Page MenuHomePhabricator

Transcluding Special:Editcount yields a higher edit count than Special:Editcount itself reports
Open, LowPublic

Description

Look at the edit count in the right box at http://en.battlestarwiki.org/wiki/User:Catrope , which is rendered using {{Special:Editcount/Catrope}}. When viewing Special:Editcount/Catrope itself, however, (just click the edit count number) the number there is significantly lower (note that you have to select the table; it's white text on a white background), the difference being 49 edits at the time of writing.

I suspect this difference is caused by the inclusion or omission of deleted revisions, since that seems like the only logical explanation. I can't check that (I'm not a sysop), but I wouldn't know what else could cause it.


Version: unspecified
Severity: enhancement
URL: http://en.battlestarwiki.org/wiki/User:Catrope

Details

Reference
bz10452

Event Timeline

bzimport raised the priority of this task from to Low.
bzimport set Reference to bz10452.
bzimport added a subscriber: Unknown Object (MLST).
Catrope created this task.Jul 3 2007, 3:50 PM

robchur wrote:

There is no such special page in the core distribution; talk to the maintainer of the extension.

brion added a comment.Jul 3 2007, 6:42 PM

The value on your edit page is cached when the page is rendered, so will not generally be up to date.

(In reply to comment #2)

The value on your edit page is cached when the page is rendered, so will not
generally be up to date.

On a closer read, you'll notice that the 'cached' number on my user page is *higher* than the number Special:Editcount reports.

(In reply to comment #1)

There is no such special page in the core distribution; talk to the maintainer
of the extension.

MediaZilla is also intended to report bugs for extensions, so please change product and component rather than marking INVALID. Reopening.

robchur wrote:

(In reply to comment #3)

MediaZilla is also intended to report bugs for extensions, so please change
product and component rather than marking INVALID. Reopening.

No, this BugZilla is for tracking issues with extensions we use on Wikimedia or which the MediaWiki development team maintain in our repository.

brion added a comment.Jul 3 2007, 6:55 PM

Rob, please note that it is in our SVN.

The transcluded case when no namespace is given calls User::edits(), which uses
the user_editcount field. This may report a different number than the counts
pulled from the revision table as it will handle deleted pages differently.

(In reply to comment #4)

Rob, please note that it is in our SVN.

Exactly, also bug 4400, bug 8388, bug 8453 and bug 9498 were handled normally, and they were about Editcount too. Re-reopening :P

(In reply to comment #5)

The transcluded case when no namespace is given calls User::edits(), which uses
the user_editcount field. This may report a different number than the counts
pulled from the revision table as it will handle deleted pages differently.

I suspected something like that. The real point here is whether the editcount should include deleted revisions. Currently, at least one of these methods is wrong.

Personally, I think deleted revisions shouldn't be included, since they, well, no longer exist. Special:Editcount should update the user_editcount field when it's viewed, as it has to recalculate the count anyway.

brion added a comment.Jul 3 2007, 7:55 PM

The user_editcount field counts how many edits you've made, and exists as a heuristic for measuring use of the account.

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptSep 20 2015, 10:18 AM
This comment was removed by SamanthaNguyen.
lcf119 added a subscriber: lcf119.Sep 25 2016, 9:43 PM

Also, the counts on the revision table may also give higher results, because page moves count as 2 (one revision on the source page, and another revision on the target page). Ideally this extension should use a table with counts for each namespace (see also T173775)