(thus
extending when AVBPrint is extended) but return FFMIN(actual_size,
INT_MAX).
But in the end, I just said that it felt wrong to have an arbitrary
limit just because of a return value type. I do not oppose any solution.
Regards,
--
Nicolas George
t" as "ts" can be understood as Transport Stream.
> I don't want "add_delay" as it can add (aka delay) or substract (aka advance).
> And I don't recommend "editpts" since at present it can't edit time stamps
> freely.
My arguments have be
rds,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subj
;hash, (uint8_t *)&data,
> sizeof(uint32_t));
> +av_hash_update(c->hashes[0], (uint8_t *)&data,
> sizeof(uint32_t));
> }
> } else
> - av_hash_update(c->hash, pk
rs have no chance of devolving into anything
except possibly garbled output on strange architectures for an obscure
format. Compilers are practical tools, not an axiomatic system.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
Andreas Rheinhardt (12019-08-08):
> Ping.
I just pushed your patch and mine.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffm
e)
> #endif
> case AV_FRAME_DATA_DYNAMIC_HDR_PLUS: return "HDR Dynamic Metadata
> SMPTE2094-40 (HDR10+)";
> case AV_FRAME_DATA_REGIONS_OF_INTEREST: return "Regions Of Interest";
> +case AV_FRAME_DATA_ENCODE_INFO: return "AVEn
he range, and for the few other cases, the caller can do the
check if necessary.
Also, make the fields that cannot be negative unsigned, and you can drop
the <0 test.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpe
ceful error: a crash, or
worse.
The assert mimics that, in a more convenient way since it gives the
reason and line number, and does not allow the bug to devolve into
something more serious than a crash.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
Juan De León (12019-08-15):
> Ping. Does anyone have any more feedback?
> If not, can anyone review this for pushing.
Less than 24 hours feel a bit short to get impatient.
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg
gether or make it an assert. A
NULL return is a compromise that is worse than either.
> I would prefer to discuss the actual data structure for now.
This is whai I can comment on.
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg-deve
ssive either?
> I couldn't help but perceive the discussion you were having and your
> intentions to post another version as ignoring others' opinions.
When some people want X and some people want not-X, somebody has to
decide. That is not ignoring, that is making
lution
is inferior.
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
ame(enum
> AVFrameSideDataType type)
> #endif
> case AV_FRAME_DATA_DYNAMIC_HDR_PLUS: return "HDR Dynamic Metadata
> SMPTE2094-40 (HDR10+)";
> case AV_FRAME_DATA_REGIONS_OF_INTEREST: return "Regions Of Interest";
> + case AV_FRAM
at is what code of conducts are for, we should start
enforcing one.
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
- Forwarded message from Nicolas George -
Date: Fri, 16 Aug 2019 13:55:18 +0200
From: Nicolas George
To: Juan De León
Subject: Re: [FFmpeg-devel] [PATCH] libavutil: AVEncodeInfo data structures
Juan De León (12019-08-15):
> I don't think it's common for size_t to wrap ar
Juan De León (12019-08-16):
> AVEncodeInfoFrame data structure to store as AVFrameSideData of type
> AV_FRAME_DATA_ENCODE_INFO.
> The structure stores quantization index for each plane, DC/AC deltas
> for luma and chroma planes, and an array of AVEncodeInfoBlock type
> denoting position, size, and
- sizeof(block)
+ n * sizeof(block). It result in allocating an incomplete info
structure, which I think is fine because nothing will access the missing
part.
In any case, you need to check that neither the multiplication nor the
division overflow. And you need to do it beforehand.
Regards,
--
N
return "Regions Of Interest";
> +case AV_FRAME_DATA_ENCODE_INFO: return "AVEncodeInfo";
> }
> return NULL;
> }
> diff --git a/libavutil/frame.h b/libavutil/frame.h
> index 5d3231e7bb..ec112c5d15 100644
> --- a/libavutil/frame.h
> +++ b/libavutil/frame.h
> @@ -179,6 +179,13 @@ enum AVFrameSideDataType {
> * array element is implied by AVFrameSideData.size /
> AVRegionOfInterest.self_size.
> */
> AV_FRAME_DATA_REGIONS_OF_INTEREST,
> +/**
> + * Extract frame and block encode info from supported decoders. The data
> + * stored is an AVEncodeInfoFrame type, which contains an array of
> + * AVEncodeInfoBlock. Described in libavuitls/encode_info.h
> + * Can be allocated in the frame directly with
> av_encode_info_create_side_data().
> + */
> +AV_FRAME_DATA_ENCODE_INFO,
> };
>
> enum AVActiveFormatDescription {
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
t;HDR Dynamic Metadata
> SMPTE2094-40 (HDR10+)";
> case AV_FRAME_DATA_REGIONS_OF_INTEREST: return "Regions Of Interest";
> + case AV_FRAME_DATA_ENCODE_INFO: return "AVEncodeInfo";
> }
> return NULL;
> }
> diff --git a/libavutil/frame.h b/l
r than a macro, I personally consider
bad style to rely on a magic value for error: return a boolean for
success and pass the size through a pointer.
But in this case, a macro would probably be the simplest.
Regards,
--
Nicolas George
signature.asc
Description: PGP sig
ave no more comments. Thanks for you efforts.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
ces beyond that in the ticket e.g. ffmpeg -i avi -i
> mp4 outfile where both files have streams with identical parameters.
Is it really a regression? I mean: was the old behaviour really better
in many cases?
Regards,
--
Nicolas George
signature.asc
D
Paul B Mahol (2018-08-23):
> Ugh, can you explain why?
Need provable claims.
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Paul B Mahol (2018-08-23):
> Try it yourself
Your patch. Still not ok.
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
r instead.
If you just remove the assert, then the invalid value is passed to the
rest of the code. You know better than me how it will react to it.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing
filters like that.
No time to look at the patch in more details now.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Nicolas George (2018-08-22):
> > Fix trac ticket #7351.
> >
> > Signed-off-by: Nicolas George
> > ---
> > libavfilter/avf_concat.c | 156 +++
> > 1 file changed, 78 insertions(+), 78 deletions(-)
>
> Will
Fix trac ticket #7387.
Signed-off-by: Nicolas George
---
libavformat/libsmbclient.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Untested, I do not have a SMB share at hand.
diff --git a/libavformat/libsmbclient.c b/libavformat/libsmbclient.c
index b68cd8bd79..3285868957 100644
Nicolas George (2018-08-30):
> Fix trac ticket #7387.
>
> Signed-off-by: Nicolas George
> ---
> libavformat/libsmbclient.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
>
> Untested, I do not have a SMB share at hand.
Confirmed to work on tra
Nicolas George (2018-08-30):
> Confirmed to work on trac. Will push soon.
Pushed.
And backported to 4.0; hope I did not make any mistake, this is my first
backport.
Regards,
--
Nicolas George
signature.asc
Description: Digital signat
ll try to remember this for the next time. Thanks.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
y to miss important mails and consumes a lot of time
and energy.
You should not have added a queue in your context at all, you should
have used the one already present using the API in filters.h.
Regards,
--
Nicolas George
signature.asc
Description: Digital
s kind of case.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
told you: use the API in filters.h, the various
ff_inlink_* functions, rather than accessing the queue directly. Better
abstraction, less maintenance later.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mail
ooks like ff_inlink_queued_frames() and ff_inlink_peek_frame()
would be just what you need.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listin
.
What do you think of it, in terms of usefulness and difficulty?
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Paul B Mahol (2018-09-30):
> No, AVERROR_BUG is better.
For consistency within a single function, an assert is the correct
choice.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-de
sting here: the consistency
between the preliminary test and the later switch statement.
If you want to argue for AVERROR_BUG, please give arguments.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing li
not a professional project, we can afford to do things
properly.
A program that crashes is sloppy, indeed. Assert failures are bad. But
AVERROR_BUG is just as bad. It should not happen after commit either.
Therefore, this is not an argument to choose one over the other.
Regards,
--
Nicolas
d frame must exist.
But we can discuss it.
> + */
> +AVFrame *ff_inlink_peek_frame(AVFilterLink *link, size_t idx);
> +
> /**
> * Make sure a frame is writable.
> * This is similar to av_frame_make_writable() except it uses the link's
Apart from that, LGTM.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
for
status, but that is unrelated.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
LTER_FORWARD_WANTED(outlink, inlink);
> +
> +return FFERROR_NOT_READY;
> }
>
> static const AVFilterPad asetnsamples_inputs[] = {
> {
> -.name = "default",
> -.type = AVMEDIA_TYPE_AUDIO,
> -.filter_frame = filter_frame
order.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
->tempo > 2.0);
> +// samples are not expected to be skipped:
> +av_assert0(read_size <= atempo->ring);
Looks unrelated. Can you explain?
>
> while (atempo->position[0] < stop_here && src < src_end) {
> int src_samples = (src_end - s
pt_set or avfilter_process_command already returns error if out of
> range. So no need to check it again.
And you are changing the bound. It should be checked with Pavel, but I
think this change is wrong.
Regards,
--
Nicolas George
signature.asc
Description: Digit
Marton Balint (2018-10-01):
> Signed-off-by: Marton Balint
> ---
> libavfilter/avfilter.c | 10 ++
> libavfilter/filters.h | 13 +
> 2 files changed, 23 insertions(+)
LGTM, thanks.
Regards,
--
Nicolas George
signature.asc
Description: Di
ersion of the patch, thanks. But I am not the
maintainer of that file.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Paul B Mahol (2018-10-03):
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/af_asetnsamples.c | 146 --
> 1 file changed, 32 insertions(+), 114 deletions(-)
LGTM, thanks.
Regards,
--
Nicolas George
signature.asc
Description: Digit
No, I do not have a student lined
for this project.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Marton Balint (2018-10-07):
> Fixes Coverity CID 1439936.
>
> Signed-off-by: Marton Balint
> ---
> libavfilter/f_cue.c | 8 ++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
LGTM, sorry for having missed it.
Regards,
--
Nicolas George
signature.asc
De
Paul B Mahol (2018-10-25):
> What about making target speed configurable? 1x .. 1.5x .. 2x and so on.
That would certainly be a useful feature.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing l
care about developing ffmpeg with clang
to come up with a solution that does not make the code less readable and
harder to maintain.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Marton Balint (2018-11-01):
> Without this FTP just hangs on eof...
>
> Signed-off-by: Marton Balint
> ---
> libavformat/ftp.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
LGTM.
Regards,
--
Nicolas George
___
ff
Paul B Mahol (2018-11-12):
> > Did you write this function?
> No
Then the authorship and origin need to be documented.
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
rds,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Paul B Mahol (2018-11-12):
> How to document copyright of multiple authors?
What is the issue? libavcodec/j2kenc.c does it quite well.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffm
Tomas Härdin (2018-11-12):
> Why not set locale instead?
That would conflict with the wishes of the application. A library should
never change global state.
(Of course, the core of the problem is that locales are global state,
this is braindead design.)
Regards,
--
Nicolas Geo
ch.c
> Will apply.
You forgot the docs.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
eturn res;
> +}
> +
> +res = dav1d_get_picture(dav1d->c, &p);
I just noticed this: the lack of AVERROR() looks very suspicious. It
seems this library uses directly -errno values, which will coincide with
AVERROR() on most systems but not all.
Regards,
--
Nicolas Geo
Paul B Mahol (2018-11-16):
> +off_t shlim, shcnt;
All these off_t should have been size_t, or possibly ptrdiff_t. off_t is
for offsets in files.
Regards,
--
Nicolas George
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.
rd.
> (Or it happened when FFFrameQueue->queued was defined size_t.)
size_t is the correct type for a large number of objects that coexist in
memory.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel m
are not URIs by trying to get FFmpeg
to open file:///dev/nul%6c for example.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
meant to be
considered URIs. The documentation was added much later, and made the
same mistake you are doing now; same goes for a few private function
names.
