As of r86257 [1], which will be deployed to Wikimedia wikis soon and will be included in the 1.17 release, sortkeys output by list=categorymembers and prop=categories are now encoded as hexadecimal strings, so "FOO" becomes "464f4f".

As previously announced, sortkeys are no longer guaranteed to be human-readable, and may in fact contain binary data (this will happen when Wikimedia switches to the UCA/ICU collation). However, outputting binary data, notably in XML, was problematic [2], so I decided to use hexadecimal encoding. This means the sortkey as returned by the API is now guaranteed to not be human-readable, even if the underlying collation uses a human-readable format (such as the uppercase collation currently in use on Wikimedia wikis). However, it will still sort correctly: if A sorts before B in the binary format, that will also be the case in the hexadecimal format.

The following things changed: * The 'sortkey' property in list=categorymembers and prop=categories is now a hexadecimal string * In prop=categories , clprop=sortkey will now also output the 'sortkeyprefix' property (human-readable part of the sortkey). list=categorymembers already provided this through cmprop=sortkeyprefix * The format of cmcontinue has changed from type|pageid|rawsortkey to type|hexsortkey|pageid . If you did not make any assumptions about the format of cmcontinue and just passed back whatever you got in query-continue, this won't affect you

Roan Kattouw (Catrope)

[1] https://secure.wikimedia.org/wikipedia/mediawiki/wiki/Special:Code/MediaWiki... [2] https://bugzilla.wikimedia.org/show_bug.cgi?id=28541

_______________________________________________ Mediawiki-api-announce mailing list Mediawiki-api-announce@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mediawiki-api-announce