[FFmpeg-devel] [PATCH 4/6] tools/python: add help script to get sess_config

[Guo, Yejun]

> -----Original Message-----
> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of
> lance.lmwang at gmail.com
> Sent: 2021年4月28日 9:19
> To: ffmpeg-devel at ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH 4/6] tools/python: add help script to get
> sess_config
> 
> On Wed, Apr 28, 2021 at 12:44:06AM +0000, Guo, Yejun wrote:
> >
> >
> > > -----Original Message-----
> > > From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of
> > > lance.lmwang at gmail.com
> > > Sent: 2021年4月27日 17:26
> > > To: ffmpeg-devel at ffmpeg.org
> > > Subject: Re: [FFmpeg-devel] [PATCH 4/6] tools/python: add help script to
> > > get sess_config
> > >
> > > On Tue, Apr 27, 2021 at 06:41:11AM +0000, Guo, Yejun wrote:
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf
> Of
> > > > > lance.lmwang at gmail.com
> > > > > Sent: 2021年4月27日 14:29
> > > > > To: ffmpeg-devel at ffmpeg.org
> > > > > Subject: Re: [FFmpeg-devel] [PATCH 4/6] tools/python: add help script
> > > to
> > > > > get sess_config
> > > > >
> > > > > On Tue, Apr 27, 2021 at 04:25:55AM +0000, Guo, Yejun wrote:
> > > > > >
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Guo, Yejun
> > > > > > > Sent: 2021年4月27日 12:11
> > > > > > > To: FFmpeg development discussions and patches
> > > > > > > <ffmpeg-devel at ffmpeg.org>
> > > > > > > Subject: RE: [FFmpeg-devel] [PATCH 4/6] tools/python: add help
> > > script
> > > > > to
> > > > > > > get sess_config
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On
> > > Behalf
> > > > > Of
> > > > > > > > lance.lmwang at gmail.com
> > > > > > > > Sent: 2021年4月26日 18:49
> > > > > > > > To: ffmpeg-devel at ffmpeg.org
> > > > > > > > Cc: Limin Wang <lance.lmwang at gmail.com>
> > > > > > > > Subject: [FFmpeg-devel] [PATCH 4/6] tools/python: add help
> script
> > > to
> > > > > get
> > > > > > > > sess_config
> > > > > > > >
> > > > > > > > From: Limin Wang <lance.lmwang at gmail.com>
> > > > > > > >
> > > > > > > > Please note the byte order of the hex data is in normal order.
> > > > > > > >
> > > > > > > > Signed-off-by: Limin Wang <lance.lmwang at gmail.com>
> > > > > > > > ---
> > > > > > > >  tools/python/tf_sess_config.py | 44
> > > > > > > > ++++++++++++++++++++++++++++++++++++++++++
> > > > > > > >  1 file changed, 44 insertions(+)
> > > > > > > >  create mode 100644 tools/python/tf_sess_config.py
> > > > > > > >
> > > > > > > > diff --git a/tools/python/tf_sess_config.py
> > > > > > > b/tools/python/tf_sess_config.py
> > > > > > > > new file mode 100644
> > > > > > > > index 0000000..e4e38bd
> > > > > > > > --- /dev/null
> > > > > > > > +++ b/tools/python/tf_sess_config.py
> > > > > > > > @@ -0,0 +1,44 @@
> > > > > > >
> > > > > > > this patch changes the order in current implementation, we'd
> better
> > > > > > > merge patch 4 and patch 5 in a single patch, to adjust the order in
> > > one
> > > > > > > patch.
> > > > >
> > > > > I'm OK with that. I think few people have used the option yet.
> > > >
> > > > yes, but we still need to keep the patch modular. There will be
> > > misleading
> > > > if people bisect the code with patch 4, without patch 5.
> > >
> > > OK, will update and merge the two patch.
> > >
> > > >
> > > > >
> > > > > >
> > > > > > and, we may remove '0x' at the beginning of value, since it is no
> > > longer
> > > > > > a hex value in math. It is the byte order.
> > > > >
> > > > > For hex string, it's preferable to prefix with '0x'.
> > > >
> > > > for example, '0x123' is not the value of 3+2*16+1*256, I'm afraid it is
> > > confusing.
> > >
> > > No, your exmple is invalid and will report error for it expects hex stringand
> > > 2 hex digits
> > > can represent 1 byte. That's why the current code will check whether
> > > sess_config_length
> > > is even.
> >
> > my example is to show it is not a valid hex value in math.
> >
> > >
> > >
> > > >
> > > > anyway, '0x' as prefix is also an option, we can output a message to
> > > explain it
> > > > in python script.
> >
> > please output a message in python script for an explanation of the order,
> thanks.
> >
> 
> I'm not clear for what's message for explanation, what's message you expect?

like the message in the commit log.

> 
> TF_SetConfig wants user to provide the config as a serialized protobuf string.
> If you turn on the print by json in the python script, that's print(list(map(hex,
> s)))
> ['0x10', '0x2', '0x28', '0x5', '0x32', '0xe', '0x9', '0x9a', '0x99', '0x99', '0x99',
> '0x99', '0x99', '0xb9', '0x3f', '0x20', '0x1', '0x2a', '0x1', '0x30', '0x38', '0x1']
> 
> So we use the hex string format like below so that we can use them in config
> easily.
> 0x10022805320e099a9999999999b93f20012a01303801