> Escape sequences, if needed, can be fixed separately.
That would break a lot of working applications and is therefore not a
good idea.
be
an API break, it is not acceptable.
You can propose patches to make FFmpeg support real URIs instead / in
addition to its old pseudo-URI syntax, but you would need to design with
API compatibility in mind.
Regards,
--
Nicolas George
signatu
James Almer (2018-11-30):
> And using string literals in an assert may generate warnings.
[citation needed]
Anyway, these warnings would be broken, the code is valid.
Regards,
--
Nicolas George
signature.asc
Description: Digital signat
James Almer (2018-11-30):
> probably Clang.
I could have guessed.
Clang: "bogus warnings, gotta catch them all".
The solution is not to make the code less convenient but to fix the
compiler. Or ignore it.
Regards,
--
Nicolas George
signature.asc
Description: Digi
int ffurl_close(URLContext *h);
>
> +/**
> + * Flush any buffered data and synchronize the resource accessed
> + * by the URLContext h. This call may be expensive.
> + * Not all types of resources support syncing, the call is a no-op
> + * if sync is not supported or needed.
>
/os_support.h
> +++ b/libavformat/os_support.h
> @@ -93,6 +93,8 @@ static inline int is_dos_path(const char *path)
> #ifndef S_IWUSR
> #define S_IWUSR S_IWRITE
> #endif
> +
> +#define fsync _commit
Defining with the arguments would be more robust. A few occasions in the
ed to work. If they want to include FFmpeg
for their users' convenience, they have to make their software
LGPL-compatible.
Because "their users' convenience" is good publicity for them, they have
to pay for it, not us.
Regards,
--
Nicolas George
Dennis Mungai (2018-12-03):
> In this case , Carl's decision to strip their code from FFmpeg is valid.
> This is a clear violation of the license terms.
Indeed.
And please stop top-posting. If you do not know what it means, look it
up.
Regards,
--
Nicolas George
signature.asc
nt upstream?
I agree.
Maybe we can make an exception when the non-free upstream is dead yet
useful, like faac was for some time. But for contributors like this NDI
thing, either they make the build components (L)GPL-compatible or they
do not enjoy the benefit of being included.
Regards,
--
pyleft clauses rescinds all rights to use the original GPLed code.
Note that they did not only infringe on FFmpeg's license, they also
infringed on x264.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg
l.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
t; + * Flush any buffered data and synchronize the resource accessed
> + * by the URLContext h. This call may be expensive.
> + * Not all types of resources support syncing, the call is a no-op
> + * if sync is not supported or needed.
> + *
> + * @return AVIO_SYNC_SUCCESS on success, AVIO_SYNC_NOT_SUPPORTED if
> + * the operation is not supported and ignored, an AVERROR < 0
> + * on error.
> + */
> +int ffurl_sync(URLContext *h);
> +
> /**
> * Return the filesize of the resource accessed by h, AVERROR(ENOSYS)
> * if the operation is not supported by h, or another negative value
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
he sync was not done, then you did not need to
sync in the first place. Syncing is about guaranteeing the user that the
data is safe; if you ignore the information that it was not done, then
you are not guaranteeing it.
Regards,
--
Nicolas George
signat
Marton Balint (2018-12-09):
> Will apply soon.
Sorry, I was busy and it slipped my mind.
I will try hard to have a careful look tomorrow.
Regards,
--
Nicolas Geoge
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-
sistency
test were added to check the "should", but I do not consider blocking.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Marton Balint (2018-11-22):
> Signed-off-by: Marton Balint
> ---
> libavformat/concatdec.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
I suppose it is valid.
Regards,
--
Nicolas George
signature.asc
Description: Digit
Marton Balint (2018-11-22):
> Signed-off-by: Marton Balint
> ---
> libavformat/concatdec.c | 37 +++--
> 1 file changed, 23 insertions(+), 14 deletions(-)
LGTM.
