Question about an inconsistency - 1

"petrum@gmail.com" <petrum@gmail.com> writes:

I have a question regarding the source code in file pg_proc.h (postgresql-9.4.4).

At line 1224 (copied below) why the 28th identifier is timestamp_eq?

DATA(insert OID = 1152 ( timestamptz_eq PGNSP PGUID 12 1 0 0 0 f f f t t f i 2 0 16 "1184 1184" _null_ _null_ _null_ _null_ timestamp_eq _null_ _null_ _null_ ));

I would expect it to be timestamptz_eq (the same as the 5th identifier
from the same line).

If timestamptz_eq actually existed as a separate C function, then yes that
would be correct. But see this bit in timestamp.h:

extern int timestamp_cmp_internal(Timestamp dt1, Timestamp dt2);

/* timestamp comparison works for timestamptz also */
#define timestamptz_cmp_internal(dt1,dt2) timestamp_cmp_internal(dt1, dt2)

AFAICS there are not similar #defines equating timestamptz_eq to
timestamp_eq and so on, probably because we don't have much need
to refer to those functions in the C code. But even if we had
such #defines, I think that pg_proc.h could not rely on them.
It has to reference actual C functions.

regards, tom lane

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Hi,

In file postgresql-9.4.4/src/backend/utils/adt/format_type.c
function format_type_internal line 159, shouldn’t be used
array_base_type instead of type_oid?

In line 153 it is searched for array_base_type and thus
shouldn’t we use it (ie., array_base_type) to report that the
type is not found. It is similar to the case of searching for
type_oid in lines 129-135.

Thanks,
Petru Florin Mihancea

Hi,

In file postgres/postgresql-9.4.4/src/timezone/zic.c
function stringzone line 2091we have

if (stringrule(result, stdrp, dstrp->r_stdoff, zp->z_gmtoff) != 0)

Is it ok to have as the 3rd argument dstrp->r_stdoff or should we
have stdrp->r_stdoff? In line 2085 dstrp is used in both arguments.

Not very sure because I do not understand the semantics of the
invoked operation, but I thought it would be better to ask.

Thanks,
Petru Florin Mihancea

"petrum@gmail.com" <petrum@gmail.com> writes:

In file postgresql-9.4.4/src/backend/utils/adt/format_type.c
function format_type_internal line 159, shouldn’t be used
array_base_type instead of type_oid?

IIRC, that was intentional. Supposing there's a pg_type row with
a corrupted typelem value, it's more useful to point at the bogus
row than to report the junk value with no context. I suppose you
could argue that it should be more like

elog(ERROR, "cache lookup failed for type %u (typelem of %u)",
array_base_type, type_oid);

regards, tom lane

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

"petrum@gmail.com" <petrum@gmail.com> writes:

In file postgres/postgresql-9.4.4/src/timezone/zic.c
function stringzone line 2091we have

if (stringrule(result, stdrp, dstrp->r_stdoff, zp->z_gmtoff) != 0)

Is it ok to have as the 3rd argument dstrp->r_stdoff or should we
have stdrp->r_stdoff? In line 2085 dstrp is used in both arguments.

Doubt it, because if the coding were like that there would be no need
to pass the third argument separately at all; the subroutine could just
access rp->r_stdoff for itself. But this isn't our code, so if you
want an authoritative answer you should ask at
https://mm.icann.org/mailman/listinfo/tz

regards, tom lane

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers