[PATCH] quartz: Don't round a <1sec difference to 0 in WAVEParserImpl_seek

[Andrew Eikum]

I agree this seems weird, but that makes me wonder why it's there in
the first place. I've kind of lost track of the quartz interfaces; can
you write a test for this? E.g. Set two positions near each other and
Get to verify it changed.

I also noticed MPEGSplitter_seek and AVISplitter_seek have similar
code. I think writing a test for one and duplicating that logic in the
others would be fine.

Andrew

On Sun, Sep 23, 2018 at 02:22:26PM -0600, Alex Henrie wrote:
> Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=34302
> Signed-off-by: Alex Henrie <alexhenrie24 at gmail.com>
> ---
> For some reason, the game sets the audio stream position to a negative
> value, then seeks to 0. If Wine decides that it doesn't actually need to
> do the seek operation, it then reads garbage data from the negative
> offset.
> 
> Unfortunately, there is still another bug with the audio in this game:
> Wine keeps reading past the end of the audio stream, causing a crash at
> the end of the audio clip.
> ---
>  dlls/quartz/waveparser.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/dlls/quartz/waveparser.c b/dlls/quartz/waveparser.c
> index fa9cd45d27..995e43196d 100644
> --- a/dlls/quartz/waveparser.c
> +++ b/dlls/quartz/waveparser.c
> @@ -211,10 +211,9 @@ static HRESULT WINAPI WAVEParserImpl_seek(IMediaSeeking *iface)
>          return E_INVALIDARG;
>      }
>  
> -    if (curpos/1000000 == newpos/1000000)
> +    if (curpos == newpos)
>      {
> -        TRACE("Requesting position %s same as current position %s\n",
> -              wine_dbgstr_longlong(newpos), wine_dbgstr_longlong(curpos));
> +        TRACE("Requesting position %s same as current position\n", wine_dbgstr_longlong(newpos));
>          return S_OK;
>      }
>  
> -- 
> 2.19.0
>