Google has changed images search result page so now thumbnails are embedded into the page code itself as binary data. This leads to Parley's Google image plugin cannot file any images.

The patch tries to change parsing strategy to obtain the address of the original image and download this image instead of the thumbnail from the search page.

Tested on several flashcard files. It takes ~2 seconds on 1MB channel to download an average image from Wikipedia. All images are of good quality, everything seems to work as expected.