jakarta-bsf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rony G. Flatscher" <Rony.Flatsc...@wu-wien.ac.at>
Subject Re: Planned change on TypeConvertorRegistry.java
Date Sat, 27 Aug 2005 19:57:54 GMT
Hi Sanka,

thank you for your feedback!

>>thinking about String-->Object type conversion to incorporate into BSF 
>>it will mostlikely make sense to do the same for the primitive datatype 
>>wrapper classes like Byte and the like, or no type conversion at all (in 
>>which case it is left to the BSF engine implementors to convert on their 
>>own). Again, such a need would occur only if a Java field/method 
>>argument needs type Object values only.
>>So what would you think, would it add too much overhead or would it be 
>IMUO I think it'll add too much of overhead to TypeConvertorRegistry.
>Adding such a conversion might result in the necessity  to add quite a number of other
type conversions (e.g. Char[] -> Object[]) which I think too much overhead. 
Well, arrays would not be primitive datatypes, so this would not pertain 
to them. Although, you are right, if thinking about char[]->String, 
byte[]->String, String->char[], or String->byte[].

OTOH, the penalty would be mostly in the lookup of the type convertors, 
which shouldn't be that bad, after all.



  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message