squeakland plugin and os-x safari

tblanchard at mac.com tblanchard at mac.com
Fri Jan 24 02:12:22 UTC 2003


Sounds like a lot of work.  Is it necessary to go to this length?
CFM is a legacy technology, likely to be abandoned at the earliest  
opportunity.
I guess this helps with IE, but it seems to me that pre OS X support  
isn't worth expending a lot of effort.
I'd think just doing a framework or bundle would be the ticket.
Or am I missing something?

On Friday, January 24, 2003, at 02:53  AM, John M McIntosh wrote:

> Well I've been busy thinking about this, debugging etc. Alas no-one  
> beat my door down wanting a copy.
>
> However what I've done is built a CFM stub to load and call a mach-o  
> plugin. That makes debugging way way easier and allow us to compile  
> with GCC etc etc to get a reasonable performance gain, which also then  
> backports to Internet Explorer for os-x.
>
> What I've determined is that the URL callback primitives don't work,  
> we make the async call, then wait forever for ok/failure which we  
> never see. Don't understand why yet. Technically we could offer the  
> plugin without that feature but that causes problems for proxy  
> installations.
>
> RIght now I'm going away to think and to revisit the URL callbacks,  
> then I need to fold this code base
> I have from 3.2.4 into a 3.5.0 release, gnuify a plugin interp.c do  
> some more testing etc.
>
> Oh and by doing this we can use pthreads to run the interpreter,  
> however by doing that I discovered some interesting drawing issues one  
> needs to solve first. Otherwise we fail back to cooperative threading,  
> but that
> has sluggish performance once you've seen the pthread model run.
>
> --
> ======================================================================= 
> ====
> John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
> Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
> ======================================================================= 
> ====
>
>



More information about the Squeak-dev mailing list