| Subject | Re: [Firebird-Architect] Re: RFC: Please unify stored procedure execution |
|---|---|
| Author | Jim Starkey |
| Post date | 2004-12-22T12:30:52Z |
Samofatov, Nickolay wrote:
I'd like to think that a half hour invested in a blr parsers (it's
straightforward Polish with two exceptions) would be a great deal faster
and probably safer. Since it would called only in the absence or
failure (type == unknown) of the flag, you would know that it
corresponded to blr known at fb2.0.
>Firebird ODBC driver makes the decision based on presence of blr_sendWhy blr_receive? Wouldn't it be a second blr_send?
>verb (correspoding to SUSPEND statement) in stored procedure BLR.
>To avoid parsing BLR directly it invokes pretty printer BLOB filter on
>SP BLR and looks for 'blr_receive' as unquoted word.
>This makes everybody using CALL syntax happy.
>
I'd like to think that a half hour invested in a blr parsers (it's
straightforward Polish with two exceptions) would be a great deal faster
and probably safer. Since it would called only in the absence or
failure (type == unknown) of the flag, you would know that it
corresponded to blr known at fb2.0.
>
>
>