FreeTDS Development Group

["Bill Thompson" <thompbil AT exchange.uk.ml.com>]

All,

Thought I would consult you before doing anything that you may disapprove
of...

basically I need to modify tds_process_row_tokens() such that information
about any "compute" rows is passed back to any calling function.
In my particular case, I am trying to amend dbnextrow() to cater for these
rows. dbnextrow() needs in turn to return either:

REG_ROW      (a normal data row has been returned)
NO_MORE_ROWS (there ain't no more data...)
FAIL
BUF_FULL
a "compute id" (to tie back this compute data to its "result" information)

as dbnextrow() is a thinnish wrapper around tds_process_row_tokens(), the
latter needs to pass similar information back to the former.

As I see it I could either:

keep the existing returns from tds_process_row_tokens() (SUCCEED or FAIL
essentially), and add extra output parameters for returning extra
information.

or:

amend the return codes from tds_process_row_tokens() to roughly match the
dbnextrow() returns.

In either case, I'm going to be amending the behaviour of a function which
may well be used natively elsewhere (tsql for example - also other people
are probably accessing the tds layer directly, albeit against our better
judgement). I'm not happy about doing this.

Any advice ?

Bill