Re: [xine-devel] Fwd: AAC in Me TV

[Paul Menzel <paulepanter@us...>]

Dear Michael, dear Franki,


Am Freitag, den 01.01.2010, 01:34 +1100 schrieb Michael Lamothe:
> Franki has sent me some patches to get better AAC support in xine-lib.
>  I develop Me TV, a DVB application for GNOME and these patches are
> claimed to help get better AAC support into xine-lib.  I have little
> knowledge of the decoders in xine-lib, are you guys able to have a
> look at the patches and apply them if suitable?

thank you for your work. I am not a Xine developer, so sorry if I am
telling you nonsense.

1. Should not the faad2 patch go to FAAD2 upstream [1]?
2. Franki, could you resend the xine-lib patch against tip of the hg
repository?
3. Franki, could you write a commit message (what problem does it fix,
what changed) for both patches and maybe format it with `hg email` [2]?
You will save the developers a lot of time doing this.
4. The faad2 patch seems to change by accident some comments (/* must be
�0000� */).
5. You seem to have marked your change additionally with a comment. I
guess this can not be committed this way. Could you resend the patch
without it?

[…]

> -  int              pts;
> +  int64_t          pts;                           // FRANKI  int -> int64_t

[…]


Thanks,

Paul


[1] http://www.audiocoding.com/faad2.html
[2] http://mercurial.selenic.com/wiki/PatchbombExtension

[xine-devel] Fwd: AAC in Me TV

[Michael Lamothe <michael.lamothe@gm...>]

Hi xine dudes,

Franki has sent me some patches to get better AAC support in xine-lib.
 I develop Me TV, a DVB application for GNOME and these patches are
claimed to help get better AAC support into xine-lib.  I have little
knowledge of the decoders in xine-lib, are you guys able to have a
look at the patches and apply them if suitable?

Cheers,

Michael


---------- Forwarded message ----------
From: Franki Jozsef <frankijoe@...>
Date: 2009/12/31
Subject: Re: AAC in Me TV
To: Michael Lamothe <michael.lamothe@...>


Hi,

It's working with xine-lib and faad2 patch. I attach these patches.
I've tried later you me-tv 1.1 branch (with AAC) and it's working also with
patched xine-lib.

Please, merge my "xine-config per channel" or make some "xine-config
for SD and HD" or similar, because limited CPU power and required
image quality need different settings (ffmpeg.loop.filter, etc) (also
for deinterlacing).

...

Cheers,
franki

[xine-devel] AAC (libfaad) patch

[Franki Jozsef <frankijoe@fr...>]

Hi,

I attach pacth for AAC audiotrack.

First in "xine_faad_decoder.c"   "pts"  type mismatch  int -> int64_t

Second in "xine-engine/audio_decoder.c"  enable change mapping of type
at same audio channel number, because mpeg-ts demuxer default type (MPEG)
inhibit AAC type when appear.

Need faad2 patch (external faad2) for enabling LATM frames (I attached also).

best regards,
franki

Re: [xine-devel] libxcb-shape.a

[Donn Washburn <n5xwb@co...>]

On 12/28/2009 08:23 AM, Darren Salt wrote:
> I demand that Donn Washburn may or may not have written...
>    
>> It seems that xine-1.1.17 needs
>>      
>> libxcb-shape.a   I have /usr/lib/libxcb-shape.so.0.0.0
>>      
>> libxcb which is not being used in newer version of Xorg. I find this
>> dropped in the development packages for openSuSE 11.3.
>>      
> Then those dev packages are, AFAICS, broken. If you have libxcb-shape.a, you
> should also have libxcb-shape.so. (BTW, which version? I have 1.5 here.)
>    
>> Do you have any plans to changes or drop this request,
>>      
> No, at least not without allowing for the unavailability of unscaled overlays
> (which doesn't seem like a good idea).
>    
I replaced xine's libtool with the one that was in openSuSE 11.3 and it 
made it almost to the end before libavcodec.h complained about 
depreciated line 3262 .

