Re: incorrect xlog.c coverage report - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: incorrect xlog.c coverage report
Date
Msg-id 20181120155042.7j6fjrbasx2s3rnc@alvherre.pgsql
Whole thread Raw
In response to Re: incorrect xlog.c coverage report  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: incorrect xlog.c coverage report
List pgsql-hackers
On 2018-Nov-20, Tom Lane wrote:

> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> > On 2018-Nov-20, Peter Eisentraut wrote:
> >> I noticed some strangeness in the test coverage reporting.
> 
> > Not sure what to make of this.
> 
> What platform and compiler do you run the coverage build on?
> 
> (I'm remembering that gcov was pretty nearly entirely broken on
> Fedora awhile back.  Maybe it's just a little broken on whatever
> you're using.)

This is Debian 9.6.  gcov says:

$ gcov --version
gcov (Debian 6.3.0-18+deb9u1) 6.3.0 20170516

This matches the gcc version string exactly.

ccache is not being used.

configure: using compiler=gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516
configure: using CFLAGS=-Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement -Werror=vla
-Wendif-labels-Wmissing-format-attribute -Wformat-security -fno-strict-aliasing -fwrapv -fexcess-precision=standard
-fprofile-arcs-ftest-coverage 
 
configure: using CPPFLAGS= -D_GNU_SOURCE -I/usr/include/libxml2 
configure: using LDFLAGS=  -Wl,--as-needed


I wondered if perhaps gcov's data files are ending in the wrong place
for some reason, but I don't know how to verify that.  I also wondered
if test results would maybe not be saved for the postmaster executable
when run under the TAP test framework ... not sure about this.

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


pgsql-hackers by date:

Previous
From: "REIX, Tony"
Date:
Subject: RE: Shared Memory: How to use SYSV rather than MMAP ?
Next
From: Robert Haas
Date:
Subject: Re: Can I skip function ResolveRecoveryConflictWithSnapshot ifsetting hot_standby_feedback=on all the time