Ali Sabil wrote:
2008/10/24 Ross Burton <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>
On Fri, 2008-10-24 at 11:09 +0100, Rui Tiago Cação Matos wrote:
> In this specific library case, since the API is so simple and you
> don't know you need it until you somehow check your app's settings
> it's a no-brainer really.
You could lazy-load the library when you need to access a URL, but the
point is that your app shouldn't have to know how to handle proxies or
check settings: it just asks the library to handle them.
From what I understood, libproxy only allows your to know which proxy
to connect to for reaching a specific resource, not how to handle the
proxy, as you are supposed to handle them yourself.
This is basically true. There are three steps to initiating a
connection to a host through a proxy.
1. Load proxy configuration
2. Evaluate proxy configuration (ex. WPAD, PAC and ignore lists)
3. Establish connection through the proxy
libproxy covers #1 and #2. #3 is protocol specific. Since libproxy
can't implement every protocol out there, it is up to your protocol
layer to do step #3. This is the job of, say, libsoup. So, if your
application uses libsoup, you shouldn't need to know anything about proxies.
However, some applications have the protocol layer built-in. An example
of this might be xchat. libproxy will do #1 and #2, but xchat will need
to program #3.
Ideally, each protocol should be split into a library which would
utilize libproxy.
Nathaniel
_______________________________________________
desktop-devel-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/desktop-devel-list
