Page MenuHomePhabricator

Investigate PNG metadata in thumbnails
Closed, ResolvedPublic


What are we stripping? What are we keeping? Could we safely strip more than what we currently are?

Event Timeline

Gilles claimed this task.
Gilles raised the priority of this task from to Low.
Gilles updated the task description. (Show Details)
Gilles added subscribers: Tbayer, MaxSem, igrigorik and 5 others.

At the very least, the useless crap from should be removed (Can be done by changing im arguments)

From T20871 it seems all structured metadata is stripped and the commons URL is added in a comment. You could save a couple bytes by using MediaWiki-extensions-ShortUrl.

There is also a longstanding request (same task) to strip less, ie. put the author and license in the thumbnail metadata.

As @Bawolff pointed out, we strip most metadata from PNG thumbnails but not "thumbnail metadata". Which manifests itself as these fields when inspecting a thumbnail with exiftool:

Thumb Document Pages            : 1
Thumb Imageheight               : 4608
Thumb Image Width               : 3072
Thumb Mimetype                  : image/png
Thumb M Time                    : 1455745807

@Bawolff which IM arguments do you know to be capable of removing these fields?

We figured out what's filtered and what could be, I think that "going the extra mile", if it's possible, should be a separate task.