| From: | Maxim Orlov <orlovmg(at)gmail(dot)com> |
|---|---|
| To: | Pavel Borisov <pashkin(dot)elfe(at)gmail(dot)com> |
| Cc: | Postgres hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com>, Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>, Ken Kato <katouknl(at)oss(dot)nttdata(dot)com>, Aleksander Alekseev <aleksander(at)timescale(dot)com> |
| Subject: | Re: Is monotonous xid8 is a right way to do? |
| Date: | 2022-04-01 12:36:28 |
| Message-ID: | CACG=ezZTvzyBscxN3xoWm14WXmg3C7sJeg3UKJVrfHYhSeHxrg@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi!
In my view, FullTransactionId type was implemented without considering 64
bit wraparound. Which seems to be unlikely to happen. Then on that basis
xid8 type was created. Details of that particular implementation
infiltrated into documentation and became sort of normal. In my opinion,
semantically, both of these types should be treated as similar
types although with different sizes. Thus, again, xid and xid8 types should
be a ring and have no min and max functions. At least, in a sort of
"conventional" way when minimal value is minimal in a mathematical way and
so for maximum.
For example, max may be implemented as max(0, 42, 18446744073709551615) =
42, which is a bit weird.
--
Best regards,
Maxim Orlov.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dagfinn Ilmari Mannsåker | 2022-04-01 12:39:07 | Re: Is monotonous xid8 is a right way to do? |
| Previous Message | Dagfinn Ilmari Mannsåker | 2022-04-01 12:33:05 | Re: Is monotonous xid8 is a right way to do? |