Alvaro Herrera wrote:
> I messed with that code some more, as it looked unnecessarily
> complicated; please see attached and verify that it still behaves
> sanely. This needs those regression tests you promised. I tested a few
> cases and it seems good to me.
I've fixed a couple things over v16:
- avoid passing every cell through psprintf, which happened due
to cont.cells being pre-initialized to empty strings.
- adjusted the loop freeing allocated_cells
and added the regression tests.
Attached is the diff over v16, tested with make check and valgrind.
Best regards,
--
Daniel Vérité
PostgreSQL-powered mailer: http://www.manitou-mail.org
Twitter: @DanielVerite