I wrote:

H> | Leon,
H> |
| L>> 1.is it a kind of record locking (if your where clause collects only 1
| L>> record ) ?
H> |
H> | actually it locks all the records retrieved (if it can) as it fetches
H> | them; usually you'll want to use it on a single-record select, but that's
H> not
H> | mandatory.
H> |
| L>> 2.does it lock all fields in the where clause or complete records
H> |
H> | it locks the record; Firebird (nor any other RDBMS that I know) cannot
H> | lock single columns.
H> |
| L>> 2.how long are records locked ? until post or until commit ?
H> |
H> | I think locks are released upon transaction end (commit or rollback).
H> |
| L>> 3.is it an alternative for thoe old recordlocking ?
H> |
H> | If you mean "dummy updates", then yes, with the additional benefits of
H> | greater efficiency and update triggers not firing twice.
H> |
H> | The behaviour of this feature varies with the transaction isolation
H> | level and flags (like wait/no_wait) you use; see the release notes and
H> | the archives of the firebird-devel list to know more.

Hans replied:

H> For now in IB_Parse.pas, comenting the line SuffStrings out, seems to work
H> for me as a temp patch

Wrong thread?

Ciao
--
Nando mailto:nandod@...