#Postgresql logo code

Moving forwardįrom Ghost 1.0 we're going to be officially throwing our weight behind MySQL.

For that reason, we are dropping official Postgres support from Ghost core. Without active community support, Postgres has always been, and always will be a second-class citizen. Knex's abstractions just aren't doing the job well enough.

#Postgresql logo code

Code that automatically works for MySQL & SQLite often needs special modifications for Postgres, meaning it is time consuming and hard to maintain. Postgres is a particular challenge because it is quite different to MySQL and SQLite. Bugs get raised by the community, but they don't get fixed. We miss things because we aren't using it. No one using Postgres is contributing to code or to QA. Over time we've reached a point where we no longer have any active core team members or regular contributors that use Postgres in production. We are able to guarantee first-class support for MySQL in a way we have never been able to for Postgres. Bugs and opportunities for improvement in those two databases are found and handled quickly. The Ghost team use Sqlite3 for development, and Ghost(Pro) uses MySQL - meaning we're running tens of thousands of blogs on MySQL all day every day. The logic was sound, but in practice this hasn't been true. There was a comment in the early thread advocating support about how having 'official support' earlier would prevent Postgres being a second-class citizen. Since then, it has become one of our biggest difficulties when developing new features, second only to our permissions system. At a certain point, it seemed natural to say we supported it. PostgreSQL support was added organically by several contributors who showed up in the early days, raised bugs, fixed bugs, got a build added to travis, and did QA to ensure that Postgres worked. Adding a long-term support strategy for Ghost v0.11, and focusing on an all new, backwards-incompatible Ghost v1.0 has given us an exciting opportunity to fix some our biggest development pain points.