[wp-hackers] Maintenance release, first gate of "is the issue major?"

[Peter Westwood]

On Wed, October 4, 2006 8:22 pm, Lloyd D Budd wrote:
> This bug hunt is an awesome success and it is not over yet. While
> going to sleep last night, I rememberd a lesson learned in similar
> contexts. We probably should have been more conservative on what made
> it into 2.0.5 .
> http://trac.wordpress.org/query?milestone=2.0.5
>

Yes and no.  Most of those changes are very small.

> Each of the fixes seem to stand on their own, but quantity of change
> increases the risk of fault.
>

I disagree with this.  At work we have always striven for Maintenance
releases to contain only low risk changes - i.e ones where we can see the
bug and test the solution easily - rather than to keep maintenance
releases small.

> In the future, particularly at this "late" maintenance release and
> where we are in 2.1 development cycle, lets consider the first gate to
> be "is the issue major?"
>

In the future with the proposed 120day model we wouldn't need this
maintenace release as we would already have released 2.1 months ago.

> This is particularly relevant if you are interested in others such as
> Debian distributing WordPress.
>
> I don't want to have a 2.0.6. I want us focus on releasing 2.1.
>

Agreed - In general though maintenance releases for WordPress are
triggered by a security issue rather than anything else.

I think we also need to be aware of the fact that in general our end-users
don't always upgrade straight away and we need to understand why:

It is because maintenace releases are difficult to upgrade to - Hopefully
this is simplyfied by Mark, myself and others who provide diff files and
instructions for generating them.

It is because maintenace releases do not fix the issues that are most
important to people - or don't contain enough fixes to be important - or
we do not publish enough easily digestible information about the fixes to
make people keen to upgrade - linking to the trac milestone for a release
in the dev blog post probably doesn't help most of our users.

It is because people are just too lazy - I can see from the small sample
of users who have installed my Verson Check plugin that there are people
out there still running 1.5.2/2.0.1/2.0.2/2.0.3 even with a big flashing
warning  on every admin page!

westi
-- 
Peter Westwood <peter.westwood at ftwr.co.uk>
http://blog.ftwr.co.uk