On Tue, 12 May 2015 04:55:40 +0000 Sebastian Dransfeld
<sebastian.dransf...@gmail.com> said:

> tir. 12. mai 2015, 01:13 skrev Carsten Haitzler <ras...@rasterman.com>:
> 
> On Mon, 11 May 2015 17:06:36 +0200 Stefan Schmidt <ste...@osg.samsung.com>
> said:
> 
> > Hello.
> >
> > On 11/05/15 16:57, Stefan Schmidt wrote:
> > > Hello.
> > >
> > > On 09/04/15 14:34, Stefan Schmidt wrote:
> > >> Hello.
> > >>
> > >> On 01/04/15 14:30, Stefan Schmidt wrote:
> > >>> Hello.
> > >>>
> > >>> I brought this up during the recent EFL Dev Day and we discussed it a
> bit.
> > >>>
> > >>> Now I want to bring it here for wider discussion and give everybody a
> > >>> chance to have its say.
> > >>>
> > >>> Background is that we struggle to have a good testing coverage with
> unit
> > >>> tests of our code base. We have initiatives to improve this (which is
> > >>> very much welcome!) but we also need to address one of the root causes
> > >>> here. If you look at our current test coverage status you will see
> that
> > >>> some parts are way better than others:
> > >>> https://build.enlightenment.org/view/Test%20Coverage/
> > >>>
> > >>> When we bring new EAPI's we should try as hard as possible to have
> good
> > >>> unit tests committed together with the new EAPI's.
> > >>>
> > >>> I want to make this mandatory for cases where unit tests are possible.
> > >>> Cases which have specific hardware requirements or need interaction
> > >>> might be harder and I'm willing to accept the fact that sometimes it
> is
> > >>> almost impossible. But in many many of our newly added EAPI's added
> > >>> units tests for them is doable. Libs like eina, eo, eolian and eet
> > >>> should be straight forward.
> > >>>
> > >>> What I propose is that we have it mandatory to add test cases for new
> > >>> EAPI's and add an explanation to the commit message if you do _not_
> > >>> submit the unit tests and give a technical reason for it.
> > >>>
> > >>> What do you folks think about it? The feedback I got during the dev
> day
> > >>> was people are happy with it as long as it covers valid exceptions.
> > >> The feedback I got was only positive and after over a week everybody
> > >> should have been able to say if he/she disagrees.
> > >>
> > >> If you bring in new EAPI's it is mandatory from now on to add test
> cases
> > >> for them. If you tried and failed due to technical problems describe in
> > >> the commit message what hindered you to add test cases for them.
> > > While this have been a bit lax while the 1.14 cycle was still going on
> > > for 1.15 we should have an eye on this during reviews.
> > >
> > > I'm running an abi_checker test from 1.14 to latest HEAD to see what got
> > > added so far and which are missing test cases. Be prepared to be
> > > uncovered shortly if you have done that. :)
> > Some pieces are still missing as I have nothign to compare them against
> > ecore-drm and I need to check why elua does not show up)
> >
> > For the rest I see this new APIs:
> >
> > efl_player.eo.h, libefl.so.1.14.99
> > efl_player_length_get ( )
> > efl_player_seekable_get ( )
> >
> > Tom?
> >
> > eina_crc.h, libeina.so.1.14.99
> > eina_crc ( char const* key, int len, unsigned int seed, Eina_Bool
> > start_stream )
> >
> > A test case has been added for this. Weel done.
> >
> > eina_evlog.h, libeina.so.1.14.99
> > eina_evlog ( char const* event, void* obj, double srctime, char const*
> > detail )
> > eina_evlog_start ( )
> > eina_evlog_steal ( )
> > eina_evlog_stop ( )
> >
> > Raster, spankies! :P
> 
> dude - i just added them... i'm still not done. also this is intended for
> efl
> internal use (see docs). they may be eapi but not meant for use outside of
> efl
> itself (EPI only because symbols need to be visible). if i could make this
> non-eapi and we had a single libefl.so - i would.
> 
> 
> Hide them with a define?

what value does this add? (other than adding more code to have to #define
something in every efl file that uses them - and i intend that to be many)?

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    ras...@rasterman.com


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
enlightenment-devel mailing list
enlightenment-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to