mssql dialect omits limit/offset wrapping on second compile; alters the statement

[Former user]

In MSSQL When doing a query such as:

            qry = select([nemsis_model.Nemsis.table](nemsis_model.Nemsis.table), order_by=[nemsis_model.Nemsis.patient_care_report_number](nemsis_model.Nemsis.ems_agency_number,),limit = limit,offset=offset)

Evaluating the query twice causes it to drop the limit and offset on the second evaluation

An easy way to show what I suspect is happening is if I log the qry twice:

        log.info(qry)
        log.info(qry)

The second case will not display the constraints on record_number.

This took a while to track down as I was trying to do a select and not getting why the results of my select differed from the logging I did right before it.

[Mike Bayer]

fortunately a test case could be devised here easily:

from sqlalchemy import *

m = MetaData()
t = Table('t', m, Column('x', ))

s = select([t](t)).order_by(t.c.x).limit(5).offset(2)

from sqlalchemy.dialects import mssql

print s.compile(dialect=mssql.dialect())
print s.compile(dialect=mssql.dialect())

[Mike Bayer]

8340361bfa93dab71b18827fa0b7e7dc91b1d736 2e7506416166ab0b867259b54674057e45f447ea

[Mike Bayer]

Removing milestone: 0.7.9 (automated comment)