[Bug 766010] jifmux fails to find EOI if it's not within the last 5 bytes

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=766010

--- Comment #11 from Mohammed Sameer <msameer at foolab.org> ---
(In reply to Thiago Sousa Santos from comment #10)
> (In reply to Mohammed Sameer from comment #9)
> > (In reply to Thiago Sousa Santos from comment #8)
> > > (In reply to Mohammed Sameer from comment #7)
> > > > Do we know what kind of data is present if we scan backwards? It is in
> > > > theory more efficient but I have no idea what kind of byte values I should
> > > > be watching for.
> > > 
> > > What do you mean? jifmux was made to deal with jpeg buffers, one jpeg frame
> > > per buffer. What other type of data could be present?
> > 
> > jifmux is as good as its code is :)
> > 
> > If I write the code to scan the stream backwards, I will have to make sure
> > that I do it correctly. I don't know what markers or bytes will follow so
> > the code I write might cause issues. This is what I meant.
> 
> I'm no jpeg expert but you should be looking for a sequence of 0xFFDX, where
> X > 9 IIRC. Your code already does something similar.

I will try doing that then :)

> > 
> > > Let's take a look at this bug from another angle, did you consider using
> > > jpegparse before the jifmux? It should be able to strip that extra padding
> > > data and seems like a more appropriate element to do it instead of having it
> > > in jifmux.
> > 
> > camerabin picks the needed elements. I don't know how to force jpegparse :-(
> 
> I guess it makes sense to modify camerabin to plug a jpegparse, then. Making
> sure that those extra padding stuff is removed.

I tried jpegparse but it does not work:

0:00:00.023888024 30678  0x8b28180 WARN               baseparse
gstbaseparse.c:3569:gst_base_parse_loop:<jpegparse0> error: No valid frames
found before end of stream
ERROR: from element /GstPipeline:pipeline0/GstJpegParse:jpegparse0: No valid
frames found before end of stream

and the output file is empty.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.