Re: [HACKERS] coverage analysis improvements - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: [HACKERS] coverage analysis improvements
Date
Msg-id CAB7nPqQKXMwVov1=bQLHZGzje5QmQbDGGEeaefHiN4uENdo63w@mail.gmail.com
Whole thread Raw
In response to [HACKERS] coverage analysis improvements  (Peter Eisentraut <peter.eisentraut@2ndquadrant.com>)
Responses Re: [HACKERS] coverage analysis improvements  (Peter Eisentraut <peter.eisentraut@2ndquadrant.com>)
List pgsql-hackers
On Wed, Aug 23, 2017 at 3:47 AM, Peter Eisentraut
<peter.eisentraut@2ndquadrant.com> wrote:
> On 8/21/17 01:23, Michael Paquier wrote:
>> Patch 0001 fails to apply as of c629324.
>
> Updated patches attached.
>
>> Which versions of lcov and gcov did you use for your tests?
>
> LCOV version 1.13, and gcc-7 and gcc-6

LCOV can be compiled from here (I have for example just changed PREFIX
in the main makefile):
https://github.com/linux-test-project/lcov.gi
And testing down to 1.11 I am not seeing issues with your patches. I
have used gcc 7.1.1.

Patch 0001 removes the .gcov files, which offer a text representation
of the coverage. Sometimes I use that with a terminal... Not sure for
the others, but that's my status on the matter. This also removes the
target coverage. Please note that on some distributions, like, err...
ArchLinux, lcov is not packaged in the core packages and it is
necessary to make use of external sources (AUR). It would be nice to
keep the original gcov calls as well, and keep the split between
coverage-html and coverage. I think as well that html generate should
be done only if lcov is found, and that text generation should be done
only if gcov is used. It is annoying to see --enable-coverage fail
because lcov only is missing, but it is not mandatory for coverage.

Patches 2, 4, 5, 6 and 9 are good independent ideas.
-- 
Michael



pgsql-hackers by date:

Previous
From: Michael Banck
Date:
Subject: Re: [HACKERS] [PATCH v1] Add and report the new "in_hot_standby" GUCpseudo-variable.
Next
From: Jeevan Ladhe
Date:
Subject: Re: [HACKERS] Default Partition for Range