Doctrine 1

All but the first migrations fail with PostgreSQL [patch included]

Details

  • Type: Bug Bug
  • Status: Closed Closed
  • Priority: Blocker Blocker
  • Resolution: Fixed
  • Affects Version/s: 1.2.1
  • Fix Version/s: 1.2.2
  • Component/s: Migrations
  • Labels:
    None

Description

The problem is in this method:

protected function _createMigrationTable()
{
    if ($this->_migrationTableCreated) {
        return true;
    }

    $this->_migrationTableCreated = true;

    try {
        $this->_connection->export->createTable($this->_migrationTableName, array('version' => array('type' => 'integer', 'size' => 11)));

        return true;
    } catch(Exception $e) {
        return false;
    }
}

When migration_version table doesn't exist, everything works like it should. But subsequent migrations fail, because after failing at creating migration_version table every subsequent queries in that transaction fail.

It works flawlessly with MySQL, but fails with PostgreSQL.

  1. File
    patch-migrations.diff
    17/Feb/10 2:17 PM
    0.9 kB
    Elnur Abdurrakhimov
  2. File
    patch-migration-pgsql.diff
    20/Feb/10 4:36 AM
    1 kB
    Elnur Abdurrakhimov
  3. File
    patch-doctrine-migration-pgsql-2.diff
    15/Mar/10 12:07 PM
    0.6 kB
    Elnur Abdurrakhimov

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Elnur Abdurrakhimov added a comment -

My previous patch broke MySQL migrations. This new patch resolves that problem.

My solution is not neat, but at least migrations work and I can continue developing my project.

Show
Elnur Abdurrakhimov added a comment - My previous patch broke MySQL migrations. This new patch resolves that problem. My solution is not neat, but at least migrations work and I can continue developing my project.
Hide
Permalink
Jonathan H. Wage added a comment -

This isn't really a good patch. I don't wanna have driver specific conditional statements like that.

What about if we started and committed our own transaction for the create table so that everything else is in another transaction and would not conflict.

Show
Jonathan H. Wage added a comment - This isn't really a good patch. I don't wanna have driver specific conditional statements like that. What about if we started and committed our own transaction for the create table so that everything else is in another transaction and would not conflict.
Hide
Permalink
Elnur Abdurrakhimov added a comment -

This time solution is neater.

Show
Elnur Abdurrakhimov added a comment - This time solution is neater.
Hide
Permalink
Elnur Abdurrakhimov added a comment -

Provided a patch with a neater solution.

Show
Elnur Abdurrakhimov added a comment - Provided a patch with a neater solution.
Hide
Permalink
Jonathan H. Wage added a comment -

Thanks for following up on this and fixing the root cause of the problem.

Show
Jonathan H. Wage added a comment - Thanks for following up on this and fixing the root cause of the problem.

People

  • Assignee:
    Jonathan H. Wage
    Reporter:
    Elnur Abdurrakhimov
Vote (0)
Watch (1)

Dates

  • Created:
    Updated:
    Resolved: