Subject | Re: Standard Conformance |
---|---|
Author | peter_jacobi.rm |
Post date | 2003-06-19T12:19:19Z |
Hi Jim, all,
--- In Firebird-Architect@yahoogroups.com, Jim Starkey wrote:
> So ask away, but unless you willing to do it yourself,
> don't expect much response.
Yes, I know how Open Source works and also just subscribed
to devel list. But I'm unclear whether I'll manage to take
a deep enough dive into the source to contribute anything
but easily factorable items like collation stuff.
What was mainly in my mind with this thread, is the issue of
'needless' non-conformance. Points which may be easy to
change for the core contributors and where there is no
technical reason against conformance.
In contrast to those points, I fully understand that you
are not eager to dump the work done in triggers or security,
for a worse but conformant implementation.
As further examples for maybe-but-not-quite 'needless'
non-conformance, look at two 'Firebird-isms' unearthed
in thread:
http://groups.yahoo.com/group/firebird-support/message/25886?threaded=1
1. CHECK constraints implicitely forbid NULL when not
overriden (SQL standard says the other way around, if I'm
not mistaken).
=> maybe very easy to change, but would give unpleasant surprises
with existing code assuming current Firebird semantics.
2. referential integrity constraint names should be, but are not,
local to the table
=> this one can be implemented without disturbing current code,
but it may be a more difficult source code surgery.
Best Regards,
Peter Jacobi