[Ffmpeg-devel] MPEG2 seeking broken

Steve Lhomme steve.lhomme
Thu Apr 20 13:44:06 CEST 2006

Michael Niedermayer wrote:
>>>>> the code submitted is nothing else as
>>>>> if(ref_picture == NULL){
>>>>> #ifdef DEBUG
>>>>>    abort();
>>>>> #endif
>>>>>    ...
>>>>> }
>>>> I guess it's up to the MPEG2 maintainer to fix that bug the FFMPEG way then.
>>> first you post a bugreport related to your code which is not reproducable
>>> in ffmpeg and ffplay
>> http://article.gmane.org/gmane.comp.video.ffmpeg.devel/29795
>> All the rest of the email is based on the false assumption that I never
>> submitted a bug report and here is the proof I did. Your memory is very
>> selective.
> its not, i did re-read the thread before replying
> the message you refer to above does say at the very top that its not
> reproduceable with ffplay, which makes it everything but a bugreport
> relevent to ffmpeg/ffplay

Quoting parts of a message without the context and following text is 
usually what trolls do. So I'll quote what you should have read:

"As a follow-up to the email I sent a few days ago, I tried ffplay
(latest from CVS) on Linux on chems1.vob. The default seeking mode of
ffplay works.

But it's not "precise" enough for me (seeking 10s away is too far for
what we are using to analyse video frames). So I added the value set by
our application to seek inside any file."

My english may not be perfect, but it means I replaced a value in ffplay 
with the value that triggers the crash in my sample. And then ffplay 
crashed. So yes, this bug is not only relevant to ffmpeg/ffplay but it 
can even be reproduced on Linux built with gdb with your code from CVS 
(no it's not *my* code).

> also theres IIRC no applyable patch with the modifications
> needed to cause the crash you speak about

The crash you can't reproduce ? Maybe the crash that doesn't exist ? The 
crash I created to bother you and fill up the ML with my trolling and to 
find a way to spend my time wasting other people's time ?

The crash exists and is due to the MPEG2 decoder, it can be reproduced 
the way I explained it, and I have better things to do that losing my 
time sending crash reports that nobody is even reading and commenting.

I've spent time debugging this code and fixed it and now it works fine 
for me. I shared it, you don't care, that's your problem.

> and again bug(report) or not iam not obliged to apply your patch and neither
> am i obliged to fix it myself in the case that i choose not to accept your 
> patch though i do feel obliged to keep trolls from wasting our time ...

So what are maintainers for ? Ignoring bug reports in the code they 
maintain ?

And who wasted his time here ? Me by trying to share my work with you ? 
Or you doing your best to keep the status-quo on this bug ?


More information about the ffmpeg-devel mailing list