Thanks to Bob and you for your accurate answer.

Cheers,
Alexandre
 

On Mon, Mar 03, 2003 at 07:31:03PM -0600, Tim Olson wrote:
> Alexandre Bergel <bergel at iam.unibe.ch> wrote:
> | Hello,
> | 
> | I am wondering why the following expression is false.
> | -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> | nil class compile: 'foobar Object new asMorph'.
> | (nil class >> #foobar) literals includesAllOf: #(new asMorph)
> | -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> 
> If you browse UndefinedObject>>foobar and view the bytecodes, you will
> see:
> 
> 13 <41> pushLit: Object
> 14 <CC> send: new
> 15 <D0> send: asMorph
> 16 <87> pop
> 17 <78> returnSelf
> 
> If you then look at Interpreter class >> initializeBytecodeTable, you
> will find that 16rCC (204) is the bytecode for:
> 
> 		(204 bytecodePrimNew)
> 
> sending "new" is common enough that it has its own bytecode, so there is
> no need for a literal reference.  If you decode the "asMorph" send,
> though, you will see 16rD0 (208) translates to 
> 
> 		(208 255 sendLiteralSelectorBytecode)
> 
> 	-- tim

-- 
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Bergel Alexandre  http://www.iam.unibe.ch/~bergel
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.