Regards,
--
Nicolas George
signature.asc
Description: Digit
_duration(cat->cur_file,
> cat->avf);
>
> if (++fileno >= cat->nb_files) {
> cat->eof = 1;
I do not think it works. If the duration of the second file changes, for
example, then the start time of all subsequent files changes too, but
this patch does not update it. Seeking will show strange results.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
st->codecpar->sample_aspect_ratio.num *
> >> st->codecpar->width,
> >> + st->codecpar->sample_aspect_ratio.den *
> >> st->codecpar->height, INT_MAX);
> There is no av_reduce_q.
But there are av_mul_q() and av_div_q(), that would
Jun Zhao (2018-12-09):
> abuffersink need to setting AV_OPT_FLAG_AUDIO_PARAM flag.
>
> Signed-off-by: Jun Zhao
> ---
> libavfilter/buffersink.c |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
LGTM, thanks.
Regards,
--
Nicolas George
signature.asc
De
ere is a catch: if
the density is 20 px/mm horizontally and 10 px/mm vertically, that means
pixels are (1/20)×(1/10) mm², hence their aspect ratio is 10/20 = 1/2,
not 20/10 = 2/1.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
--
> tests/ref/fate/gifenc-rgb4_byte | 346
> tests/ref/fate/gifenc-rgb8 | 346
> 8 files changed, 1217 insertions(+), 1228 deletions(-)
Please document why you need to rewrite the encoder in order to get
muxing working.
Regards,
--
Nicolas G
Paul B Mahol (2018-12-12):
> It needs that because decoder & demuxer work that way.
This answer do not say anything. Please elaborate enough as to be
convincing. And put it in the commit message.
Regards,
--
Nicolas George
signature.asc
Description: Digital si
s delay between two packets.
> Codec copy need to have same behavior between demuxer and
> muxer to work correctly.
Thank you, this is convincing. In the commit message please.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
AV_CODEC_PROP_LOSSLESS,
> .mime_types= MT("image/gif"),
> },
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
ase in the seek code for that.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
FAICS, you case amounts to the fact that nobody spontaneously told you
they use this filter.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/list
fy one.
>(except for signalling precise speaker
> coordinates, which I don't think anything uses).
I think somebody uses it, because somebody felt the need to include it
in the standard. Therefore, we need to support it.
Regards,
--
Nicolas George
signatur
> thinks, but critically judge everything we're implementing here.
I am not suggesting following anything blindly. But files using this
feature can exist out there, so the question is whether to accept an API
that cannot handle them without loss of information.
Regards,
--
N
e simply a per-channel metadata dictionary, it would allow
both the speaker coordinate in this spec and the free-form label that I
want. Alternatively, we can use a more typed approach, like side data.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
les. Subtitles are not just a bunch of
pixels that get overlaid on top of the video. Well, they could be, but
it's not what the users expect. Subtitles are often hand-written,
partially or completely, and read directly. A tool that mangles it would
be useless for
ect not, because I
have looked for them and not found.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visi
es not mean they have to live in a different library.
Regards,
--
Nicolas George
signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link
nd lavf would greatly benefit from
> having more internal structure
Again, splitting the libraries is not necessary to have better code
organization.
Let me be clear: the organization in thematic sub-directories is fine.
The terrible waste is to produce sepa
hoops and
waste effort.
> Hard disagree on that one, but also quite off-topic, let's get back on
> subtitles.
This is an important discussion too. I changed the subject for people
who want to track accurately.
Regards,
--
Nicolas George
signature.asc
Description: PGP sign
gt; +}
> +
> +return next;
> +}
> +
> +static int flif16_parse(AVCodecParserContext *s, AVCodecContext *avctx,
> + const uint8_t **poutbuf, int *poutbuf_size,
> + const uint8_t *buf, int buf_size)
Alignment is off.
>
g.org/doxygen/trunk/gif__parser_8c_source.html
> I will probably prefix them with 'FLIF16_STATE_' then.
The name of the values are fine. I meant the name of the enum itself,
FLIF16ParseStates compared to the name of the type, flif16_states.
By the way, the name of enums is usually sin
301 - 400 of 4940 matches
Mail list logo
