Re: serving very large files from allegroserve

Nick Levine wrote:

> Am I right in thinking that ACL8 <=> using IO
timeouts?
> 

acl8 should be using io timeouts.  look for :io-timeout on
the 
*features* list to be sure.



> Isn't it just. Suggestions for verifying my claims and
then debugging
> it?

You can't trace mp:with-timeout since it's a macro but you
can trace
a function used by the expansion of with-timeout:


(trace excl::positive-wait)

The value passed to positive-wait should not be 300 but some
very
large value (i.e. 100 days worth of seconds).


_______________________________________________
lispweb mailing list
lispwebred-bean.com
http
://www.red-bean.com/mailman/listinfo/lispweb

Some considerable time later... I went looking in the
appropriate
locator-exact structure's info slot: a hash-table which as
I
understand it contains a file-entity for each file implicity
published
when serving something under a publish-directory.

The file I'd been trying to download had timeout slot nil,
another
file in the same directory had timeout slot 8640000.

So all I have to do is remhash the offending entry and
hopefully when
I go home this evening I can listen to the remains of this
audio file.

So, what I really want to know, now that I've wasted over
two hours
digging into this (and everyone on this list has had to
listen to me
doing it), is: where did I go wrong? All I ever did was
publish a
directory (with no :timeout arguments) and I got my fingers
stepped
on. What's the route whereby the server can create a
file-entity
without its inherited timeout?

That's all from me.

- n



_______________________________________________
lispweb mailing list
lispwebred-bean.com
http
://www.red-bean.com/mailman/listinfo/lispweb