World: r3wp

regarding proxy settings: please post that on the 1.3.3 checklist 
if you want me to take a look at that. if you could provide a sample 
configuration script for testing, that would be great

do-browser in Mozilla: haven't looked at it yet. it would require 
access to both the Javascript interpreter as well as an "evaluate" 
(or execScript) method that would parse and interpret the script.

if anyone knows of such a feature in the Mozilla API, please let 
me know

networking through the browser: you mean initiating HTTP requests 
for everything? that would slow performance waaay down.

As far as web stuff goes, my customers want the "real life" apps. 
 Security needs range from completely disabled to something like 
java's, but being able to save files locally (save-as requestor).

Henrik, do you mean x # of events per second?

That's kind of tricky to do. It might be worth thinking about for 
the 3.0, or better yet: how about speeding VID up?

From an architecture perspective, an application should be able to 
handle as many events as the OS throws at it. Especially for graphics-intensive 
applications: those extra mouse events make a huge difference.

Pekr, regarding proxy settings: is there a problem with proxy settings 
via the registry?

Pekr: actually, I'm sure there is, that's why you are asking. Can 
you explain the problem in a little more detail? You need that auto-config 
script interpreted? Maybe it would be easier to configure REBOL to 
rad the auto-configure script than to try to get IE to parse it and 
get the details via an API. Thoughts?

Question: Did someone say there was a wrapper for Mozilla plugins 
to run on IE? If so, can you post the URL please?

josh, it's probably not appropriate to talk about for the plugin, 
so I commented on it in the REBOL3.0 blog

(the events thing)

np

(I found a wrapper for ActiveX controls to run on Mozilla, but not 
the other way around)

doesn't IE already do netscape-style plugins?  I've never had any 
troubles running any plugin with IE.  It''s always been the other 
way around, where IE was the only browser that did ActiveX, well, 
until Neptune.   I guess I should try to make demo page with latest 
Moz plugin release before I pipe up.

I think Microsoft had patent problems with Netscape-style plugins 
(I may be thinking of something else).

Brian - yeah, I do remember something like patents involved, but 
mainly imo it was MS move with IE 5.5, to move out from NS kind of 
plug-ins - the rest of the world is still using NS plugin architecture, 
that is why Mozilla plug-in works with Netscape, Opera and possibly 
others.

ScotT - interesting - I posted some link and said it earlier, that 
RT could concentrate upon NS plug-in only .... as there is (somewhere 
:-) wrapper ActiveX, which can wrap to NS plug-ins ....

I would have to scan some links ......

JoshM - yes, I need the script to be interpreted by browser imo .... 
I will post you a source-code privately, as there are live IPs and 
I am lazy to change :-)

JoshM: please go to plugin-1 group, scroll up a bit to find my bold 
messages, and read on a bit, there are some findings. To post some 
notes - probably the most significant link is - 

http://www.mozilla.org/projects/plugins/


The plugin extension adobe, mozilla, apple, opera, macromedia and 
sun agreed upon is npruntime

How to host NS plug-ins in IE:

http://www.mozilla.org/projects/plugins/plugin-host-control.html

http://www.mozilla.org/projects/plugins/npruntime.html

does not npruntime extension allow for both the ability to call into 
plug-in (being scriptable from the browser pov - probably java-script) 
and to access browser objects? Would be probably nice if rebol plug-in 
would support such new calls too ....

couldn't ECMAScript's 'eval be used for kind of 'do-browser? http://www.devguru.com/technologies/ecmascript/quickref/javascript_index.html

.... I sent privately my findings regards get-net-info being outdated 
and how control panel/internet settings/proxy influences registry 
keys ....

privately to JoshM ..... :-)

Networking: as far as the browser can do the same, i would use browsers 
io. Or is it that much slower than rebols handler? Advantages are: 
proxy, cache. Maybe there are some extensions people like, about 
adds or blacklist or whatever? if such things exists they would be 
shared too. Its security , and a computer which knows my "no"s looks 
smarter.

that is questionable, Volker. The thing is - if you "limit" (or extend 
- https) Rebol's functionality to that of browser - do we talk rebol 
hjere anymore?

proxy can be imo solved by importing settings. We need to find out, 
how to query browser for its settings etc.

... still we need imo ability to use networking/security model of 
browser, for those who need it (networking via browser). The question 
is with security - should plug-in allow default rebol security? That 
is the question ....

Proxy settings: I think the simplest thing to do here is to modify 
whatever REBOL code is detecting the proxy settings to import the 
automatic proxy configuration script. REBOL is far more suited to 
parsing text files and interpreting information than C code. As a 
side note, we're not importing any settings via the plugin itself, 
we're just letting REBOL do whatever it does normally.

Browser networking: I could be completely out of my mind, but I'm 
pretty sure that the browser sets up and tears down connections with 
each HTTP request. In other words, there's no "open connection" that 
we can just tap into -- the browser is initiating new requests for 
each page. To move to an architecture like that for plugin networking 
calls doesn't make much sense to me.......maybe I'm mis-understanding....

Pekr: if Mozilla implements eval, then yes.

Mozilla: I'm looking at npruntime now. Looks interesting.

JoshM: eval is ECMA script (java script) facility, so mozilla should 
have it. In fact, the idea came from the old plug-in group, where 
Micha posted some code which was supposed to work, should I find 
it?

Mozilla: I think the problem with hosting the Mozilla control in 
IE with a wrapper is two-fold:

 (1) one more wrapper layer slows things down (a problem for high-performance 
 apps, e.g. graphics-intensive)

 (2) according to that page, there's no scripting support. that's 
 a major bummer because we need at least do-browser, not to mention 
 the possibility of DOM access.

Pekr: sure. thanks.

Mozilla scripting: As a side note, right now we're implementing scripting 
via xpcon. I didn't know about this new npruntime model, so we may 
switch to that for the next version of the plugin.

ok, I thought that the wrapper allows that, but there is aparently 
some limitation .... well, at least, having ns plug-in seems to work 
for the rest of the world, getting npruntime support on-board would 
be nice addition too ...

Mozilla: This opens up a good discussion. What I really like about 
npruntime is the native access to the DOM. On IE, COM gives us that 
too. The problem/question is -- maybe you guys have ideas -- how 
to expose those C++ reference counted objects to REBOL clients?

to REBOL code, I should say

Only Carl might know the answer to that question, but at least you 
can see the problem :)

Josh, thanks for explanation about the delayed events. What a pity 
you cannot improve this also in Rebol2.x version as this problem 
sheds bad light upon Rebol/View apps (which uses drag'n'drop and 
simmilar features).

Yes....we're looking into that. I'm sure there's a way.

There's always a way :)

Henrik: event throtling: maybe you could use the event time filtering/compression 
made by Gabriele or Romano for that purpose? Or write your own.

Josh: so you basically need to replace WndProc handler with a plugin 
specific one right?

Well....the problem is that the plugin has to handle its own window 
events because of the COM architecture. And REBOL has to handle its 
own events for other reasons. Soo.....maybe instead of an invisible 
proxy window, we can make REBOL's window a real window that sits 
on top of the plugin window, then it would get its events directly.

That's what I'm thinking, anyway. This is still all prototype-stage.