On Sat, May 11, 2013 at 2:51 AM, Justin Clift <jclift@xxxxxxxxxx> wrote:
>
> Heh, I personally have no idea with gcov/lcov stuff, so other people that
> do have a clue with it would be the ones to respond. :)
>
My sincerest apologies if I misunderstand here, but... ( i cant tell
if your kidding or not about not having a clue).
gcov/lcov basically shows you how much of your codebase actually gets
executed during your testing. First you create a baseline ('before')
snapshot of your project. Then you run all of your tests. And then you
take another 'after' snapshot of your project. By comparing the two
snapshots you can now tell exactly what code was executed during your
tests. This tells you how effective your tests are, and in which areas
you may need to do more testing.
So in the glusterfs case here, for example shows that about 80% all of
'socket.c' gets executed which is good. No immediate need to devise
any new tests that explicitly utilize it.
http://lbalbalba.x90x.net/lcov/glusterfs/rpc/rpc-transport/socket/src/socket.c.gcov.html
On the other hand, 'read-only-common.c' does not get executed even
once. So you may need to create a test that uses that code.
http://lbalbalba.x90x.net/lcov/glusterfs/xlators/features/read-only/src/read-only-common.c.gcov.html
Anyway, again my apologies if you knew this already.
Hope this helps,
Regards
John Smith.
