Subject | Re: Questions about migration from IB to Firebird |
---|---|
Author | Alexander V.Nevsky |
Post date | 2004-02-09T19:20:56Z |
--- In firebird-support@yahoogroups.com, Olivier Mascia <om@t...>
wrote:
relict - IB6.0x - and forgot about 4 points (first 3 are not related
to client but can require to polish own code):
1. Improved in FB1 dialect 3 ambiguity check - usage in one query
columns with the same name from different tables raises syntax
exception, IB6 accept such a query but can return wrong result.
2. IB6 allows to make in PSQL Char+Char and raises exception in
runtime if this chars can't be interpreted as numerics. FB1 in dialect
3 rejects this on syntax level and strongly requires concatenation.
3. Improved check of size of concatenation of long chars/varchars. FB1
don't allow to step over 64K limit checking declared types sizes.
4. Several bugs were fixed not only in engine but in gds32 too in
FB1. I can recall exactly only bug with execution several SPs with
different amount of parameters in one transaction, but there were
more. This is argument not for necessity but for desirability to
change client ASAP.
Anyway, I intentionally recommend to test pair database-application
after migration before deploing solution to production. I did'nt
touched IB6 for years too so can forget something too ;)
Best Regards,
Alexander.
wrote:
> On 9/02/2004 17:00 GMT+1, develop111 wrote:migration
>
> > 1- Can I access a Firebird server from a client computer that have
> > IB 6 client files installed, or I also have to instal Firebird
> > client on this computer ? (What I'm looking for is an easy
> > to Firebird).Guys, seems all of you for a long period of time did'nt touched this
>
> FB 1.0x : go ahead.
relict - IB6.0x - and forgot about 4 points (first 3 are not related
to client but can require to polish own code):
1. Improved in FB1 dialect 3 ambiguity check - usage in one query
columns with the same name from different tables raises syntax
exception, IB6 accept such a query but can return wrong result.
2. IB6 allows to make in PSQL Char+Char and raises exception in
runtime if this chars can't be interpreted as numerics. FB1 in dialect
3 rejects this on syntax level and strongly requires concatenation.
3. Improved check of size of concatenation of long chars/varchars. FB1
don't allow to step over 64K limit checking declared types sizes.
4. Several bugs were fixed not only in engine but in gds32 too in
FB1. I can recall exactly only bug with execution several SPs with
different amount of parameters in one transaction, but there were
more. This is argument not for necessity but for desirability to
change client ASAP.
Anyway, I intentionally recommend to test pair database-application
after migration before deploing solution to production. I did'nt
touched IB6 for years too so can forget something too ;)
Best Regards,
Alexander.