On Mon, May 14, 2012 at 02:09:10PM -0700, Alex Converse wrote:
> On Mon, May 14, 2012 at 2:07 PM, Måns Rullgård <[email protected]> wrote:
> > Janne Grunau <[email protected]> writes:
> >
> > > Fixes fate on cross-compiled/wine mingw32 configs. The diff comaprison
> > > ignores space changes too.
> > > ---
> > >  tests/fate-run.sh |    2 +-
> > >  1 files changed, 1 insertions(+), 1 deletions(-)
> > >
> > > diff --git a/tests/fate-run.sh b/tests/fate-run.sh
> > > index a45e9a9..367b152 100755
> > > --- a/tests/fate-run.sh
> > > +++ b/tests/fate-run.sh
> > > @@ -61,7 +61,7 @@ stddev(){
> > >  }
> > >
> > >  oneline(){
> > > -    val=$(cat "$2")
> > > +    val=$(tr -d '\r\n' < "$2")
> >
> > NAK
> 
> I like it

I don't.

These shell scripts are tricky enough as-is.  These workarounds start
piling up quickly and before you know it you start breaking one fringe
setup or the other with every little innocuous-looking change.

I really hate having to be a curmudgeon about this, but I've been through
this with our and MPlayer's configure and it quickly becomes a spiraling
path down to madness...

> > We should not be doing hacks like this.  They make the code obscure, and
> > it is hard to predict where they will be needed.
> >
> > The systems currently failing are frankensetups which should be fixed to
> > present proper data to the scripts.  For example, they could override
> > the system cat by setting PATH.  We don't use cat on any binary data, so
> > that should work fine.  I verified that all tests pass with this in
> > place of the normal cat:
> >
> > #! /bin/sh
> > /bin/cat "$@" | tr -d '\r'
> 
> I think assuming we'll never use cat on binary data is more invasive than 
> this.

Give me a shell one one of those Frankensetups, I'll fix it.

Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to