I added ParserModifyImageHTML to core to make it easier for PageImages
to identify images in the lead section.
This also allows PageImages to stop writing to LinksUpdate properties.
It is one of only two extensions that do that.
Depends-On: I24528381e8d24ca8d138bceadb9397c83fd31356
Bug: T176520
Bug: T296895
Change-Id: I472f4a023969bfde6298eb56112c16d2ae842199
$thumb might be an instance of MediaTransformError, which as the name
implies, is not really usable as a thumbnail and should be skipped
instead.
Bug: T290973
Change-Id: Ie818bd295bfe06c7c0dc0a2ea7dd273c7e0e416f
Replace instances of "blacklist" with "denylist" throughout extension.
Bug: T277955
Depends-On: Ib4985ec2fcb22eafad8f3a7cf9fc3161782a71db
Change-Id: Ibe460cb9691d56a9e83686b53c7629b5404af6fb
As we switch ParserCache to JSON, we can no longer serialized
class instances in extension data. PageImages was writing the
full set of properies it received from Parser into it's extension
data, some of which are sometimes class instances. Instead, only
write the nessessary subset of data into extension-data.
This change is completely forward and backward compatible.
Since before this change we were already writing the same array,
but with many additional unused properties.
Bug: T266251
Change-Id: Ieb4a139465159611e6b3a99c4b68c3c174b1944f
SearchResultProvideThumbnail was introduced in T250144. In order to fill
search results with the "thumbnails" field it should be implemented in proper extension.
Bug: T250207
Change-Id: I81d5d70f5237d6399b8ec6bec3058d12e40ca142
