gst-streaming-server stalls

Tim-Philipp Müller t.i.m at zen.co.uk
Sun Jan 13 15:01:25 PST 2013


On Sun, 2013-01-13 at 23:06 +0100, marcin at saepia.net wrote:

Hi,

> I see that you've taken this a bit personally and that wasn't my aim.

Nope, not at all. This has nothing to do with 0.10 or 1.x or which one
is better for your specific use case at this point (I couldn't say), or
you appreciating the effort.

What you wrote just doesn't make sense in my view, but perhaps you meant
to write something else. I was merely trying to understand what it is
that 'frightens' you (a strong emotional term that does not quite fit
with your self-professed preference for basing decisions on hard data).


> It is normal that new software has tons of bugs. It is awesome that
> you fix them so quickly.

My point was that *all* software has tons of bug. Even the 5-year old
0.10.x series has tons of bugs (thousands per year fixed, thousands not
fixed). You need to test your use case and confirm things work well for
you, and then constantly re-test when you make changes, with any
version.

However, I would argue that 1.0 is not really new software. It is 0.10
with some API changes. Not more, not less. 95% of the code base is the
same a before.


> So it is also normal in such case that I prefer 0.10 with ca. 530
> fixed bugs than 1.0 with unknown amount of bugs yet to be fixed. I
> cannot take this risk unless I will be forced to.

Those 530 bugs are of course fixed in 1.x as well, plus many hundreds of
other known bugs that will never be fixed in 0.10. To me personally the
idea that 0.10 is somehow generally less risky is ludicrous, but I
understand it is hard to convey why this is so. In any case, as I said,
I was trying to understand the basis of your assessment, not trying to
persuade you of one or the other, since I simply don't know your use
case or what is best for it.

I think I have my answer now though, thanks.

 Cheers
  -Tim



More information about the gstreamer-devel mailing list