Gerrit Code Review <gerrit@wikimedia.org> committed
rEEDIe2d8dec217cc: Final NoteDb migration updates (authored by
Turnstep).
Final NoteDb migration updates
Gerrit Code Review <gerrit@wikimedia.org> committed
rEEDIf508f10129ba: Update patch set 1 (authored by
Turnstep).
Update patch set 1
Gerrit Code Review <gerrit@wikimedia.org> committed
rEEDIb4fce827e707: Update patch set 1 (authored by
Turnstep).
Update patch set 1
Gerrit Code Review <gerrit@wikimedia.org> committed
rEEDI5139d2a55b7d: Create change (authored by
Turnstep).
Create change
Gerrit Code Review <gerrit@wikimedia.org> committed
rEEDIf92469027c72: Final NoteDb migration updates (authored by
Turnstep).
Final NoteDb migration updates
Gerrit Code Review <gerrit@wikimedia.org> committed
rEEDI21fb7d102f8f: Create change (authored by
Turnstep).
Create change
Rename tables mwuser to user, and pagecontent to text. Table names are always quoted, so
it doesn't matter if they are keywords. A table prefix can be used if desired to avoid having
table names that are keywords.
Not sure how I missed this task until now. I'm mostly okay with this, but the timestamp change in particular seems a step in the wrong direction.
A good extension name, that explains intuitively what it does and avoids generic words
Ok, well it was just a quick idea. If I get some time I will test this out but most of my test boxes cannot handle 1.28 at the moment due to the recent PHP requirement.
What about simply changing the $conn->query("createdb") to $conn->doQuery in PostgresInstaller.php? We lose some random sanity checks and profiling, but this is a one-time create database call so those are not really important.
A basic first step for questions like this is to see what the major distros (especially the LTS ones) are providing for node versions. As a point of data, looks like CentOS 7 is providing 0.10.46. On the other hand, we usually end up installing a local node anyway, so it won't affect us very much.
As a heavy "third-party MediaWiki operator", I am fine with requiring mbstring. It's standard, in all packaging systems, easy to enable, and seems a low cost for a high payoff.
FWIW, this is now a full-blown extension:
Are there any temporary workarounds? Willing to monkeypatch as heavy as needed....
"If one other person (any pre-existing Phabricator user) wants to put their name down as being unequivocally opposed to releasing 1.27.0 with a PHP 5.5 requirement" Barring better arguments for 5.5 than "has some neat stuff we would like to maybe use", and as someone who has hacked production code to get MW installed on machines with a too-low-but-still-works PHP versions :), I'm opposed for now.
Worked around this with a cronjob executing runJobs and turning off all async calls, but this is a very nasty bug.
No time to fix, but leave a note about odd behavior of this extension for…
Use the new sequence name
Postgres schema updates, per request at ticket 18537.
Modified version of patch from Brad Jorsch, updates Postgres schema to match…
Quick Postgres support. Still have to remove some mysqlisms before it will…
Use correct schema file when running an update.
Postgres version of schema.
Postgres: fix incorrect code in getIndexInfo so that column names are now…
Postgres changes: ALTER TABLE uses TYPE, not ENGINE Compare column type names…
Quick Postgres conversion.
First pass at Postgres support: schema transalation only, app code not tested.
Quick Postgres per request.
We don't want the quoted table name when we are doing a direct lookup via a…
First pass at Postgres support.
Add space before GROUP BY call.
Postgres version of schema - untested.
Remember checkbox setting, properly quote SQL.
Use an ORDER BY to enhance randomness
A custom field may have multiple entries, don't show disabled ones.
Add missing rt-badcfield message.
Allow filtering by custom fields. Redo the way we build queries.
Extract epoch to make all date columns sortable.
Allow limit=xx as an alias for l=xx
Allow tables to be sortable, both globally and per invocation. Bump to version…
Extension for interaction with the Request Tracker system.
Trim the value passed to switchHook.
Quick schema for Postgres
Applied patch from Brad Jorsch, per bug 15834. Fixed Postgres schema and some…
Use single quote, not double, inside SQL statement.
Postgres version of schema. Extension not tested with Postgres yet.
Adjust GROUP BY for databases that need all items from SELECT
Postgres version of the schema.
Add Postgres version of schema.
Add Postgres support. Refactor touchpages concept.
Change hard-coding to useIndexClause()
Postgres version of schema for Drafts extension.
Postgres version of table
Fix syntax errors in last commit.
Skip words if they don't contain all letters.
Put old indexes back in, minor cleanups.
No sense in storing timestamp in three separate indexes.
Don't use non-standard FORCE INDEX, but call useIndexClause() instead
Fix GROUP BY for databases not using implicitGroupby
Existing index should be fine for Postgres.
Load alternate sql file when using Postgres. Make sure we check that blockStart…
IPv6 xff should contain a "xff". Fix array offset.
Use database timestamp() function for things passed directly to db.
Revert: gotta stop editing early in the morning.
Set default for acr_deleted, as code sometimes inserts a null.
Pretty sure this was supposed to match the line below it.
Add primary key, fix errant commas
Use properly formatted timestamp versions when building list of notices.
Use databases concept of a timestamp, not MW's
Use the correct vars for not_start and not_end in the update.
Postgres version of the schema.
Fix some minor space vs. tabs issues.
Change to more standard CASE WHEN THEN form.
Don't try to access array if it's empty.
Remove odd WHERE clause condition and standardize DB call, per bug 14989.
No time to fix, but leave a note about odd behavior of this extension for…
Use the new sequence name