Tom Lane wrote:
> Ron Mayer <rm_pg@cheapcomplexdevices.com> writes:
>> Once this settles I suppose I should post a ECPG patch that's based
>> off of these Decode/Encode interval functions too?
>
> Yeah, if you want. I think you'll find that the datetime code has
> drifted far enough since ecpg forked it that you'll be looking at a
> pretty huge diff :-(
Merging of the interval style into ecpg attached.
I blindly copy&pasted code from src/backend/utils/adt/datetime.c
into src/interfaces/ecpg/pgtypeslib/interval.c and made the minimal
changes (pg_tm -> tm; adding constants; etc) to make the regression
tests pass; and mentioned that in the comments.
I know little enough about ecpg that I can't really tell if these changes
are for the better or worse.
One thing in the patch that's probably a bug is that the
constants in src/include/utils/dt.h and src/include/utils/datetime.h
under the section "Fields for time decoding" seem not to match, so
when I copied some constants from datetime.h to dt.h to make it compile
the values I copied are probably wrong. Unfortunately I know little
about ecpg or the history of dt.h to know what the right values should
be.