Permalink
Commits on Nov 28, 2018
  1. Merge pull request #443 from mozilla/pb/lints

    philbooth committed Nov 28, 2018
    #443
    r=vbudhram
  2. feat(scripts): lint-ignore tables that already have foreign keys

    philbooth committed Nov 28, 2018
Commits on Nov 7, 2018
  1. Merge pull request #419 from mozilla/pb/411

    philbooth committed Nov 7, 2018
    #419
    r=vladikoff
  2. chore(scripts): lint-ignore consumeRecoveryCode_2 and setPrimaryEmail_3

    philbooth committed Oct 29, 2018
    These two scripts use ROW_COUNT() but I couldn't see an obvious way to
    work round them. In the case of consumeRecoveryCode_2, ROW_COUNT() is
    used as an interim signal for aborting the remainder of the procedure.
    In the case of setPrimaryEmail_3, the UPDATE is in a transaction which
    causes affectedRows to be zero upon returning.
  3. feat(scripts): add ROW_COUNT() checks to the procedure-linting script

    philbooth committed Oct 29, 2018
    In https://bugzilla.mozilla.org/show_bug.cgi?id=1499819, it was found
    that `ROW_COUNT()` causes db replication to lag. As such, it makes sense
    to grep for `ROW_COUNT()` in our stored procedure lint script.
    
    The script was called `explain-warn.js` before, but in light of the new
    functionality it's renamed here to the more generic `procedure-lint.js`.
Commits on Nov 2, 2018
  1. Merge pull request #431 from mozilla/pb/393

    philbooth committed Nov 2, 2018
    #431
    r=shane-tomlinson
  2. Merge train-124 into master

    philbooth committed Nov 2, 2018
    #430
    r=shane-tomlinson
  3. chore(db): use mariadb-friendly drop index syntax

    philbooth committed Nov 2, 2018
    Sometimes we get self-hosters or contributors showing up who want to use
    MariaDB. Even though the official line is that we don't support it, it
    happens often enough that we may as well just fix the places where we're
    incompatible. Afaik, that is just the `DROP INDEX` syntax in a couple of
    places.
    
    Note there are also occurrences in the reverse migrations, which I
    haven't bothered to fix. Seeing as they're not invoked in normal running
    and need to be manually uncommented before running anyway, it seems
    reasonable for them to be manually edited if anyone needs to do so.
  4. Release v1.124.1

    philbooth committed Nov 2, 2018
  5. Merge pull request #429 from mozilla/pb/deps

    philbooth committed Nov 2, 2018
    #429
    r=shane-tomlinson
  6. fix(package): update deps

    philbooth committed Nov 2, 2018
Commits on Nov 1, 2018
  1. Merge pull request #425 from mozilla/pb/fix-explain-script

    philbooth committed Nov 1, 2018
    #425
    r=shane-tomlinson
Commits on Oct 30, 2018
  1. fix(scripts): stop the explain script tripping over git grep colours

    philbooth committed Oct 30, 2018
    The explain script was using `cut` to split paths from git grep before
    the colon character. This fails on system where control characters such
    as colours are emitted before the colon.
    
    A more robust approach is to use a regex to pull out the specific text
    we're looking for. The script is already hard-coded with other paths
    like `lib/db/mysql.js`, so matching against the expected migration path
    doesn't seem like too much of a stretch to me.
  2. Merge train-124 into master

    philbooth committed Oct 30, 2018
    #426
    r=shane-tomlinson
  3. Release v1.124.0

    philbooth committed Oct 30, 2018
Commits on Oct 16, 2018
  1. Merge pull request #408 from mozilla/pb/test-fetch-bounce-order

    philbooth committed Oct 16, 2018
    #408
    r=vladikoff
  2. fix(mem): ensure emailBounces are stored most-recent first

    philbooth committed Oct 16, 2018
Commits on Sep 17, 2018
  1. Merge pull request #392 from mozilla/pb/explain-ci-wip

    philbooth committed Sep 17, 2018
    #392
    r=rfk
  2. chore(scripts): disable the explain script in production

    philbooth committed Sep 17, 2018
  3. feat(scripts): add an ignore file for the explain script

    philbooth committed Sep 17, 2018
    Some of our existing stored procedures fail in the explain script but we
    don't want those failures to show up. In the future we may have reasons
    for wanting to ignore other procedures too. This change lets us ignore
    them while still running the script in CI.
  4. chore(scripts): tweak some old migrations to fix explain errors

    philbooth committed Sep 17, 2018
    Some of our old migrations don't conform to assumptions made by the
    explain script. This change just tweaks some of the argument names so
    that they stop showing up as errors in that script. There are no logical
    changes to the stored procedures.
  5. feat(scripts): add script to automate MySQL EXPLAIN checks

    philbooth committed Aug 24, 2018
    Adds a script that pulls out SELECT queries from our stored procedures
    and then runs EXPLAINs for them in MySQL. Warnings are emitted for any
    bad smells that are found in the EXPLAIN results, e.g. `filesort`.
Commits on Aug 28, 2018
Commits on Aug 21, 2018
  1. Merge pull request #386 from mozilla/pb/memdb-device-commands

    philbooth committed Aug 21, 2018
    #386
    r=vladikoff
  2. chore(db): ensure mem db behaves like mysql db

    philbooth committed Aug 21, 2018
    In 06554f5, we temporarily disabled
    insertion of device commands into the MySQL db. We didn't do the same
    with the memory db though, which means the auth server tests see
    different behaviour depending on which backend they're running against.
    This change forces the memory backend to behave the same.
Commits on Aug 19, 2018
  1. Merge train-118 into master

    philbooth committed Aug 19, 2018
Commits on Aug 18, 2018
  1. Release v1.118.1

    philbooth committed Aug 18, 2018
  2. chore(db): stop calling the upsertAvailableCommands procedure

    philbooth committed Aug 18, 2018
    #383
    r=jbuck,eoger,vladikoff
Commits on Jul 20, 2018
  1. fix(tests): move local utils tests so they get run by npm t (#377) r=@…

    philbooth authored and vladikoff committed Jul 20, 2018
    …vladikoff
    
    Fixes #376.
    
    The tests in test/lib/utils.js weren't being exercised by npm t. Moving them into test/local fixes that.
    
    Because they weren't being exercised, one of the tests had got into a failing state. Looking at it, I think it's a redundant test now because there is no keyspace argument to generateRecoveryCodes. Hence it's deleted here.
Commits on Jul 11, 2018
  1. Release v1.116.0

    philbooth committed Jul 11, 2018