[TYPO3-english] Oracle & Typo3 + TV

[Xavier Perseguers]

Hi,

> Most of the time the person/company who encounters the problem also has 
> to meet a deadline. In the example of Oracle + TV, it means that the 
> problem has to be solved before the normal site building can be started 
> (otherwise you have to migrate the database also). This means that there 
> are only a couple of days available (already extra time in the project) 
> to solve the problem or alternative ways to build the site have to be 
> found.

Exactly! In my case, I had the *requirement* to run a CMS on Oracle. 
MySQL was simply not an option. I was already glad to be able to run it 
on a Linux system because the other possible requirement would not have 
been Windows + Apache but Windows + IIS...

I knew that other DBMS were supported, I read the TYPO3 wiki explaining 
how to make it work and... that's it. I did not had extra time to make a 
proof-of-concept or the rest but I initially took more than it seemed to 
be needed as time of reserve in the budget in order to accomodate those 
"little extra" when doing something that is not out-of-the-box. I had to 
  invest some more personal time to make it work but in the end I did it.

Now, the sad story is that it takes so long to get *all* my patches into 
the Core.

> Even small patches can take a long time before they are accepted. In the 
> mean time you cannot update the extension without applying your own 
> patch(es) to that new version.

This is why patches should be taken into consideration and bug fixes 
have a higher priority than new cool features, even if I know that me 
too would like to get some new feature into Core.

Having a patched system for a while or on a dev environment is OK but it 
is really a pain to have to upgrade to a new version (not "major" but 
minor) and have to still apply a few patches.

>> Third, no core devs except Xavier is interested in Oracle by looking 
>> at the number of reminders for Oracle support in DBAL,
>> who is going to test TV against oracle???
> 
> Most people won't even try using Oracle after seeing the long list of 
> problems. DBAL is a nice idea, but it isn't really a DataBase 
> Abstraction Layer (it's more like a DataBase Compatibility Patching Layer).
> Trying to be compatible with a whole range of DBMSs on query level means 
> that each DBMS is used in a suboptimal way. Hopefully this is a thing of 
> the past with TYPO3 v5, but for the near future I can't really recommend 
> anything other than MySQL for TYPO3.

I would not be so strict but for sure I would not cry a big Hourra if 
someone is willing to use something else than MySQL.

> It is however pretty odd that TYPO3 officially supports a range of 

This is the main problem. Once all my patches are in, 4.3 will be 
(again) compatible (at least to some working level) with Oracle but 4.2 
won't although some previous versions were.

But again, I could propose TYPO3 because it was said to be compatible 
with other DBMS including Oracle which, again, was a *must have*.

> DBMSs, but core devs apparently don't have these available for the 
> necessary positive votes from a core dev?

This is sadly true. But remember, even if I was quite sure of that, I 
did patches and submitted them. They were nearly all accepted and those 
who were not will certainly be in in a while. Sometimes it is not 
possible to actually test those patches if you don't have a system at 
hand but /testing by reading/ *is* possible.

I'm always astonished to see how often developers *run* their code to 
see what comes out, even after a minor modification...

-- 
Xavier Perseguers
DBAL Member
MVC ExtJS Samples Leader

http://forge.typo3.org/projects/show/extension-mvc_extjs_samples