On Wed, Aug 4, 2010 at 7:38 AM, MZMcBride z@mzmcbride.com wrote:
Ariel T. Glenn wrote:
I"m all for the change, but it would have to be announced well in
advance of rollout and coordinated with other folks. For example,
I
have a check against rev_len (in bytes) when writing out XML dumps,
in
order to avoid rev id and rev content out of sync errors that we
have
run into multiple times in the past. That code would need to be
changed
to count characters of the text being used for prefetch
instead of
bytes.
Are character counts between programming languages generally consistent? And is there a performance concern with counting characters vs. counting bytes? Another post in this thread suggested that it might be up to five times slower when counting characters. I've no idea if this is accurate, but even a small increase could have a nasty impact on dump-processing scripts (as opposed to the negligible impact on revision table inserts).
Well, why not add the field, but don't populate it for old entries? This way new revisions have both char and byte counts while old revisions only have bytecount populated. Then, when you edit a page, MW checks: has the now-old revision the charcount field populated? If not, update the data, if yes, just save the new revision.
Marco