"Gokulakannan Somasundaram" <[EMAIL PROTECTED]> writes:
> On 11/2/07, Tom Lane <[EMAIL PROTECTED]> wrote:
>> [ splorfff... ] The grammar support alone will cost ten times that.
> But i guess(I may be wrong), you may be wrong about the grammar support
> part.
Well, a crude estimate is that SELECT ... AS OF Savepoint would take
about the same amount of grammar and nodetree infrastructure as SELECT
... FOR UPDATE does. Try grepping for all the code that deals with
LockingClause (the raw-grammar representation of FOR UPDATE) and
RowMarkClause (the transformed representation). Hits on the list fields
SelectStmt.lockingClause and Query.rowMarks would be relevant too
(code touching these lists wouldn't necessarily mention the list member
node type anyplace).
1000 lines may be an overestimate, but it's probably nearer the mark
than 100 is.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?
http://archives.postgresql.org
