[TYPO3-dev] Are performance improvements part of maintaince releases?

[Lars Houmark]

Hi everyone,

First of all, I think rupi have not only documented the fix in mention 
and how little the risk is. At the side of that he has also documented, 
that the "rest of the universe", with far more mission critical 
applications, and also projects that we either look up to or compare 
our project with, is in fact doing loads of performance improvements in 
patch-level releases.

So with all this good feedback, could we try to get a consensus on 
performance improvements in patch-level releases?

Here is my take, which I what I understand from the general reply here 
- though some had a different angle. Feel free to improve on the below 
text. I hope the release managers and the core team leader will take 
this up, decide on it, and write into the necessary documents in order 
for it to take effect.

--- start of proposal ---

Performance improvements is normally not considered valid for 
patch-level releases, unless the following is generally the case:

- The performance improvement is dramatic and not just 10-12%.
- The bugfix is simple and not complicated (* draft note, could be elaborated)
- The RFC sender will document that the changes is not breaking 
functionality or behavior
  - In order to prove that, the following documentation is accepted as valid:
     - Unit tests attached to the RFC
     - Written documentation of the old and new behavior related to the change

The documentation needs to be accepted and voted for along with the 
patch itself, after the same rules of normal voting.

As with any other bug, the release manager and the core team leader has 
the possible of vetoing the addition of a performance bugfix to a 
patch-level release.

--- end of proposal ---

Thanks for any comments :)

-- 
Lars Houmark
Member of the TYPO3 Security Team