mahara-contributors team mailing list archive

[Lester Caine <lester@xxxxxxxxxxx>]

I now have a fairly completely port of the mahara code working on Firebird. Attached are the modified files, but they have not been tested on Postgres/MySQL as yet ( I don't have any machines with them on :) )
Further tidying up will be required, but in theory the Oracle and MSSQL should now also not be fore off functional!

Things tidied while porting ...

LIMIT removed and the ADOdb limit function enabled.

{} complimented with `` wrapping for field names

In a few places the timestamp processing had the table alias outside the
function, which would not work for MySQL and I suspect Postgres.

ADOdb functions need to be used for sequence management. This will tidy
up additional blocks of dodge code since the code already exists.


** Attachment added: "Mahara codebase modified to use backtick for field esaping"
   http://launchpadlibrarian.net/47652954/Mahara_Firebird.tar.gz

-- 
Firebird support development
https://bugs.launchpad.net/bugs/572189
You received this bug notification because you are a member of Mahara
Contributors, which is subscribed to Mahara.

Status in Mahara ePortfolio: New

Bug description:
Attached are a pair of drivers to add Firebird as an option on the supported databases. Class goes to xmldb library, and the adodb driver replaces the one currently provided.

The next step on this exercise is to work out how to handle wrapping of field names with double quotes. The normal practice has been to use 'backtick' as a hook to identify both field and table names that need processing. This is then replaced or removed as required before the SQL is run. This would - I think - tidy up a problem I am seeing currently in the code base where some fields have had double quotes added in the scripts, or bodge case statements adjust things for MySQL or Postgres .... A generic fix would allow the availability of other databases to be expanded? And I am happy to look at handling that.