[FFmpeg-devel] [PATCH] Fix a few warnings in libavutil test programs

[Diego Biurrun]

On Thu, Jan 10, 2008 at 11:58:43AM +0100, Michael Niedermayer wrote:
> On Thu, Jan 10, 2008 at 11:32:31AM +0100, Diego Biurrun wrote:
> > On Wed, Jan 09, 2008 at 04:24:18AM +0100, Michael Niedermayer wrote:
> > > On Wed, Jan 09, 2008 at 01:33:47AM +0000, M??ns Rullg??rd wrote:
> > > > Michael Niedermayer <michaelni at gmx.at> writes:
> > > >
> > > > > On Wed, Jan 09, 2008 at 12:14:19AM +0100, Diego Biurrun wrote:
> > > > >> Here is a patch that takes FFmpeg one step closer to a C
> > > > >> programming textbook ;-)
> > > > >>
> > > > >> Diego
> > > > >>
> > > > >> P.S.: Yes, I can add Makefile rules for all test programs if
> > > > >> wanted.
> > > > >
> > > > > You may apply this if and only if you write a fully working
> > > > > regression test for them so that make test, does run them and
> > > > > checks that they suceeded.
> > > >
> > > > WTF?  Please read my previous email again.
> > >
> > > previous? your? No, i dont see a relation between your previous
> > > mail and this thread. A thread in which i try to convince diego to
> > > do something usefull instead of fixing commented out code.
> >
> > I stumbled over that stuff while reviewing a patch from Diego#2, so I
> > thought I might as well make it usable.  How is that not useful?
>
> it worked fine without that change, so no you dont make anything
> useable its 99.9% nitpickery.

aes did not compile before I added #undef random, mathematics and
integer had the test code under '#if 0' and not under '#if TEST', so
they needed editing to compile.

But that was not the point I was trying to make: I stumbled over this
while reviewing a patch from Diego#2 (which was wrong and I committed
the right fix instead).  How is reviewing a patch not useful?

Since I had already compiled and partially edited those files I thought
I might as well silence a few warnings.  It's not like this takes more
than 5 minutes extra...

> If OTOH you would make them part of make test that would be usefull and
> could prevent actual bugs.

A prerequisite for that is having them compile.

Diego