[Seaside-dev] Seaside2.8a1-mb.508

Michael Lucas-Smith mlucas-smith at cincom.com
Sat Oct 20 20:01:22 UTC 2007


>> This change was an effort toward the code freezeness actually
>>     
>
> The freeze was two months ago. As far as I am concered this amout of
> changes you guys rush in on the finish line seriously endagers a
> timely 2.8 release. The idea of the freeze was to give code a time to
> sabilize and fix bugs, not to add features.
>
>   
Undeclareds and Undefineds -are- bugs.
>> we want
>> Seaside and its commonly used components to load cleanly in to Cincom
>> Smalltalk without undeclareds and undefineds, etc. In this case, Async
>> can depend on Scriptaculous if Scriptaculous is loaded. We cannot
>> introduce the Cincom Smalltalk #{} syntax for variable bindings, so
>> instead we pushed it to the "platform compatibility" layer -
>> SeasidePlatformSupport.
>>     
>
> Huh? If I look at I see it only used in one place, this is importing
> JavaScripts and to quote form the code it is a:
> "Naughty, dirty, awful hack"
> We are not talking about having parts that use Scriptcaculous and
> other that don't. If you have that, just make two packages and one
> depend on Scriptcaculous.
>   
We're not changing the code or the logic of the code. It's in code 
freeze, we're making the particular way the code works more platform 
independent. We're being quite careful not to change the -logic- of the 
code.. because it's in code freeze.
> To sum it I don't see the need for #ifDefined:do:
>   
Different platforms do "is this class there" differently. For example, 
in VA you use abrAsClass on the symbol (there are a few other variants 
too). The need is quite clear as far as I can see.


More information about the seaside-dev mailing list