[FFmpeg-devel] [PATCH v2] examples: add flac_test
nfxjfg at googlemail.com
Wed Apr 15 11:58:34 CEST 2015
On Wed, 15 Apr 2015 11:34:28 +0200
Nicolas George <george at nsup.org> wrote:
> Le sextidi 26 germinal, an CCXXIII, wm4 a écrit :
> [ I skip over the usual jealousy rants. ]
> > Your attitude also shows an amazing chauvinism: who cares about the API
> > users (and any downstream projects), as long as the command line tools
> > are fine?
> And you fail to understand. Once again.
> The command-line tools should be able to test the API. All of it. If there
> is a point of the API that is not possible to test with the command-line
> tools, the command-line tools must be extended to be able to. Period.
How hard is it to understand that a program built for a specific
purpose can not be made to use an API in all possible ways? Unless you
add dozens of redundant code paths to ffmpeg.c (do you want this?), you
won't be able to get full API coverage, and more importantly, you won't
be able to test all kinds of ways the API could be used in alternative
Artificial test programs can cover multiple ways of using the same API,
and we can make automated tests. This is an improvement over
downstreams breaking just because some implementation detail changed,
and they didn't do the same as ffmpeg.c
> Inventing a specific test that nobody will maintain is a waste of time.
No it won't. And they will be maintained, because they will be part of
FATE. (Unlike all our broken example programs, which -hilariously- some
people actually try to use as a base for their code.)
> > Because your world view seems to be restricted to the ffmpeg git repo.
> It seems to me this is the topic of this mailing-list.
The concerns of API users have a place on this list too.
More information about the ffmpeg-devel