Last modified: 2008-01-15 15:40:08 UTC
I would propose enable following parameters to the default rollback summary: $3: number of revisions reverted $4: revid reverted to $5: timestamp for revid reverted to $6: revid reverted from $7: timestamp for revid reverted from
Can't you just get all this information by looking at the page history? WONTFIX?
(In reply to comment #1) > Can't you just get all this information by looking at the page history? > WONTFIX? > But that would require slower client side checks.
(In reply to comment #2) > But that would require slower client side checks. Expanding these variables in *every* rollback summary would slow down human-made rollbacks, which I suspect are more numerous than bot-made ones.
(In reply to comment #3) > (In reply to comment #2) > > But that would require slower client side checks. > Expanding these variables in *every* rollback summary would slow down > human-made rollbacks, which I suspect are more numerous than bot-made ones. > Not really, as long as the revision count one is not included. The useful ones seems to be: $1: reverted user $2: user reverted to $3: revid reverted to $4: timestamp for revid reverted to
Actually I think only $6 and $7 require fetching additional DB rows.
(In reply to comment #5) > Actually I think only $6 and $7 require fetching additional DB rows. > The "number of revisions reverted" would need one.
(In reply to comment #6) > The "number of revisions reverted" would need one. Ouch, yeah, that one would be particularly expensive.
(In reply to comment #7) > (In reply to comment #6) > > The "number of revisions reverted" would need one. > Ouch, yeah, that one would be particularly expensive. > wouldn't the rollback routine allready have such a number?
(In reply to comment #8) > wouldn't the rollback routine allready have such a number? No. All it cares about is the revision being reverted to: it fetches the text from that revision, then creates a new revision identical to it. Whatever may or may not have happened in between doesn't matter.
(In reply to comment #9) > (In reply to comment #8) > > wouldn't the rollback routine allready have such a number? > > No. All it cares about is the revision being reverted to: it fetches the text > from that revision, then creates a new revision identical to it. Whatever may > or may not have happened in between doesn't matter. > aha, ok. PS: get on IRC!
Added in r29794: $3: revid of the revision reverted to $4: timestamp of the revision reverted to $5: revid of the revision reverted from $6: timestamp of the revision reverted from Already present: $1: User whose edits are reverted to $2: Victim My custom rollback message right now (to give you guys an idea) is: Reverted revision $5 ($6) and earlier revisions by $2, reverted to revision $3 ($4) by $1
