Primitive numbering

[Gérard Bunel]

Hi All,

I am on the way to try implementing Behavior>>selectors as a primitive.

My question is: Is there some rule to define primitive number I should use for this ?

In fact I already did it using p500 but not sure that's correct numbering.

My previous post about Class hierarchy problem was related to the implementation of Behavior>>allSelectors which uses Behavior>>selectors.

Gerard

[James Ladd]

Glad you a doing this.

Use p500 to start with and when I get the pull request I can adjust.

Again - thanks for this

Sent from Hyperspace.

--
You received this message because you are subscribed to the Google Groups "Redline Smalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to redline-smallt...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Gérard Bunel]

Hi James,

I wanted to push my modifications, but Git requires a login/password I've never set anywhere.

What's the procedure ? (BTW I'm not familliar with Git. I generally use Subversion)

[James Ladd]

Hi Gerard,

You need to setup a github account then fork the Redline Smallalk repo.

You can then clone/checkout your repo and work on it, commit and push is

back to your repo.

When you have done all the work you want to do you can then use github to
issue a pull request. That is where I see what you have done and then

merge or reject the changes.

You wont loose your current changes if you back them up first, then clone your
repo and copy the files back.

Please let me know if you need more help.

- James.

[Gérard Bunel]

Hi James,

Did you get my changes ?

 

Gerard

[James Ladd]

No - ill check again ASAP 

Sent from Hyperspace.

--

[Gérard Bunel]

So James,

Did you get them ?

I'm currently trying to implement a solution to the Class Hierarchy problem by creating a special Class called PrimNil (which is a singleton) so that this object can override some PrimObject behavior.

basically:

equals() : which return true when compared to PrimObject.NIL

and also, by implementing getAttribute(int index) in PrimObject and overiding in PrimNil by returning Primobject.NIL.getAttribute(index)

This way PrimObject.PRIM_NIL can be seen as nil in the smalltalk part.

All current tests are OK. So I'm expecting this behavior OK. What do you think about this ?

Gérard

[Gérard Bunel]

So James,

Did you get them ?

I'm currently implementing a solution for the Class hierarchy problem I mentioned in a previous post.

The solution is based on the implemenattion of a specific class: PrimNil which sole instance is PrimObject.PRIM_NIL.

This object answers true when compared to PrimObject.NIl

And returns attributes from PrimObject.NIl when requested (typically requested by IndyBootstrap I also changed so that it requests attributes through a getAttribute(int index) method instead of a direct access to the attributes field)

This way iterating over Class Hierarchy does not throw NPE and so that I can implement the allSelectors method in the Behavior class.

Currently all tests are OK, but I wanted to have your opinion about this implementation first.

Wanted also to know if my previous push was OK before doing the next one.

Gérard

Le mardi 7 mai 2013 23:45:06 UTC+2, jamesl a écrit :

[James Ladd]

It's Friday night - ill try and look at this ASAP - I don't see a pull request but ill check again - thank you for your patience 

Sent from Hyperspace.

[James Ladd]

I just looked and I have not received any pull requests.

However, you should probably have a look at
http://www.redline.st/blog/2013/05/25/making-things-simpler.html

There were some BIG changes.