[OpenJDK 2D-Dev] RFR: 8149815: Misleading

[Brian Burkhalter]

Hi Bharath,

On May 19, 2016, at 11:06 AM, Bluv Nallakaluva <bharath.nallakaluva at oracle.com> wrote:

> I think the change looks fine.
>  
> However I want confirm the reason for the check  “type == TIFFTag.LONG”  besides  type == TIFFTag.TIFF_IFD_POINTER .
> From this http://partners.adobe.com/public/developer/en/tiff/TIFFPM6.pdf. , it is clear that IFD /LONG type can be used for the tag “SubIFD”. (this entry points to a Thumbnail IFD)
> Here the IFD entry corresponding to this tag should have the node name “TIFFIFD” which makes sense.
>  
> So there might be other  IFD entries that would have a tag whose type can be LONG and still they have the value as the address to the actual IFD just like the above “SubIFD”
> And for these entries the node name being “TIFFIFD”   is valid.
>  
> Can you confirm if my understanding is correct ?

Yes that is correct. There were sub-IFDs in existence prior to Adobe’s PM spec and they used type LONG. We encountered some more than a decade ago.

> Also I think I caught one more place where this instanceof is used in TIFFFieldNode,
> The following excerpt from TIFFFieldNode constructor. I believe this one should also be fixed.
>  
> isIFD = field.getData() instanceof TIFFDirectory;

You are correct: good catch - thanks.

I’ll update the webrev.

Thanks,

Brian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/2d-dev/attachments/20160519/bee36730/attachment.html>