[TYPO3-core] Regression in 4.5.23

[Steffen Müller]

Hi.

Let me ask some questions here.

On 22.02.2013 11:02 Ernesto Baschny wrote:
> 
> This is a no go in my eyes. The 4.5 backport was clearly not a
> "no-brainer" backport at all and should have passed the normal reviewing
> and voting system (two core devs, two testers).
> 

How to validate if something is a no-brainer?
Who's responsible for that?
How can this be supported technically?


> Especially in 4.5 LTS people *expect* things not to break and if
> something in the session handling isn't working since start of 4.5
> (2011), people have probably made some work around the problem in the
> meantime anyway. Fixing it just for the sake of having it fixed with the
> huge risk of a regression should not happen at this state of the 4.5
> life-cycle, where we prefer stability (and security) opposed to
> backporting every tiny fix. "Less change is better".
> 

This exceptional definition of LTS is new to me. AFAIK it neither has
been postulated nor has it lead to a more strict backporting policy in
gerrit review progress.

So far we had the policy: No new features in releases, but backported
bugfixes.

How can we implement this in gerrit?

I guess announcing this explicitly in a separate thread would help.

> 
> Instead of concentrating on backporting all stuff to 4.5, let's work on
> having master (6.1) stable and bug free in order to have a real stable
> next LTS (6.2)!

True, but backporting bugfixes to actual stable LTS is also neccessary.

Who will decide which bugfix is worth LTS backport?
Which are the criteria for this decision?


-- 
cheers,
Steffen

TYPO3 Blog: http://www.t3node.com/
Twitter: @t3node - http://twitter.com/t3node