gmane.comp.video.videolan.vlc.devel

Re: HLS player

Kevin Cadieux — 2013-06-19T01:53:55

The VLC console shows suspicious messages that are probably linked to my
problem:

main warning: buffer too late (-16083945 us): dropped


I get many of these in a row. Any ideas anyone?


On Tue, Jun 18, 2013 at 9:02 PM, Kevin Cadieux <kcadieux< at >gmail.com> wrote:

_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: HLS player

Kevin Cadieux — 2013-06-19T01:02:11

Alright, thanks for the explanation.

I am having a strange little problem right now. I compiled VLC on Ubuntu
using these guidelines: http://wiki.videolan.org/UnixCompile
Everything was fine until I started VLC and loaded a sample HLS stream. The
HLS feed kept jumping 10 to 15 seconds ahead every 2 seconds. For some
reason, VLC is calling the Read function of the HLS module way too fast
which causes the skipping ahead (I think). Does anyone have an idea why VLC
would be doing that? I didn't modify the source at all and the problem does
not occur when I use the real VLC release.

Thank you,

Kevin
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

[PATCH 2/2] vout: remove the dummy display pool

Rémi Denis-Courmont — 2013-06-18T20:07:20

With the previous commit, the vout wrapper's display pool is no longer
used ever if direct rendering is not in use. Removing it saves one
picture buffer.
---
 src/video_output/vout_wrapper.c |   15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

diff --git a/src/video_output/vout_wrapper.c b/src/video_output/vout_wrapper.c
index 8bcb118..68099d9 100644
--- a/src/video_output/vout_wrapper.c
+++ b/src/video_output/vout_wrapper.c
< at >< at > -113,16 +113,9 < at >< at > static void NoDrInit(vout_thread_t *vout)
     if (sys->display.use_dr)
         sys->display_pool = vout_display_Pool(sys->display.vd, 3);
     else
-        //sys->display_pool = picture_pool_Reserve(sys->decoder_pool, DISPLAY_PICTURE_COUNT);
-        sys->display_pool = picture_pool_NewFromFormat(&sys->display.vd->source, DISPLAY_PICTURE_COUNT);
+        sys->display_pool = NULL;
 }
-static void NoDrClean(vout_thread_t *vout)
-{
-    vout_thread_sys_t *sys = vout->p;
 
-    if (!sys->display.use_dr)
-        picture_pool_Delete(sys->display_po

[RFC 2.2] [PATCH 1/2] voit: avoid picture copy whendecoder and output format do not match

Rémi Denis-Courmont — 2013-06-18T20:07:19

When the decoder output and the display video formats are different,
the picture copy comes from "free" during the conversion from the
earlier format to the latter one. So do not bother with copying the
picture before the conversion.

The straight copy is still required however in the case that the video
formats are identical but the display pool is inadequate (too few
pictures, slow memory bus...).
---
 src/video_output/video_output.c |   42 +++++++++++++++++++--------------------
 1 file changed, 21 insertions(+), 21 deletions(-)

diff --git a/src/video_output/video_output.c b/src/video_output/video_output.c
index 8c4b329..2e5de0f 100644
--- a/src/video_output/video_output.c
+++ b/src/video_output/video_output.c
< at >< at > -986,37 +986,37 < at >< at > static int ThreadDisplayRenderPicture(vout_thread_t *vout, bool is_forced)
             return VLC_EGENERIC;
     }
 
-    picture_t *direct;
-    if (!is_direct) {
-        direct = picture_pool_Get(vout->p->display_pool);
-        if (direct) {
-            VideoFormatCopyCr

Re: Crystal HD: Software Deinterlacing possible?

Jean-Baptiste Kempf — 2013-06-18T12:10:28

On 18 Jun, Philip Höhlein wrote :

What deinterlacing mode are you using now?

Best regards,

Re: Crystal HD: Software Deinterlacing possible?

Philip Höhlein — 2013-06-18T12:05:57

Deinterlacing is working now, thanks to Steinar for his patch.
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: HLS player

Rémi Denis-Courmont — 2013-06-17T06:00:59

On Sun, 16 Jun 2013 19:57:28 -0400, Kevin Cadieux <kcadieux< at >gmail.com>
wrote:
Access
demuxing

Time-based seeking requires that the HLS plugin implements the
(access_)demux interface. The stream_filter interface only supports byte
range-based seeking, which AFAIK is inadequate for HLS.

Re: HLS player

Kevin Cadieux — 2013-06-16T23:57:28




What do you mean exactly by "Prepare the transition to a Demux or an Access
Demux module"? So far the HLS module doesn't seem to be doing any demuxing
at all. Are you saying it should take on new responsibilities?


Kevin
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: vlc_UrlParse: end username at the last < at > ratherthan the first

Rémi Denis-Courmont — 2013-06-16T13:37:44

Le dimanche 16 juin 2013 16:33:26, Rémi Denis-Courmont a écrit :

Another way to look at it: the URL is invalid if it has more than one < at > 
character in the authority. Which one the software accepts is outside the 
scope of the specification.

Alternatively, we could reject the URL completely. But I fear that would 
introduce a lot of new poorly tested error code paths.

Re: vlc_UrlParse: end username at the last < at > ratherthan the first

Rémi Denis-Courmont — 2013-06-16T13:33:26

Le dimanche 16 juin 2013 16:05:58, Francois Cartegnie a écrit :

There are huge work-arounds in the URL parsing and decoding code. For example, 
non-ASCII bytes are accepted by decode_URI(), even though they are clearly 
invalid. However, they cannot break a valid (corner) case. Valid hostnames or 
IP addresses do not contain < at >.

