Re: incorrect xlog.c coverage report

On 2019-May-31, Alvaro Herrera wrote:

> On 2019-May-30, Michael Paquier wrote:
>
> > On Wed, May 29, 2019 at 12:09:08PM -0400, Alvaro Herrera wrote:
> > > Are there objections to doing that now on the master branch?
> >
> > Adding the flush call just on HEAD is fine for me. Not sure that
> > there is an actual reason to back-patch that.
>
> Okay ... I did that (patch attached), and while my new __gcov_flush()
> shows as covered after I run the src/test/recovery tests, the function I
> mentioned earlier (validateRecoveryParameters) is not any more covered
> after the patch than it was before.

I forgot to mention that this patch produces a new warning:

/pgsql/source/master/src/backend/tcop/postgres.c: In function 'quickdie':
/pgsql/source/master/src/backend/tcop/postgres.c:2737:2: warning: implicit declaration of function '__gcov_flush'; did you mean 'pq_flush'? [-Wimplicit-function-declaration]
__gcov_flush();
^~~~~~~~~~~~
pq_flush

I couldn't find a way to squelch that. gcc devs in their infinite
wisdom don't provide a prototype for it, apparently.

Another thing I thought about was adding a configure test for the
function, but a) apparently the function is very old so it's not
necessary, and b) it fails anyway apparently because of that warning.

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services