Thanks Mukul, thanks Michael for pointing out this important detail.
Actually, I didn't pay my attention to those top level comments. my
bad!. Well, I just wished that it'd be better if there were any
indication about interning parameters in method-level comments as
well. But that's OK.
Thanks.

On 7/26/10, Michael Glavassevich <[email protected]> wrote:
>
> Hi Ishan,
>
> There is a comment at the top of the NamespaceSupport:
>
> /**
>  * Namespace support for XML document handlers. This class doesn't
>  * perform any error checking. and assumes that all strings passed
>  * as arguments to methods are unique symbols. The SymbolTable class
>  * can be used for this purpose.
>  */
>
> and at the top of NamespaceContext (the interface which NamespaceSupport
> implements):
>
> /**
>  * Represents an interface to query namespace information.
>  * <p>
>  * The prefix and namespace must be identical references for equal strings,
> thus
>  * each string should be internalized (@see String.intern())
>  * or added to the <code>SymbolTable</code>
>  *
>  * @see <a
> href="../../../../../xerces2/org/apache/xerces/util/SymbolTable.html">
>  * org.apache.xerces.util.SymbolTable..</a>
>  */
>
> which describes what is expected of the user for every method they call on
> a NamespaceContext (which has a string argument). Strings for XML names and
> URIs are interned just about everywhere in Xerces, with strings compared
> using '==' for performance reasons in many other places. It's something
> that all Xerces developers should be aware of. It's an integral part of the
> design.
>
> Thanks.
>
> Michael Glavassevich
> XML Parser Development
> IBM Toronto Lab
> E-mail: [email protected]
> E-mail: [email protected]
>
> Ishan Jayawardena <[email protected]> wrote on 07/26/2010 07:44:25 AM:
>
>> Hi Michael and Mukul,
>> Actually, while I was developing the SCD parser, I came across this
>> same problem that Mukul had. I was using these getURI() and
>> declarePrefix() methods and I couldn't get the code working until I
>> changed the input parameters into interened parameters, which I did
>> without any underlying details about those methods as a result of a
>> long trial and error process. But I guess this is not a very good
>> characteristic of the API and instead we have to tell the programmers
>> explicitly that when using such methods, they are supposed to pass
>> interned parameters. WDYT?
>> Thanks.
>>
>> On 7/24/10, Mukul Gandhi <[email protected]> wrote:
>> > On Sat, Jul 24, 2010 at 11:01 PM, Michael Glavassevich
>> > <[email protected]> wrote:
>> >> FYI: You missed one spot. I've just fixed it.
>> >
>> > oops :(
>> >
>> > I just noticed that. Thanks for the correction.
>> >
>> >
>> > --
>> > Regards,
>> > Mukul Gandhi
>> >
>> > ---------------------------------------------------------------------
>> > To unsubscribe, e-mail: [email protected]
>> > For additional commands, e-mail: [email protected]
>> >
>> >
>>
>>
>> --
>> Best Regards,
>> Ishan Jayawardena.
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [email protected]
>> For additional commands, e-mail: [email protected]


-- 
Best Regards,
Ishan Jayawardena.

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to