Subject: From: Date:

RE: [ECS] Web server "events" needed ? Rob Hicks Thu, 10 Feb 2000 08:43:34 -0800

---

Hi Mark,

I'm sorry to have written my last comment so lamely. Let me take another
stab.

As I understand what ECS is doing now, it is dynamically creating web pages
based upon the various monitoring groups that have been defined and serving
them up using an embedded web server. (Note that I haven't got it to work
yet, because I have not yet disabled or tried to reconfigure my existing web
server). If my assumption is correct, then, if the ECS embedded web server
could be disabled-without disabling the dynamic creation of the group web
pages, any web server could respond to client requests for the ECS generated
group pages. It would simply be a matter of configuring the existing web
server to know where to find them. Pointing either Microsoft's Personal Web
Server or IIS to the right director is fairly easy. Does that make better
sense now?

Regards,
Rob

-----Original Message-----
From: Mark Gilmore [mailto:omnip@usit.net]
Sent: Thursday, February 10, 2000 7:19 AM
To: ecs-list@netbloc.com
Subject: Re: [ECS] Web server "events" needed ?

Hi Rob,
I understand the port-conflict problem, but don't understand
your question: What is a "regular web server process request" ?

NOTE TO ALL: A recurring problem in all this web/javascript
stuff is that many assume that I know MUCH more than I do :-).
I happen to have a knack for tweaking working examples to my
liking. But this can give the false impression that I actually
know what I'm doing :-). So please consider this in subsequent
communications related to web-severs/JAVAScript/JAVA/CGI
(i.e. be very specific).

Tanx!
--
Mark Gilmore
Omnipotence (ECS home automation software)
http://www.usit.com/omnip
423-745-0026


Rob Hicks wrote:
>
> Hi Mark.
>
> It looks like you're making good progress.
>
> I can't speak to the read/write events, but I believe can shed some light
on
> the close event, if I understand your question correctly. Older web
browsers
> (and servers for that matter) tried to maintain an open session unless the
> connection was closed. If not, they maintained a "keep alive" for that
> connection and when they tried to maintain too many "keep alives" would
run
> out of memory and die. Newer servers and browsers automatically time out
> each keep alive.
>
> Also, I have a question. I assume that ECS is using port 80 for HTTP.
> However, if a web server already exists on the ECS server machine, ECS
won't
> be able to process any client requests because both ECS and the
preexisting
> web server will be trying to listen on the same port. Have you thought
about
> allowing just a regular web server process requests for ECS?
>
> Regards,
> Rob
> 801-319-1766
> robhicks@utah-inter.net
>
> -----Original Message-----
> From: Mark Gilmore [mailto:omnip@usit.net]
> Sent: Thursday, February 10, 2000 5:52 AM
> To: ecs-list@netbloc.com; Chris Armbruster; Dario
> Subject: [ECS] Web server "events" needed ?
>
> Hi all,
> In servicing web-based Client requests, the example I referenced
> defined "Events" for ACCEPT/READ/WRITE/CLOSE. I am still confused
> as to the diff btw READ/WRITE events (all I know is that when i get
> a WRITE Event, that I sometimes need to wait a second or so for I
> get the request packet from the Client). And I have YET to see a
> CLOSE Event.
>
> QUESTION:
> Do I really need these "events" at all ?
> I.E. - couldn't I just call accept() (to check for new connections)
> and recv() (to check for Client requests), and react appropriately
> (withOUT the use of these "events") ??
>
> tanx
> --
> Mark Gilmore
> Omnipotence (ECS home automation software)
> http://www.usit.com/omnip
> 423-745-0026

---