[MPlayer-dev-eng] Performance of libdvdread/libdvdcss

Vicente Sendra visenri at yahoo.es
Sat Sep 1 13:25:43 CEST 2012


--- El sáb, 1/9/12, Reimar Döffinger <Reimar.Doeffinger at gmx.de> escribió:

> De: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
> Asunto: Re: [MPlayer-dev-eng] Performance of libdvdread/libdvdcss
> Para: mplayer-dev-eng at mplayerhq.hu
> Fecha: sábado, 1 de septiembre, 2012 12:15
> On Sat, Sep 01, 2012 at 10:20:05AM
> +0100, Vicente Sendra wrote:
> > But we are doing things that we should not do, because
> we are using both functions with the os file handle (ioct)
> and functions with the C run-time file descriptor (read).
> And this is a bad thing if we read this:
> > http://support.microsoft.com/kb/99173/en-us
> 
> Reading Microsoft advices is usually useless because they
> don't try to
> put in enough information to allow you to understand it.
> There's no reason why using (these kinds of) ioctls natively
> and reads
> through the libc interface should be an issue.
> 
> > I'm going to try the mixed solution, reading with
> win2k, closing, then reading with libc with keys from
> win2k.
> 
> Why don't you try the much easier approach of changing the
> random access
> flag first?
> Whoever is going to review the patch is certain to ask you
> _why_ it
> makes a difference, and even if your approach works you'll
> have a
> very hard time figuring out why from there.

Sorry, i didn't say anything about that, but i tried both values (FILE_FLAG_SEQUENTIAL_SCAN and 0), and no difference was observed.

PS: Being used to visual studio and delphi, it's a real pain in the ass to go searching where is every function or struct declared. Do you know any IDE that can be used with MinGw that it's worth a try (and if possible free).


More information about the MPlayer-dev-eng mailing list