[saga-rg] Re: ISSUE 108
John Shalf
JShalf at lbl.gov
Fri Apr 28 13:25:24 CDT 2006
I would pose that question to the SIDL/Babel developers at LLNL, but
I think it may be an oversight in their document. There should be
some convention for handling the class methods more rigorously.
-john
On Apr 27, 2006, at 12:50 PM, Andre Merzky wrote:
> Ah, I was not aware of that point in SIDL - that makes
> sense.
>
> That leaves out constness for class methods, as in:
>
> string get_attribute (string key) const; // is const
> string set_attribute (string key, string val); // is not const
>
> Is there some SIDL convention for these as well? I'm afraid
> I parsed through the SIDL doc a couple of times, but it
> still leaves me puzzled more often than not :-(
>
> Thanks, Andre.
>
>
>
> Quoting [John Shalf] (Apr 27 2006):
>> Cc: SAGA RG <saga-rg at ggf.org>
>> From: John Shalf <jshalf at lbl.gov>
>> Subject: Re: [saga-rg] Re: ISSUE 108
>> Date: Thu, 27 Apr 2006 12:11:24 -0700
>> To: Andre Merzky <andre at merzky.net>
>>
>>
>> I think this question is addressed implicitly by the adherence to
>> SIDL interface definitions. In SIDL, you define variables as "in",
>> "out" or "in/out". A variable that is exclusively "in" is implicitly
>> a const and can be trivially mapped as such in language that support
>> const. So, we *do* specify const-ness -- we just use a language-
>> independent way of expressing the function of the subroutine
>> parameters.
>>
>> -john
>>
>> On Apr 27, 2006, at 9:37 AM, Andre Merzky wrote:
>>> My opinion: we should define constness in the language
>>> bindings.
>>>
>>> Reason: not all languages support const
>>>
>>> Counter-argument: state constness of objects and parameters
>>> should not vary from one binding to the next. Well, that
>>> can also be solved by synchronizing the bindings in terms of
>>> constness.
>>>
>>> Cheers, Andre.
>>>
>>>
>>>
>>> Quoting [Andre Merzky] (Apr 20 2006):
>>>> Date: Thu, 20 Apr 2006 18:55:33 +0200
>>>> From: Andre Merzky <andre at merzky.net>
>>>> To: SAGA RG <saga-rg at ggf.org>
>>>> Subject: ISSUE 108
>>>>
>>>> 108) explain why we don't specify constness. Or should we?
>>>> - OPEN, URGENT
>>>>
>>>> Obvious again. Opinions?
>>>>
>>>> Cheers, Andre.
>>> --
>>> "So much time, so little to do..." -- Garfield
>>>
>
>
>
> --
> "So much time, so little to do..." -- Garfield
More information about the saga-rg
mailing list