I think we can
- get all plugins,and for each
- write a file in that format
- read it back and apply the test
It is also worth verifying that the writer abort checks are in sync with
the reader aborts,
ie happen at such equivalent points as might exist.
-phil.
On 08/15/2016 11:30 AM, Sergey Bylokhov wrote:
Is it possible to unify the test for all our plugins? I assume they
should work in the same way. I am not sure but probably the image can
be generated at runtime?
On 11.08.16 21:59, Jayathirth D V wrote:
Hi,
Please review the following fix in JDK9 at your convenience:
Bug : https://bugs.openjdk.java.net/browse/JDK-4924727
Webrev : http://cr.openjdk.java.net/~jdv/4924727/webrev.00/
Issue : When we issue ImageReader.abort() in
IIOReadProgressListener.imageStarted(), reading is not aborted and it is
continued.
Root cause : After IIOReadProgressListener.imageStarted() call in
PNGImageReader.java when we enter decodeImage() we call
clearAbortRequest() which will clear the abort request from
IIOReadProgressListener.imageStarted().
Solution : clearAbortRequest() documentation mentions that it should be
called before reading of image starts, so it should be called before
IIOReadProgressListener.imageStarted()(In PNGImageReader.java it is
processImageStarted(0) in readImage()). So moved clearAbortRequest()
call from decodeImage() to readImage(). Also we should call
abortRequested() in PNGImageReader.java at places mapping to
IIOReadProgressListener and not randomly at end of functions or at
places related to IIOReadUpdateListener, updated the code for the same.
Observation not related to this issue : We don’t have call similar to
IIOReadProgressListener.readAborted() in IIOReadUpdateListener, but user
can call ImageReader.abort() from IIOReadUpdateListener methods. Is
there a need to add similar method in IIOReadUpdateListener? Any inputs
on this also would be helpful.
Thanks,
Jay
