Ops Review

[Jonathan Worthington]

chromatic wrote:
> Ops to remove:
> 	set_addr/get_addr
>   
get_addr I know we use in two places in Rakudo (for .WHERE and also for 
default object stringification). We can always make it a dynop, but I 
suspect other languages might want to be able to print the memory 
address of a PMC too for similar reasons.

> 	pushmark/popmark/pushaction
>   
languages/dotnet/ (which is likely somewhat broken due to lack of time 
on my part at the moment...thanks to everyone who has patched it to 
make/keep it a lot less broken than it could have become) relies on 
pushmark/popmark in translating the dotnet exception to work on Parrot. 
It may be possible to track what exceptions are on the stack and deal 
with this another way, I don't know (I do remember it was a royal pain 
to get right...)

> 	interpinfo
>   
Is this not similar to or the same as $P0 = new 'ParrotInterpreter'?

> Object ops:
> 	- could add/remove parent become class methods?
>
>   
They already exist as methods too. Thing is, then we'd have no way from 
the ops to call those vtable methods...not sure how big a deal that is, 
but vtable methods without a way to call them feel odd to me.

Don't see much else to disagree with,

Jonathan