[Typo3-dev] Versioning Fix

[Peter Kühn]

Hi yall,

at first: thnx so much for the fantastic time spend with you in 
kitzbühel!!!

to the versioning prob:
imho it might be necessary to simply differentiate between the 
(theoretical) discussion, whether an auto_increment should be able to 
become negative or not and the situation outside: there are thousands of 
typo-installations runing on mysql < 4.0.18.
actually this was the main reason why mathes and i wrote this patch - we 
needed the versioning for a project that is live at one of our customers 
and the admin doesnt want to update mysql.

i think that the versioning feature is absolutely great - a huge step 
forewards and a handle to make typo even more popular if it would work 
out of the box everywhere and this means from my point of view 
without(!) the need to change the setup that serves well for typo in 
generall.

anyway: the second reason to write the patch was that stucki told me, he 
wants this issue to be solved in 3.7.1 so we just wanted to share a 
suggestion as RFC.

have a nice day
pekue



Kasper Skårhøj schrieb:
> Hi Mattes,
> 
> 
>>I mean - please don't feel offended, but using SQL that works fine on 
>>some versions of one (!) DB somewhat messes up my understanding of high 
>>quality code.
>>
> 
> 
> When a field is a SIGNED integer, why can I not make its value negative? All IDs are at all times unique and it is therefore clearly a bug that the auto_increment feature in MySQL messes up the next-index value. This is THE BUG they fixed in later versions.
> 
> So your argument is that high quality code should work on buggy implementations? At best you can claim that a requirement of MySQL 4.0.18+ is maybe a bit strict, nevertheless only required for versioning.
> 
> Why it doesn't work on Oracle etc. I don't know.
> 
> - kasper
> 
>