Re: pg_dump sometimes misses sequence parameters

Alexey Bashtanov <bashtanov(at)imap(dot)cc> writes:
> The bug affects REL_10_STABLE and master branches.
> 9.4..9.6 are unaffected.

> To reproduce:
> psql -c 'DROP SEQUENCE IF EXISTS foo'
> psql -c 'CREATE SEQUENCE foo INCREMENT BY -1 MINVALUE
> -9223372036854775808 MAXVALUE 9223372036854775807'
> pg_dump -t foo > tmp
> psql -c 'DROP SEQUENCE foo'
> psql <tmp

> The last psql call fails with "START value (9223372036854775807) cannot
> be greater than MAXVALUE (-1)",
> as pg_dump does not record MAXVALUE properly.

Ugh.

> The reason is atoi is used and those large numbers are interpreted as 0
> and -1 respectively.
> I'd propose to use atol instead of atoi, please see the patch attached.

That will only fix it on platforms where long is 64 bits. I think that
converting these strings to integer at all is a dumb idea. It would be
much better to write the first two tests like

if (is_ascending && strcmp(minv, "1") == 0)
minv = NULL;
if (!is_ascending && strcmp(maxv, "-1") == 0)
maxv = NULL;

It's tempting to try to make the remaining tests look similar, but
I'm not quite sure how without writing out the exact decimal values
of the constants, which probably isn't better. I experimented with
stuff like "strcmp(minv, CppAsString2(PG_INT32_MIN))" but that
doesn't quite do what we need.

regards, tom lane