| Lea_WMDE | |
| Mar 10 2017, 1:05 PM |
| F6859436: pasted_file | |
| Mar 22 2017, 5:07 PM |
On loading the SpecialPage, check whether this file has already been uploaded to commons. If it hasn't, proceed to the page. If it has, instead of the normal page, return the following to the user:
The file you are currently trying to upload already exists on Commons.
You can find it under this link: <link>
<Image as it has been found on Commons>
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Handle duplicate files gracefully | mediawiki/extensions/FileImporter | master | +214 -14 |
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | None | T140462 Correctly move files from Wikipedia to Commons (2013) | |||
| Resolved | thiemowmde | T146666 Build a commons extension that moves files to Commons | |||
| Resolved | thiemowmde | T156792 Implement SpecialPage for FileImporter extension | |||
| Resolved | Addshore | T160166 Handle duplicate files gracefully |
Change 342845 had a related patch set uploaded (by Addshore):
[mediawiki/extensions/FileImporter] WIP Handle duplicate files gracefully
@Lea_WMDE I have a few questions here that need clarifying.
The current file is by far the most important one. Therefore, I would only indicate to the user if that exact file exists on Commons already. For all other cases, I would not do any action as the file as a predecessor of the current file is outdated for the one case the moved file is used for, but on Commons might be still current in other situations. Therefore I would be ok with having it double
I guess really a decision needs to be made from a matric roughly as below:
| Look at the most recent revision being imported ONLY | Look at all revisions being imported | |
| Look at the most recent revision on commons ONLY | ||
| Look at all revisions on commons (not including deleted files) | ||
| Look at all revisions on commons (including deleted files) | ||
We need a tick in 1 box for the intended behaviour.
I would suggest this.
| Look at the most recent revision being imported ONLY | Look at all revisions being imported | |
| Look at the most recent revision on commons ONLY | x | |
| Look at all revisions on commons (not including deleted files) | ||
| Look at all revisions on commons (including deleted files) | ||
Any thoughts as to why this would be a bad/good idea?
Just to be nitpicking; what does "this file has already been uploaded to commons" actually mean? An identical file, a grossly similar file, what about a transformed or cropped file?
@jeblad: I strongly assume an identical file. Everything else would most likely be a) hard to check and/or b) give too many false positives/negatives.
Yup, with the way this is currently implemented only EXACT duplicate files will be stopped at this point.
And indeed, doing anything else would probably be too hard / not worth it.
Change 342845 merged by jenkins-bot:
[mediawiki/extensions/FileImporter] Handle duplicate files gracefully
@Lea_WMDE @Jan_Dittrich for reference I have attached a screenshot of how this currently looks below.