[Typo3-dev] DBAL

[Martin T. Kutschker]

Karsten Dambekalns wrote:
> 
> What we can see here, is that switching to text fields instead of blob
> fields will introduce the need to add the handling of the different
> behaviours of case sensitivity to the abstraction layer.

What we also see is that if you *need* case-insensitivity you cannot 
assume you have it. Quite the contrary, most DBs are case-sensitive, so 
we need an abstraction layer for this anyway. Every "simple" search that 
is now case-insensitive on Mysql, won't be on any other RDBMS.

 > Whether this
> is worth the hassle could only be decided by looking at how this might
> speed up things - I could image that lookups on text field *might* be
> faster than on blob fields (depends on how the DB handles indexing on
> different field types).

No real problem with BLOBs. I just wondered what the reason for them 
was. No I've been told, so it's fne with me.

Masi

PS: While we're at it. I remember Oracle has two types of [VAR]CHARS. 
The "standard" ones and the N* (or NS*) variants. I recollect that the 
former where 7bit ASCII (?) and the latter did some language/charset 
thingies. Don't forget you need the charset/language support in the DB 
to have case-insensitivity. Hehe. Makes me wonder what the minimum 
version number of Mysql and the other RDBMSs will be after all these 
issues have been addressed.

PPS: I love the Turkish I! small i with dot => capital I *with* dot, 
small  i without dot => capital I *with* dot. So case is language 
dependent...