Re: Remove extra Logging_collector check before calling SysLogger_Start()
От | Bharath Rupireddy |
---|---|
Тема | Re: Remove extra Logging_collector check before calling SysLogger_Start() |
Дата | |
Msg-id | CALj2ACW2pVV8zx_eC_f3WHwrbc1WZ0vV4vDrs1fYZWRAT9TaxQ@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: Remove extra Logging_collector check before calling SysLogger_Start() (Daniel Gustafsson <daniel@yesql.se>) |
Список | pgsql-hackers |
On Fri, Dec 3, 2021 at 1:43 PM Daniel Gustafsson <daniel@yesql.se> wrote: > > > On 3 Dec 2021, at 08:58, Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com> wrote: > > > It seems like there's an extra Logging_collector check before calling > > SysLogger_Start(). Note that the SysLogger_Start() has a check to > > return 0 if Logging_collector is off. This change is consistent with > > the other usage of SysLogger_Start(). > > > > /* If we have lost the log collector, try to start a new one */ > > - if (SysLoggerPID == 0 && Logging_collector) > > + if (SysLoggerPID == 0) > > SysLoggerPID = SysLogger_Start(); > > I think the code reads clearer with the Logging_collector check left intact, > and avoiding a function call in this codepath doesn't hurt. In that case, can we remove if (!Logging_collector) within SysLogger_Start() and have the check outside? This will save function call costs in the other places too. The pgarch_start and AutoVacuumingActive have checks outside PgArchStartupAllowed and AutoVacuumingActive. Regards, Bharath Rupireddy.
В списке pgsql-hackers по дате отправления: