https://bugzilla.wikimedia.org/show_bug.cgi?id=164
--- Comment #145 from Roan Kattouw <[email protected]> 2009-11-19 22:13:25 UTC --- (In reply to comment #144) > Some SQL engines (like Oracle, Sybase, Informix, but probably not all index > formats supported in various versions of MySQL) also store the rowID's of the > component tables from which columns are extracted in the SELECT clause, in > order to allow the columns under the read cursor to be updatable, but won't do > that for columns computed from expressions, which are not updatable. This just > requires opening the cursor "FOR UPDATE" (and sometimes, you can also limit > the > list of table columns that need to be updated within the cursor loop, which > remains open during the transaction, in order to minimize the extra storage > for > dependant rowID's). > Let's not make assumptions about individual DB engines here, but use SQL queries that unambiguously state what we want without relying on such behavior. About the whole page ID thing: as long there's a sorting tiebreaker of some kind making every entry unique (not necessarily visibly unique in the UI, just unique to the software), paging will be fine. Whether that tiebreaker is the page ID or the page title or the water level in the San Francisco Bay at the time of the page creation measured in 128ths of inches doesn't matter all that much from the implementation side, although of course only the second is remotely understandable for users. -- Configure bugmail: https://bugzilla.wikimedia.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikibugs-l
