From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Thomas Munro <thomas.munro@enterprisedb.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Allowing printf("%m") only where it actually works
Date: Mon, 28 May 2018 11:14:21 -0400
Message-ID: <2685.1527520461@sss.pgh.pa.us> (raw)
In-Reply-To: <20180528150125.j2x3mp62ssbeg3h5@alvherre.pgsql>
References: <20180528150125.j2x3mp62ssbeg3h5@alvherre.pgsql>
Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> This seems to say that we oughta assign GetLastError() to saved_errno
> during errstart, then use %m in the errmsg() instead.
No, because in some parts of the code, errno does mean something,
even in Windows builds.
I think the right fix is to leave %m alone, and instead:
(1) save GetLastError's result along with errno during errstart.
(2) provide a function, say errlastwinerror(), to be used where
elog/ereport calls currently call GetLastError():
elog(ERROR, "something went wrong: error code %lu",
errlastwinerror());
What it does, of course, is to reach into the current error stack
level and retrieve the saved result.
We could use some hack along the line of what Thomas suggested
to enforce that this function is used rather than GetLastError().
It's amusing to speculate about whether we could actually cause
GetLastError() appearing within elog/ereport to get mapped into
this function, thus removing the need to change any code in call
sites. But I suspect there's no adequately compiler-independent
way to do that.
I wonder whether we need to back-patch this. I don't recall
seeing any field reports of misleading Windows error codes,
but I wonder how many people even look those up ...
regards, tom lane
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-hackers@postgresql.org
Cc: tgl@sss.pgh.pa.us, alvherre@2ndquadrant.com, thomas.munro@enterprisedb.com, pgsql-hackers@lists.postgresql.org
Subject: Re: Allowing printf("%m") only where it actually works
In-Reply-To: <2685.1527520461@sss.pgh.pa.us>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox