Last modified: 2010-07-02 07:24:24 UTC
Nearly 30% of the revisions marked as new in codereview, predate this extension, and don't really need explicitally reviewing Requesting flag to be able to mark these old!
<MaxSem> 6709 <Reedy> heh, ta Max <AzaToth> ok <MaxSem> for the reference: select count(*) from code_rev where cr_status='new' and cr_timestamp > 20081101000000 <Reedy> It's still nearly 30% of the "new" commits Nearly 30% postdate codereview even. Meaning 70% predates
Created attachment 7544 [details] Rough patch Basic patch... Needs a different CSS colour setting for old (maybe?) Also, something in relation to brions "fixme" - // $wgExtNewFields[] = array( 'code_rev', "$base/archives/codereview-cr_status.sql" ); // FIXME FIXME this is a change to options... don't know how Need a way to alter cols in extensions. Or something
Further poking.. Can we do something with $wgDatabase->fieldInfo() ...? Do we get enough info? Or can we get enough info (see if 'old' exists?) http://www.php.net/manual/en/function.mysql-fetch-field.php#20082 Suggests if we get the flags, we'll get it... enum_value_exists for updaters.inc...?
CC'ing brion (Hope he doesn't mind! :P)
Just use the updatelog table.
Can we just do something like $wgExtNewFields[] = array( 'code_rev_enum_add_old', "$base/archives/codereview-cr_status.sql" ); and get away with it?
Created attachment 7546 [details] Minor improved patch Use separate file for enum update. Changed addition to wgExtNewFields to use 3 array indicies. Use "code_rev_enum_add_old" to do the addition, hopefully should mean that it doesn't get run again when added to update log No support attempted for SQLite. The update may work fine.. And can just be applied to both. MaxSem? ;)
Created attachment 7547 [details] Moar with updatelog shizz Addition of insert to code to updatelog Add call to update_row_exists in efCodeReviewSchemaUpdates as conditional to find out if update already done
Reviewed by MaxSem on irc. Minor tweaks made. Will commit
r68871