Alan Hudson wrote:
> Paul Sandoz wrote:
>
>> Alan Hudson wrote:
>>
>>> Seems I'm running off the end of the octetBuffer after some number of
>>> encoding algo runs.
>>>
>>> I don't see any place that decrements _octetBufferStart. It also
>>> seems like the ensureCapacity may not be called in enough places.
>>> I'm pretty tired now so it will likely be Monday before I can dig
>>> into why this one breaks.
>>
>>
>>
>> Not enough info to tell if this is a encoding or decoding issue.
>>
>> _octetBufferStart should never be decremented. It defines the starting
>> index the octets of encoding algorithm data in the _octetBuffer.
>>
>> _octetBufferLength defines the length of the octets of encoding
>> algorithm data in the _octetBuffer.
>>
> Its happening during decoding. I can somewhat solve the problem by
> making _octetBuffer much larger... but eventually it always gets full.
> I don't see any logic for actually reseting back to beginning of this
> array. Ie we can't just grow this array forever?
>
There could be something wrong with the Decoder.ensureOctetBufferSize
method.
When content is encountered (UTF-8 encoded string, encoding algorithm
data) etc it is checked to see if the buffer can hold the required
amount according to the length of the content. It should not keep
growing....
>>
>>> I'm pretty close to having Xj3D work with FI again.
>>
>>
>>
>> Great. Sorry for not making it easier for you :-(
>>
> No problem. I expected as much moving from one version to another. Its
> been pretty smooth so far. This issue is the most vexing at the moment.
>
OK.
Paul.
--
| ? + ? = To question
----------------\
Paul Sandoz
x38109
+33-4-76188109