Iterators/vectors (was: [Biocorba-l] BSANE and bioCORBA)

[Michael Dickson]

I'll add my "vote" for the original patterns.  We discussed precisely 
the issues Antoine raised during the initial discussion which is why we 
settled on the patterns we did.

If I had to change anything about it (and we don't :)  I'd replace the 
typesafe iterators with one generic one.  Its a small point but the 
other methods generates lots of stubs and skeletons (for languages that 
have them).  I generally prefer type safety but in practice this proved 
to be a bit of a pain.  So the code you'd likely see would look like:


Object obj;
while (someIterator.next(&obj)) {
  TypedObj tobj = TypedObjHelper.narrow(obj);
  tobj.dostuff();
}

Instead of:

TypedObject obj;
while (someIterator.next(&tobj)) {
  tobj.dostuff();
}

I don't feel religious about it but I'd clearly prefer it.

Mike

Juha Muilu wrote:

>Hi Antoine,  
>Nice to have you on the list!
>
>So we will keep the iterator/list pattern as used in the BSA. 
>
>I have found it quite good to implement and use.
>
>
>Antoine van Gelder wrote:
>
>>On 31 May 2001 23:39:19 +0100, Martin Senger wrote:
>>
>>>  Although I thing that it is not big issue I feel that implementation
>>>with 'out' parameter and returning boolean is easier and also it saves one
>>>network call.
>>>   I do not like raising any Exception as a "normal" behaviour - that
>>>makes the codeclumsy and less readable.
>>>
>>Iterator patterns that return bool are much cleaner as it allows for
>>code such as:
>>
>>Object obj;
>>while (someIterator.next(&obj)) {
>>  obj.dostuff();
>>}
>>
>>Instead of:
>>
>>while (someIterator.has_more()) {
>>  Object obj = someIterator.theObject();
>>}
>>
>>Which is heavier on both network calls and object constructor calls.
>>
>>Or even worse if I understood Brad correctly:
>>
>>try {
>>  while (someIterator.hasNext()) {
>>    SomeObj obj = someIterator.the_object();
>>  }
>>} catch (SomeException) {
>>  // extremely cpu expensive way of ending a loop adding huge
>>  // overhead to even the simplest algorithm not to mention
>>  // the pain you cause in terms of memory management for
>>  // c++ developers.
>>}
>>
>> - a
>>
>>_______________________________________________
>>Biocorba-l mailing list
>>Biocorba-l@biocorba.org
>>http://www.biocorba.org/mailman/listinfo/biocorba-l
>>