[Uim] lazy-loading

[YamaKen]

At Tue, 18 Jan 2005 14:32:53 +0900,
ekato at ees.hokudai.ac.jp wrote:
> 
> On 2005/01/18, at 14:10, YamaKen wrote:
> > Since the full-validation requires full implementation of
> > custom.scm, some consideration are required. I'm not sure
> > whether we should do full-validation of custom variables that
> > received via helper protocol, at now.
> 
> Hmm, how about just sending changed variables only instead of all 
> variables?  But this may be also difficult...

That is not the problem.

I meant full-validation as existence, type and range checking.
Although the existence of a custom variable can be checked with
custom-rt.scm, others are not. And custom type 'key' requires
type information to execute define-key.

Additional type information in the protocol message or type
information management is required to resolve it. But I'm not
sure what solution is the best, yet.


Thank you for the constructive discussion :)

-------------------------------
YamaKen  yamaken at bp.iij4u.or.jp