A bunch of compressed old entries on es.wiktionary.org are corrupted. For example, history entries on the Portada are corrupt from June 16 through August 11 and September 24 through October 25.
A spot-check of corrupt entries shows a byte pattern very reminiscent of a text conversion from Latin-1 to UTF-8. I need to know when this wiki was converted, and exactly what was done; are there prior backups? Is the conversion reversible? Might we have the same problem on other wikis?
-- brion vibber (brion @ pobox.com)
On Nov 27, 2004, at 10:19 PM, Brion Vibber wrote:
A bunch of compressed old entries on es.wiktionary.org are corrupted. For example, history entries on the Portada are corrupt from June 16 through August 11 and September 24 through October 25.
It seems shaihulud converted this wiki from latin-1 to UTF-8 on November 14. Some (but not all) old entries were compressed at the time, and the compressed byte streams were damaged by the conversion script. (This corrupt is not reversible, as there are four byte values which get converted to the same sequence indicating an invalid character.)
I've found the backup dump used to make the conversion and should be able to recover all damaged entries from it. es.wiktionary.org is locked to new edits temporarily; it should be back up within a few hours.
-- brion vibber (brion @ pobox.com)
On Nov 27, 2004, at 11:49 PM, Brion Vibber wrote:
I've found the backup dump used to make the conversion and should be able to recover all damaged entries from it. es.wiktionary.org is locked to new edits temporarily; it should be back up within a few hours.
Done; es.wiktionary.org is back online for editing.
-- brion vibber (brion @ pobox.com)
wikitech-l@lists.wikimedia.org