[OpenJDK 2D-Dev]  Review request for 8041460, PNGImageWriter creates metadata with incorrect bitDepth
philip.race at oracle.com
Tue Jul 8 04:43:36 UTC 2014
PS the comments referring to bug id et al seem too much.
I am not sure we need a new comment here just because there was a bug.
That may be something sustaining may want to do in 'old' releases but
the ongoing clutter for the mainstream release isn't something we want.
On 7/7/14 2:45 PM, Phil Race wrote:
> Since its returning an array index into theIHDR_bitDepths array I
> think this should be written as
> IHDR_bitDepth =IHDR_bitDepths[getEnumeratedAttribute(node,
> "bitDepth", IHDR_bitDepths)];
> .. but ensuring the line <=80 chars long
> Also note that this is a complete duplicate of
> Always remember to look for duplicates !
> So I have closed 8041460 and you should push under 4991647 - once
> the code is fixed and reviewed.
> I have assigned 4991647 to you.
> On 6/23/2014 6:09 AM, Andrew Brygin wrote:
>> Hello Mikhail,
>> the fix looks fine to me.
>> On 6/9/2014 10:32 PM, mikhail cherkasov wrote:
>>> Hi all,
>>> please review the fix:
>>> for the following bug:
>>> The problem is that during merging of metadata trees we take index
>>> of IHDR_bitDepths
>>> and set this value to IHDR_bitDepth, but IHDR_bitDepth requires a
>>> real number of bits.
>>> IHDR_bitDepths - store all valid values: "1", "2", "4", "8", "16",
>>> so to get
>>> valid value we need: 1 << index.
More information about the 2d-dev