Re: Fix overflow in DecodeInterval

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Joseph Koshakow <koshy44@gmail.com>
Cc: Andres Freund <andres@anarazel.de>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-04-03T16:03:23Z
Lists: pgsql-hackers
I wrote:
> Joseph Koshakow <koshy44@gmail.com> writes:
>> I think I know that the issue is. It's with `ParseISO8601Number` and
>> the minutes field "1.".
>> Previously that function parsed the entire field into a single double,
>> so "1." would
>> be parsed into 1.0. Now we try to parse the integer and decimal parts
>> separately. So
>> we first parse "1" into 1 and then fail to "." into anything because
>> it's not a valid decimal.

> Interesting point, but then why doesn't it fail everywhere?

Oh ... a bit of testing says that strtod() on an empty string
succeeds (returning zero) on Linux, but fails with EINVAL on
AIX.  The latter is a lot less surprising than the former,
so we'd better cope.

(Reading POSIX with an eagle eye, it looks like both behaviors
are allowed per spec: this is why you have to check that endptr
was advanced to be sure everything is kosher.)

			regards, tom lane



Commits

  1. Fix portability issues in datetime parsing.

  2. Fix overflow hazards in interval input and output conversions.

  3. Add a couple more tests for interval input decoding.