[squeak-dev] Re: Pragmas (Re: The Inbox: Morphic-phite.429.mcz)

Igor Stasenko siguctua at gmail.com
Mon Apr 26 21:33:33 UTC 2010


On 27 April 2010 00:23, Andreas Raab <andreas.raab at gmx.de> wrote:
> On 4/26/2010 1:56 PM, Igor Stasenko wrote:
>>>>
>>>> If pragmas would be the code, then i expect them to look like code i.e.:
>>>
>>> I never claimed that that "pragmas are code". In fact, I said
>>> *specifically*
>>> the opposite.
>>>
>>
>> Ok, then i wonder, what is the subject of discussion?
>
> The subject of the post was an attempt to clarify what pragmas are and what
> they aren't. Given that even Eliot is confused about it by referring to
> "primitive pragmas" that seemed worthwhile all by itself.
>
>> If its only about terminology, then its not interesting. We may call
>> it pragma, method annotation or whatever.
>> I think that more improtant is how we handling it and what we can do with
>> it.
>
> Terminology is important and the terminology we currently have is
> *extremely* confusing. People refer to "pragmas" and whenever a third party
> hears that they think "oh, compiler stuff, better stay away from it" not
> realizing that that's not what these are. This is how this discussion
> started after all.
>
Ok.
Observing a Squeak's history i seen an attempts to allow methods to
use an arbitrary objects
as literals. Or, by analogy, a method's properties can carry any
objects you may need.
I thinks this is a most generic form of what current 'pragmas' could evolve to.
So, from this point of view, pragmas is a subset of 'additional
arbitrary method state', which having a formally
defined structure 'instance of Pragma class'.

> Cheers,
>  - Andreas
>


-- 
Best regards,
Igor Stasenko AKA sig.



More information about the Squeak-dev mailing list