On 02.01.2016 01:52, Rostislav Pehlivanov wrote:
>> To the very best of my knowledge you are allowed
>> to relicense code that was published under a two-
>> (or three-) clause BSD license under the GPL (and
>> the LGPL) as long as you keep the copyright
>> notice.
The BSD license is a very liberal
On 01.01.2016 21:16, Rostislav Pehlivanov wrote:
> Thanks for the feedback.
> Fixed all the issues you reported. The new revision is in this email.
Thanks. Unfortunately you introduced a much more serious crash problem.
Also, there is now a typo "libbavcodec" (two 'b') in the subject.
> --- /dev
Oh and by the way the title of the article is "Maintaining
Permissive-Licensed Files in a GPL-Licensed Project: Guidelines for
Developers". I read most of it and I'm under the impression that this is
about keeping BSD code inside LGPL code but still allowing the BSD code to
be used outside of the p
I'm not sure about that, the page doesn't list what happens if an LGPL
project uses some parts of a BSD project. The decoder was started from
scratch and the parts which were taken from libdaala (DCTs, PVQ and the
entropy decoder) have had major modifications done to them. So I still
think it's fin
>To the very best of my knowledge you are allowed
>to relicense code that was published under a two-
>(or three-) clause BSD license under the GPL (and
>the LGPL) as long as you keep the copyright
>notice.
Yes, that's what I've concluded as well.
Basically, the license says:
>Redistributions of so
Carl Eugen Hoyos ag.or.at> writes:
[...]
We were apparently both wrong:
https://www.softwarefreedom.org/resources/2007/gpl-non-gpl-collaboration.html
So please place the original (two-clause BSD) license
including the copyright statement under your (new) LGPL
license.
Sorry for the noise, Ca
Rostislav Pehlivanov gmail.com> writes:
> Fair enough, I will leave only the BSD-2 clause license
> and I'll put the files as an exception when I send the
> decoder as another RFC.
This would be fine but please see below: If you prefer
the LGPL, please use it!
> Ronald has agreed to review t
Fair enough, I will leave only the BSD-2 clause license and I'll put the
files as an exception when I send the decoder as another RFC.
Ronald has agreed to review the decoder (on a technical side of things) if
he has time this weekend so I'll wait for him to try to finish his review.
In any case,
Rostislav Pehlivanov gmail.com> writes:
> The reason why the FFmpeg license was added as well
> was to keep FATE happy
Then please add the new license to FATE but remove
the unintended license from the new source files.
You could also leave the test broken: I believe that
this should not be
The reason why the FFmpeg license was added as well was to keep FATE happy
and not have to put an exception for it (since all the exceptions are just
for empty template files). The way I understand it is that the license
above is the license for the file and all the code within it, while the
licens
Rostislav Pehlivanov gmail.com> writes:
> >this would fail on beos acording to doc/errno.txt
> Fixed, replaced with EINVAL in my private tree (
> https://github.com/atomnuker/FFmpeg
Sorry but the licensing simply makes no sense now:
The files should be - at you choice - either
licensed under t
>this would fail on beos acording to doc/errno.txt
Fixed, replaced with EINVAL in my private tree (
https://github.com/atomnuker/FFmpeg) and in the patch attached with this
email. Not sure how appropriate EINVAL is but other decoders use it too.
Also as suggested moved the vector functions to daal
On Fri, Jan 01, 2016 at 08:16:12PM +, Rostislav Pehlivanov wrote:
[...]
> +static av_cold int daala_decode_init(AVCodecContext *avctx)
> +{
> +int i, r_w, r_h, err = 0;
> +DaalaContext *s = avctx->priv_data;
> +
> +/* Inits a default QM, if the file isn't using the default it will
On Fri, Jan 1, 2016 at 12:16 PM, Rostislav Pehlivanov
wrote:
> Thanks for the feedback.
> Fixed all the issues you reported. The new revision is in this email.
>
> I've replaced the daalaent_cphase with a macro instead. Saves a few lines.
>
> I don't think it's a problem having large functions mar
14 matches
Mail list logo
