On Fri, Sep 3, 2010 at 4:33 PM, OQ <overlordq(a)gmail.com> wrote:
I also was going through the
schema and noticed there was a decent disparity between table indexes
on what is indexed and what the indexes cover wrt MySQL and Postgres.
This is almost certainly deliberate, no? PostgreSQL supports
different types of indexes and can use them in different ways. For
instance, it can filter using one index and sort using another,
although not as efficiently as if it were one index.
On Fri, Sep 3, 2010 at 4:43 PM, Max Semenik <maxsem.wiki(a)gmail.com> wrote:
Most updates for MySQL will just work for SQLite.
You can use php maintenance/sqlite.php --check-syntax <filename>
to check it.
Error: SQLite support not found
Can't check SQL syntax: SQLite not found
Backtrace:
#0 /var/www/git-trunk/phase3/maintenance/sqlite.php(123):
Sqlite::checkSqlSyntax(Array)
#1 /var/www/git-trunk/phase3/maintenance/sqlite.php(48):
SqliteMaintenance->checkSyntax()
#2 /var/www/git-trunk/phase3/maintenance/doMaintenance.php(104):
SqliteMaintenance->execute()
#3 /var/www/git-trunk/phase3/maintenance/sqlite.php(133):
require_once('/var/www/git-tr...')
#4 {main}
Sigh, stupid distributors. Anyway, after installing php5-sqlite, both
of the patches now fail because CHANGE isn't supported. I guess it
would work if I just changed CHANGE to DROP then ADD? That would wipe
out cl_sortkey's contents on MySQL, but I suppose it's not a big deal.
Anyway, if the patches usually work without modification, why doesn't
SQLite just use the same patches as MySQL, with special-casing for
exceptions? Even if the status quo were documented (it's not), it's
not ideal to expect people who never use SQLite to go out of their way
and take extra steps to keep it up-to-date.