Re: Merge compact/non compact commits, make aborts dynamically sized
От | Simon Riggs |
---|---|
Тема | Re: Merge compact/non compact commits, make aborts dynamically sized |
Дата | |
Msg-id | CA+U5nMKH2xUzcHyLGCXPJMvF8kMt_mvQ=D9arx+1z-O61D1Jkw@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: Merge compact/non compact commits, make aborts dynamically sized (Heikki Linnakangas <hlinnakangas@vmware.com>) |
Список | pgsql-hackers |
On 24 February 2015 at 18:51, Heikki Linnakangas <hlinnakangas@vmware.com> wrote: > On 02/20/2015 05:21 PM, Andres Freund wrote: >> >> In the attached patch I've merged compact/noncompact code, made aborts >> use similar logic to avoid including useless bytes and used both for the >> 2pc equivalents. > > > +1 for this approach in general. +1 also The compact commit record idea wasn't useful enough to stick with, especially in the light of logical rep, possible future extensions to the commit record and the prevalence of subtransactions. XactEmitCommitRecord and XactEmitAbortRecord are somewhat hard to understand. More comments or clearer design please. The record format itself needs more explanation than I see in the patch. Might be worth documenting when and how those fields would be populated also (my bad). It might be useful to have the optional field descriptions be named with a _opt suffix to make it clearer. -- Simon Riggs http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, RemoteDBA, Training & Services
В списке pgsql-hackers по дате отправления: