have
the source.  With the source I can make it (for it in almost*) do what
I need; if someone else finds it useful fine, if not well I'm happy <g>.

(Also, I'm a programmer for fun as well as for job.  I can't /not/ want to
tinker with things <g>...)

Again, thanks muchly to you and everyone else for making it possible for
me to do this...

Chris C