Follow-up to r15791.
You can lie to me, but not to your installer.
Make DatabasePostgres::tableExists to check for real table names, not
faked ones.
DatabasePostgres is currently lying to the rest of the MediaWiki that
"mwuser" table is actually called "user" and that "pagecontents" is
called "text". While MediaWiki does not care, the installer (and updater
do).
This allows us to overcome first hurdle in getting MediaWiki 1.7.3 to
update to trunk on PostgreSQL and uncover further bugs.
For this commit to actually do something, we rename those tables when
upgrading to match what we have in maintenance/postgres/tables.sql
And by the way, tell installer not to check for "user" table, since most
PostgreSQL users will have "mwuser" instead. Picking "archive" instead.
Patchset2: wrapped commit message at 72 chars
Change-Id: Ic874e92bb1adda3430d3292423d4f3617c25456c
You can lie to me, but not to your installer.
Make DatabasePostgres::tableExists to check
for real table names, not faked ones.
DatabasePostgres is currently lying to the rest
of the MediaWiki that "mwuser" table is actually
called "user" and that "pagecontents" is called
"text". While MediaWiki does not care, the
installer (and updater do).
This allows us to overcome first hurdle
in getting MediaWiki 1.7.3 to update to trunk
on PostgreSQL and uncover further bugs.
For this commit to actually do something,
we rename those tables when upgrading to match
what we have in maintenance/postgres/tables.sql
And by the way, tell installer not to check
for "user" table, since most PostgreSQL users
will have "mwuser" instead. Picking "archive"
instead.
* Made PG throw a DBQueryError when it gets a query error, instead of DBUnexpectedError. Apparently this mistake goes back to r14625, when exceptions were first introduced. Did it by removing reportQueryError(), the DatabaseBase version works fine.
* Fixed several places where there was an attempt to check for a query error by checking if the result of query() was false. This never worked. Used try/catch instead.
* Made the DBConnectionError messages go on one line so that they don't mess up the formatting in the installer.
* In DatabasePostgres::selectDB(), only disconnect and reconnect if the DB name is actually changing.
* Made DatabasePostgres::schemaExists() less weird and scary.
* Added DatabasePostgres::roleExists() for use by the installer.
* Removed the PostgreSQL-specific hack to make _InstallUser have a default other than "root". Made _InstallUser into a proper DBMS-specific internal variable instead, since every DBMS we support so far needs a different default.
* Removed the $dbName parameters from openConnection/getConnection, and got rid of $this->useAdmin. Implemented a more sophisticated caching scheme instead. Partial revert of r89389 and r81440.
* When connecting as the install user before DB creation, and when testing the web user's credentials, try a few different database names and use whichever one works.
* Instead of connecting as the web user to create tables, I used SET ROLE. It seems cleaner and more like what the other DBMSes do during installation. "SET ROLE wikiuser" requires the same privileges as "CREATE SCHEMA ... AUTHORIZATION wikiuser", so it's unlikely to break anything.
* In the area of web account creation, fixed various minor logic errors and introduced more informative error messages at the submit stage, pre-install. Show a helpful error message if the web user exists already and the install user can't do the relevant SET ROLE.
* Split schema creation out to a separate install step.
* When creating an account as a non-superuser, add the administrative account to the new account's group. This is necessary to avoid a fatal error during installation (bug 28845).
* Removed code which alters an existing web user to have appropriate search paths and permissions. This may break other apps and is not necessary. As in other DBMSes, If the web user exists, it is the responsibility of the sysadmin to ensure that it has appropriate permissions.
* Rewrote setupPLpgSQL() to use the query builder functions.
The sections whose visibility was being controlled by the checkbox had an initial visibility hardcoded in the page output, which had been manually set to match the expected initial default state. I've switched these to instead check what the variable is set to and base the initial style visibility on that.
Could be better to encapsulate these settings together in some way, but this'll do for now.
