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

[georg kuehnberger]

hola, Rupert, typo3-devs,

Rupert Germann wrote:
> hi,
> again:
> Are performance improvements part of maintaince releases?

they are,

> or in other words: 
> is a function that makes 50 identical sql requests when rendering a page a bug?

it is.

- funny things happened within the thread on the core list, Rupert 
mentioned; - Based on an RFC, a commit and revert, arguing started on 
various levels (AFAI understood),
though the IMHO the point was missed, which I feel is odd.

Lets face reality:

1.) Following the meaningful principle of stable and relible software, 
each single release (patch/minor/major), will efectively introduce more 
/ different code & features, which is great! However "more code & 
features" wont improove performance, but to the contrary. Anybody 
running TYPO3 since version 3.5 knows what I'm referring to;

2.) Performance IS an Essential Function of the TYPO3-core!
Please consider it as such!

3.) Thus,
- Performance-Problems of the core ARE: (guess what): Problems.
- Problems are BUGs.
- Problem-fixes, thus are Bugfixes;
eg:
- the reduction by +50 avoidably db-calls IS a BUG-Fix - no doubt.
- a patch provided, which avoids "useless" db-calls and still guarantees 
the prooven functionality, is a bugfix.

- This IS completely in-line with: 
http://typo3.org/teams/core/resources/release-workflow/
     "The Core team is permanently fixing problems in the development 
branch of this version."

- Let me re-state: "performance-bottlenecks ARE PROBLEMs",
and thus they should be solved in each new release, and backported in 
case appropriate.

4.) @typo3-core-team:
- Please consider the avoidable consumption of ressources as valid & 
patchable bugs, and thus
- add "Performance" AS a Core-Function(al requirement) to 
http://typo3.org/teams/core/resources/release-workflow/ .

 > let the flamewar begin
no need;

 > PS: at the end of this discussion I'd like to see a document which 
clarifies which improvements are bugs and which are not.

Hi need! as stated above & below;

just my 2 pesos,
regards Georg
--- not being a core-dev, though simply stating the obvious, that 
performance IS a funtion; the opposite is a bug and thus has to be fixed.

PS: XCLASS-ing core-parts for sure is no the best practice; however 
sometimes unavoidable, eg. performance-features not going into the core.
The call for a removal of those XCLASS-ing extensions from the TER, is 
no solution eigther; they provide quite some insight on TYPO3-core 
shortcomming; better learn from them, integrate the ideas into the core 
and then consentually "discard" them.

PPS: We repeatedly had the pleasure to debug T3-sites doing 1.000-10.000 
single db-calls just to generate one page, where roughly 50% were 
redundant function and db-calls (again min. 50% caused by 
T3-Core-Functions), which IS bad as we know.), so lets not hinder 
improovement.