From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
cc: pgsql-www@lists.postgresql.org
Subject: Re: Bogus reports from coverage.postgresql.org
In-reply-to: <20180310040627.nwckurme2w42onp3@alap3.anarazel.de>
References: <30022.1520642188@sss.pgh.pa.us>
 <20180310040627.nwckurme2w42onp3@alap3.anarazel.de>
Comments: In-reply-to Andres Freund <andres@anarazel.de>
	message dated "Fri, 09 Mar 2018 20:06:27 -0800"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <15757.1520656953.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 09 Mar 2018 23:42:33 -0500
Message-ID: <15758.1520656953@sss.pgh.pa.us>
Precedence: bulk

Andres Freund <andres@anarazel.de> writes:
> On 2018-03-09 19:36:28 -0500, Tom Lane wrote:
>> I happened to notice that if you look at
>> https://coverage.postgresql.org/src/backend/optimizer/util/predtest.c.g=
cov.html
>> it claims that the two occurrences of
>> elog(ERROR, "predicate_classify returned a bogus value");
>> at lines 492 and 803 are reached.

> Is it possible that the compiler figured that the patch is unreachable
> and removed the code?

Maybe, but then it's even less reasonable to claim that the lines
were reached.

In general, lcov seems to ignore lines for which no code was generated,
which is fine by me.

			regards, tom lane