Note that the way these are all done now means that the sections can never be used when JS is off; if a fully-functional installer is desired in a JS-free environment, it may be necessary to change these further to default to shown, and do the initial hiding in JavaScript (could have 'flashing' issues though: the bits showing up, then quickly hiding. laaame but oh well)
More importantly, this adds support for installing extension tables *at* install time. Previously, the best we could do was to add the require()s to LocalSettings. However, without a subsequent update, wikis will probably start having errors immediately after install (this is really bad!). So I've added an interface to the DatabaseUpdater passed to LoadExtensionSchemaUpdates. An extension can now call addNewExtension() which will allow the extension to be enabled on install and not just update. Example committed in CodeReview
** Moved the responsibility for calling setupSchemaVars() on install to the DatabaseInstaller subclass. This allows it to be called after setupDatabase() has completed, as required by Oracle and PostgreSQL.
** Fixed OracleInstaller::getConnection() so that it respects $this->useSysDBA correctly.
** In OracleInstaller, added some more variables to the list of schema vars, which are needed by user.sql and tables.sql
** In SearchOracle, specify the database name when calling ctx_ddl.sync_index(). This fixes a fatal error in the createMainpage step, caused by the schema name not being equal to the current user.
* In oracle/tables.sql, fixed a couple of indexes with missing table prefixes.
* Improved debugging output in DatabaseInstaller::getConnection() and Installer::createMainpage().
* In DatabaseBase::selectDB(), set $this->mDBname correctly, as in DatabaseMysql.
* Fixed a bug causing the engine and charset settings to not be properly preserved when adding new tables on upgrade.
* Fixed total breakage of SQLite upgrade, by reusing the administrative connection to the SQLite database instead of creating a new one when wfGetDB() is called. Added LBFactory_Single to support this.
* Introduced a "schema variable" concept to DatabaseBase to avoid the use of globals for communication between the installer and the Database. Removed a lot of old global variable names from Database::replaceVars(), most were only added on a whim and were never used.
* Introduced DatabaseInstaller::getSchemaVars(), to allow schema variables to be supplied by the DatabaseInstaller child classes.
* Removed messages config-mysql-egine-mismatch [sic] and config-mysql-charset-mismatch. In the old installer it was possible for users to request a certain character set for an upgrade, but in the new installer the question is never asked. So these warnings were shown whenever a non-default character set or engine was used in the old database.
* In MysqlInstaller::preUpgrade(), fixed the incorrect strings used to identify the MySQL character sets: mysql5 instead of utf8 and mysql5-binary instead of binary.
* On install, initialise the site_stats table, using code copied from the old installer. Unlike the old installer, use SiteStats to increment the user count when the initial user is added.
* Fixed several instances of inappropriate call-by-reference.
* Replaced call_user_func_array() with call_user_func() where possible, it is shorter and simpler.
* Moved the caching boilerplate for DatabaseInstaller::getConnection() to the base class, and have the derived classes override an uncached function openConnection() instead. Updates r80892.
* In MysqlInstaller::getLocalSettings(), escape PHP strings correctly with LocalSettingsGenerator::escapePhpString().
* Reduce timeout for checks in dirIsExecutable() to 3 seconds, so that it doesn't take 30s to run when apache is in single-threaded mode for debugging.
* MySQL and SQLite have been tested and they appear to work. PostgreSQL upgrade is totally broken, apparently it was like that before I started. The Oracle code is untested.
* Move setup_database() from DatabasePostgres to PostgresInstaller
** Add install step after interwiki for the COMMIT
** Put the table creation test (should we abstract this?) in setupDatabase()
** The leftover stuff in initial_setup() also belongs in setupDatabase(), I think?
* Added associated files for tooltip display
* Added associated css rules for tooltip display
* Updated installer's help system in (most?) places to display help as tooltips instead of show/hide system
* Beginnings of visual changes (e.g., less boxes, more color)
* Modified several of the installer's widget display methods
* Added a couple new i18n values
* Fixed small bug in installer logic:
- Password validation on new users was being executed even if there wasn't a $user object
