On Son 13.08.2006 16:26, Ralph Mitchell wrote:
>
>Not really. To do it right would require building a large part of a
>browser. For example:
>
>3) I had some success pushing javascript through the Spidermonkey
>javascript interpreter, to process a proxy autoconfig file. Even then,
>I found some of the "builtin" functions were not actually built into
>the interpreter. I had to dig them out of the Mozilla source, where
>they were stored as javascript. Are you prepared to dig out all the
>other "builtins" that aren't??

I have also try to use *only* the js-engine but then i have seen that
you also need the DOM part of mozilla and ....

My conclusio was that, if somebody want to make a cli-program which work
a little bit as a browser then you should use a browser engine, for
example gecko, or go the way which http://bfilter.sourceforge.net/ goes,
build your own engine.

Which way you decide i think it's a lot of work for the person(s)? which
start to develop such a solution on top of libcurl ;-)

A nice soulution would be libevent => libcurl -> *multi* => gecko but
*this* is a huge project, imho ;-)

Regards

Alex
Received on 2006-08-14