Page MenuHomePhabricator

ProofreadPage does not provide scans for files wich 0x0 size in metadata, despite of the size being set manually in the index
Open, Needs TriagePublicBUG REPORT

Description

List of steps to reproduce (step by step, including full links if applicable):

  • using a file with broken metadata, eg.

https://commons.wikimedia.org/wiki/File:PL_Hudson_Jej_naga_stopa.pdf

What happens?:
scans of its pages are not provided in the Page namespace, eg.
https://pl.wikisource.org/w/index.php?title=Strona:PL_Hudson_Jej_naga_stopa.pdf/43&action=edit&redlink=1
while they can be still generated by the system, eg.
https://upload.wikimedia.org/wikipedia/commons/thumb/7/7c/PL_Hudson_Jej_naga_stopa.pdf/page43-1000px-PL_Hudson_Jej_naga_stopa.pdf.jpg
Even if the requested scan width being set in the index:
https://pl.wikisource.org/wiki/Specjalna:Przenie%C5%9B/Indeks:PL_Hudson_Jej_naga_stopa.pdf
(1000px here; this value is intended to override the default with even if scalling up is requested)

What should have happened instead?:
Scan should be genereted and provided, if it is possible. The manually set value should always override the size from metadata. This would also make the system more immune to various errors, like in the provided example.

Software version (if not a Wikimedia wiki), browser information, screenshots, other information, etc.:
any

Event Timeline

and one more file (on frwikisource) -> the books is a file (restored, because clearly Public domain, but the original data were wrong)...

the file does not display properly on commons or on wikisource

please also see T298423 -> which seems to be the same problem