Ben,

That certainly is an approach.

One key goal for Redline is for people using Java to be able to use their 
debugging tools and step from Java code into Smalltalk, and if that Smalltalk 
code calls other Java code then step into that code.

Redline generates 100% compatible bytecode and class files (in memory) making 
this possible 

Redline also supports the generation of inline bytecode so no primitives just 
the equivalent Jvm bytecode 

- James

Sent from my Commodore 64

> On 2 Jan 2017, at 3:46 am, Ben Coman <[email protected]> wrote:
> 
>> On Sat, Dec 31, 2016 at 5:53 AM, James Ladd <[email protected]> wrote:
>> This is very helpful thank you.
>> 
>> Maybe I can rephrase things to help capture what I'm trying to do.
>> 
>> In the Smalltalk-80 blue book there is a set of Kernel-Object/Kernel-Classes 
>> defined and I'm trying to find a full running implementation of these 
>> classes in a Smalltalk environment OR recreate
> 
> 
> Just my wandering mind.... an interesting experiment might be running
> SqueakJS [1]
> on top JVM8's Nashorn javascript engine [2][3], and then gradually
> converting parts to it to native Java code as profiling dictates.
> 
> cheers -ben
> 
> 
> [1] http://freudenbergs.de/bert/publications/Freudenberg-2014-SqueakJS.pdf
> [2] https://ariya.io/2014/03/nashorn-the-new-rhino-on-the-block
> [3] https://www.infoq.com/articles/nashorn
> 
> ---------------
> On Sat, Dec 31, 2016 at 5:14 PM, [email protected]
> <[email protected]> wrote:
>> Isn't there a typo in the first expression?
> 
> I've looked but don't see it.  btw the first '$' is the shell prompt,
> the second '$' is a variable holding the vm location, since my
> download of bootstrap.image wasn't located under pharolauncher.
> 
> But I do notice the second expression should have been...
> $  $PHARO bootstrap.image eval "Object class allSubclasses size"
> 

Reply via email to