Page MenuHomePhabricator

Files/images should be fetched in batch by parser, not one at a time
Open, MediumPublic


File objects are looked up in the Parser one at a time as they're encountered. This isn't too bad on Wikimedia sites because all the file repositories are on the local network, but for third-party sites running with InstantCommons on, pages with lots of images can take a *long* time to render.

Parser::fetchFileNoRegister() does an individual wfFindFile() call. It could be *much* more efficient to call RepoGroup::findFiles() in batches; for InstantCommons images this would allow for batching the API requests, reducing the round-trip time significantly.

Version: 1.22.0
Severity: normal



Event Timeline

bzimport raised the priority of this task from to Medium.Nov 22 2014, 2:09 AM
bzimport set Reference to bz54033.
bzimport added a subscriber: Unknown Object (MLST).