-- 
73 de Donn Washburn
307 Savoy Street     Email:" n5xwb@... "
Sugar Land, TX 77478 LL# 1.281.242.3256
Ham Callsign N5XWB   HAMs : " n5xwb@... "
VoIP via Gizmo: bmw_87kbike / via Skype: n5xwbg
BMW MOA #: 4146 - Ambassador
       " http://counter.li.org " #279316

Re: [xine-devel] libxcb-shape.a

[Darren Salt <linux@yo...>]

I demand that Donn Washburn may or may not have written...

> It seems that xine-1.1.17 needs

> libxcb-shape.a   I have /usr/lib/libxcb-shape.so.0.0.0

> libxcb which is not being used in newer version of Xorg. I find this
> dropped in the development packages for openSuSE 11.3.

Then those dev packages are, AFAICS, broken. If you have libxcb-shape.a, you
should also have libxcb-shape.so. (BTW, which version? I have 1.5 here.)

> Do you have any plans to changes or drop this request,

No, at least not without allowing for the unavailability of unscaled overlays
(which doesn't seem like a good idea).

-- 
| Darren Salt            | linux at youmustbejoking | nr. Ashington, | Doon
| using Debian GNU/Linux | or ds    ,demon,co,uk    | Northumberland | Army
| + http://www.youmustbejoking.demon.co.uk/ & http://tartarus.org/ds/

What is worth doing is worth overdoing.

[xine-devel] libxcb-shape.a

[Donn Washburn <n5xwb@co...>]

It seems that xine-1.1.17 needs

libxcb-shape.a   I have /usr/lib/libxcb-shape.so.0.0.0

libxcb which is not being used in newer version of Xorg.  I find this dropped in the development packages for openSuSE 11.3.  Do you have any plans to changes or drop this request,

-- 
73 de Donn Washburn
307 Savoy Street     Email:" n5xwb@... "
Sugar Land, TX 77478 LL# 1.281.242.3256
Ham Callsign N5XWB   HAMs : " n5xwb@... "
VoIP via Gizmo: bmw_87kbike / via Skype: n5xwbg
BMW MOA #: 4146 - Ambassador
       " http://counter.li.org " #279316

[xine-devel] How to compile without cdrom support?

[Andrey Chinkov <achinkov@sm...>]

I use Ubuntu-8.04 and want build xine-lib from sources.
I want to compile xine-lib without cdrom input support.
How should I configure xine-lib? I.e. what parameters and flags should be
passed to ./configure script?
Thanks in advance.

[xine-devel] I'm unable to initialize 'auto' video driver.

[Andrey Chinkov <achinkov@sm...>]

Hi. I am on Ubuntu-8.04.

I did the following steps consistently:

1) Download xine-lib-1.1.17.tar.bz2, unpack it to
/home/gosh/Desktop/xine-lib-1.
1.17

2) Install libxine with this commands:

cd /home/gosh/Desktop/xine-lib-1.1.17
./configure --prefix=/home/gosh/Desktop/xinelib_installed --disable-shared
--enable-static
make
make install

3) Download muxine.c (to /home/gosh/Desktop/xine_mini)

4) Compile muxine.c with this command:

cd /home/gosh/Desktop/xine_mini
gcc muxine.c -o xinemini -lxine -lrt -lz -ldl -lX11 -I
/home/gosh/Desktop/xinelib_installed/include -L
/home/gosh/Desktop/xinelib_installed/lib

5) Execute in terminal:

gosh@...:~/Desktop/xine_mini$ ./xinemini
/mnt/W_DRIVE/Trial/video/KMT_germans_net.mpg
muxine (xine � player) v0.0.6
(c) 2003 by Daniel Caujolle-Bert <f1rmb@...>.
I'm unable to initialize 'auto' video driver. Giving up.
gosh@...:~/Desktop/xine_mini$

My question is: How to resolve "I'm unable to initialize 'auto' video
driver. Giving up." message in muxine.c?

Thanks in advance.

[xine-devel] Crystal HD ( BCM70012 ) decoder plugin

[gimli <gimli@da...>]

Hello,

the decoder card got my attention since i discovered that there is
a private XBMC svn branch called "crystalhd".