AFAICT, this change cannot break a valid use; in particular it will not break 
properly encoded usernames. Feel free to erase it, but at least Chroma works 
like that too.

Unfortunately, there are still quite many bugs in the URL handling of VLC (see 
section 5 of the RFC), most notably regarding reference resolution. This patch 
gets it neither closer nor further from correctness, I think.

Re: vlc_UrlParse: end username at the last < at > ratherthan the first

Francois Cartegnie — 2013-06-16T13:05:58

X-Git-Url:
http://git.videolan.org/?p=vlc.git;a=commitdiff_plain;h=886de66d58d930e46ef5cb2dc875d2aea5b83f71


Not a "better idea" but mandatory: RFC 3986 section 3.2.1
'< at >' being a 'reserved' character: RFC 3986 section 2.2

You would usually reject that kind of workaround, no ?

Francois
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: Translations for VLC 1.2.0

Christoph Miebach — 2013-06-16T09:38:22

Hello!

On 14.06.2013 16:02, Mr. Phan Anh wrote:

I can send you the latest pot file of course or you can download it 
yourself, but I strongly suggest you send me (no need for both mailing 
lists) your broken file and a short description what is going wrong 
before you start again from scratch.

The Vietnamese translation has more than 2000 translated messages that 
are used for 2.1 as well. I can delete the fuzzy ones for example, so 
you would only have to review them. And deleting/keeping a translation 
is very much faster than re-typing it.

And even if you deleted your local copy, I could offer you some kind of 
backup from our repository.

Best regards

Christoph
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: [PATCH] ISDB-S was not working.

HIRANO Takahito — 2013-06-15T18:12:35

I synced to the HEAD and confirmed that it works.  Thanks for the quick fix!

2013/6/16 Rémi Denis-Courmont <remi< at >remlab.net>:
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: [PATCH] ISDB-S was not working.

HIRANO Takahito — 2013-06-15T17:34:32

Thanks for applying the patch.

Actually, it turned out that my tuner does not support DTV_TONE.
That's why  my modification around sec_setup worked for me.

Can we add an option not to call DTV_TONE or can we just ignore the
error from DTV_TONE?

2013/6/11 Rémi Denis-Courmont <remi< at >remlab.net>:
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: Translations for VLC 1.2.0

Mr. Phan Anh — 2013-06-14T14:02:16

Please send me back the English File, I have trouble with the file which
was being done.

I will retranslate the VLC from scratch with the update.

So please send me the latest update english file translation to me.
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: [PATCH] ISDB-S was not working.

Rémi Denis-Courmont — 2013-06-15T17:53:03

Hello Takahito-san,

Le samedi 15 juin 2013 20:34:32, HIRANO Takahito a écrit :

Sure, fixed.

Re: "Race condition" in livehttp module

Enrique Arizón Benito — 2013-06-14T12:08:19

Hi,

Thanks for your help! I'll wait them for the patch.

Regards!

Enrique

On Thu, Jun 13, 2013 at 8:13 PM, Ilkka Ollakka <ileoo< at >videolan.org> wrote:

_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: "Race condition" in livehttp module

Ilkka Ollakka — 2013-06-14T07:30:02





That should be fixed already in 2.1 nightlies

Re: vlc head build fails < at > `make`: "error: too few arguments to function ‘vaCreateSurfaces’", with libavcodec 55.16.100

2013-06-14T04:00:40

I need _my_ newer libva in /usr/local/...

Not a problem, there's a 1.1.1 distro release in /usr/...

  rpm -qa | grep -i libva-devel
    libva-devel-1.1.0-6.1.1.x86_64

Unfortunately,  'pointing' to it by adding 

  export LIBVA_CFLAGS="-I/usr/include"
  export LIBVA_LIBS="-L/usr/lib64 -lva"

to the vlc build env fails to work.

The build fails in the same way, same error -- STILL pulling the VA
includes and libs from /usr/local

those flags SHOULD do the trick; need to dig as to why they're not
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: vlc head build fails < at > `make`: "error: too few arguments to function ‘vaCreateSurfaces’", with libavcodec 55.16.100

Masaru Nomiya — 2013-06-14T02:27:26

Hello,

In the Message; 

  Subject    : Re: [vlc-devel] vlc head build fails < at > `make`: "error: too few arguments to function ‘vaCreateSurfaces’", with libavcodec 55.16.100
  Message-ID : <1371176041.19094.140661243737593.6E647867< at >webmail.messagingengine.com>
  Date & Time: Thu, 13 Jun 2013 19:14:01 -0700

darx< at >sent.com has written:




It's too new.

Mine is 

$ pkg-config --modversion libva
0.33.0

Please install libva-1.1.1.tar.xz.

Regards,

---
┏━━┓彡 Masaru Nomiya             mail-to: nomiya < at > galaxy.dti.ne.jp
┃＼／彡
┗━━┛ "Bill! You married with Computers.
          Not with Me!"
         "No..., with money."
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel

Re: vlc head build fails < at > `make`: "error: too few arguments to function ‘vaCreateSurfaces’", with libavcodec 55.16.100

2013-06-14T02:14:01

reading < at >

  "Build Failure With VA-API 0.34.0"
   https://trac.videolan.org/vlc/ticket/8186

Iiuc, the problem's not libavcodec API version, but that of vaapi.

on my system,

pkg-config --modversion libva
  0.34.0

The bug's ~ 4 months old.  Is use of libva v0.34.0 in a VLC-master build
still a no-go?  Or is there a fix/workaround in a branch/fork somewhere?

--darx
_______________________________________________
vlc-devel mailing list
To unsubscribe or modify your subscription options:
http://mailman.videolan.org/listinfo/vlc-devel