Re: Operation log for major operations - Mailing list pgsql-hackers

From Ted Yu
Subject Re: Operation log for major operations
Date
Msg-id CALte62w+yo33GxQ9FRc4Ho2Cp-phg0-SBiPoWr15Lc1H8gumJA@mail.gmail.com
Whole thread Raw
In response to Re: Operation log for major operations  (Dmitry Koval <d.koval@postgrespro.ru>)
Responses Re: Operation log for major operations
List pgsql-hackers


On Fri, Jan 20, 2023 at 1:19 AM Dmitry Koval <d.koval@postgrespro.ru> wrote:
Thanks, Ted Yu!

 > Please update year for the license in pg_controllog.c

License updated for files pg_controllog.c, controllog_utils.c,
pg_controllog.h, controllog_utils.h.

 > +check_file_exists(const char *datadir, const char *path)
 > There is existing helper function such as:
 > src/backend/utils/fmgr/dfmgr.c:static bool file_exists(const char *name);
 > Is it possible to reuse that code ?

There are a lot of functions that check the file existence:

1) src/backend/utils/fmgr/dfmgr.c:static bool file_exists(const char *name);
2) src/backend/jit/jit.c:static bool file_exists(const char *name);
3) src/test/regress/pg_regress.c:bool file_exists(const char *file);
4) src/bin/pg_upgrade/exec.c:bool pid_lock_file_exists(const char *datadir);
5) src/backend/commands/extension.c:bool extension_file_exists(const
char *extensionName);

But there is no unified function: different components use their own
function with their own specific.
Probably we can not reuse code of dfmgr.c:file_exists() because this
function skip "errno == EACCES" (this error is critical for us).
I copied for src/bin/pg_rewind/file_ops.c:check_file_exists() code of
function jit.c:file_exists() (with adaptation for the utility).

--
With best regards,
Dmitry Koval

Postgres Professional: http://postgrespro.com
Hi,
Maybe another discussion thread can be created for the consolidation of XX_file_exists functions.

Cheers

pgsql-hackers by date:

Previous
From: Marco Slot
Date:
Subject: Re: [PATCH] Use indexes on the subscriber when REPLICA IDENTITY is full on the publisher
Next
From: Robert Haas
Date:
Subject: Re: Non-superuser subscription owners