r215211 changed ImageDecoder::isSizeAvailable() to call encodedDataStatus() in the subclasses, which forces a partial decoding of the image to get the size from the file. But ImageDecoder::isSizeAvailable() is also being called inside the decoders in some places. One of this places is the setSize() method inside each decoder. So a call to isSizeAvailable() when the image size is not known causes the image to be partially decoded to get the size, but as a result setSize() gets called, which calls isSizeAvailable() again, which triggers a new decoder because the size has not been set, and we have another infinte loop.
Created attachment 306808 [details]
Comment on attachment 306808 [details]
View in context: https://bugs.webkit.org/attachment.cgi?id=306808&action=review
> + REGRESSION(r215211): [GTK] Lots of image related tests are timign out, causing the test bot to exit early
Fix the typo. timign -> timing
Committed r215227: <http://trac.webkit.org/changeset/215227>