A shout-out to SQLAlchemy

Posted on

At work-work we write Python apps, using SQLAlchemy as ORM and database layer. I tend to run our stuff on OpenSUSE with SQLite for quick development purposes, but production tends to be MS SQL Server and Windows XP (there are migrations ongoing ..). Today we got a bug report that certain columns in the database were being created as VARCHAR(max) (that’s a SQL-Serverism) and that this was leading to errors elsewhere in the application when reflection of the database was used to (re-)retrieve the types of columns.

So I spent a few hours testing and writing up a careful bug report with a reproducible test case. That’s the time my job is (most) fun, writing up something that contributes outside our own walls.

Turns out we were doing something that wasn’t quite right, but the SQLAlchemy developers agreed that it was a not-unreasonable thing to do and within six hours there was a fix, committed and scheduled for the next release. So here’s a shout-out to Mike B. at SQLAlchemy for his quick work. (And I’m glad the effort of making a good-as-I-can bugreport paid off.)