I followed one forum thread in the the XBMC forums very close	:

http://blog.xbmc.org/forum/showthread.php?t=55042

About three till four weeks ago a leaked Linux driver and library
appeared on a vendor side who uses the decoder card in his products.

One accidental commit from "davilla" in the XBMC svn contained the
header files for the CrystalHD card.

This header files and the leaked driver/library made it possible
to start developing on Linux.

How to get the driver, library and header files can be found in
another XBMC forum thread :

http://xbmc.org/forum/showthread.php?t=62708

My development tree with an initial H264 video decoder plugin can
be found here :

https://sourceforge.net/apps/trac/archvdr/browser/branches/

The plugin is based on informations from the XBMC crystalhd branch.

http://xbmc.org/trac/browser/branches/crystalhd

Some functions and ideas for the H264 NAL splitting and converting
where taken from the vdpau hd tree.

http://hg.debian.org/hg/xine-lib/xine-lib-1.2-vdpau


cu

Edgar (gimli) Hucek

[xine-devel] [PATCH] TTA demux: use _x_demux_send_data

[Kelvie Wong <kelvie@ie...>]

# HG changeset patch
# User Kelvie Wong <kelvie@...>
# Date 1260952072 28800
# Node ID 957e20fda6281198420048b162ee9542bde2d6fc
# Parent  08ca808df71c5b3cf7009daebdaf5f2a973ef75d
TTA demux: use _x_demux_send_data

To reduce code duplication.

diff -r 08ca808df71c -r 957e20fda628 src/demuxers/demux_tta.c
--- a/src/demuxers/demux_tta.c	Tue Dec 15 20:49:16 2009 -0800
+++ b/src/demuxers/demux_tta.c	Wed Dec 16 00:27:52 2009 -0800
@@ -122,39 +122,17 @@
 
   bytes_to_read = le2me_32(this->seektable[this->currentframe]);
 
-  while(bytes_to_read) {
-    off_t bytes_read = 0;
-    buf_element_t *buf = NULL;
-
-    /* Get a buffer */
-    buf = this->audio_fifo->buffer_pool_alloc(this->audio_fifo);
-    buf->type = BUF_AUDIO_TTA;
-    buf->pts = 0;
-    buf->extra_info->total_time = (int)(le2me_32(this->header.tta.data_length) * 1000.0 / le2me_32(this->header.tta.samplerate)); /* milliseconds */ 
-    buf->decoder_flags = 0;
-
-    /* Set normalised position */
-    buf->extra_info->input_normpos =
-      (int) ((double) this->currentframe * 65535 / this->totalframes);
-
-    /* Set time */
-    buf->extra_info->input_time = (int)(FRAME_TIME * this->currentframe)*1000;
-
-    bytes_read = this->input->read(this->input, buf->content, ( bytes_to_read > buf->max_size ) ? buf->max_size : bytes_to_read);
-    if (bytes_read < 0) {
-      this->status = DEMUX_FINISHED;
-      break;
-    }
-
-    buf->size = bytes_read;
-
-    bytes_to_read -= bytes_read;
-
-    if ( bytes_to_read <= 0 )
-      buf->decoder_flags |= BUF_FLAG_FRAME_END;
-
-    this->audio_fifo->put(this->audio_fifo, buf);
-  }
+  _x_demux_read_send_data(this->audio_fifo,
+                          this->input,
+                          bytes_to_read,
+                          /*pts*/ 0,
+                          BUF_AUDIO_TTA,
+                          /*decoder_flags*/ 0,
+                          (int) ((double) this->currentframe * 65535.0 / this->totalframes),
+                          (int)(FRAME_TIME * this->currentframe)*1000,
+                          (int)(le2me_32(this->header.tta.data_length) * 1000.0 /
+                                le2me_32(this->header.tta.samplerate)),
+                          this->currentframe);
 
   this->currentframe++;
 

Re: [xine-devel] [PATCH] TTA demux: Fix a buffer overflow

[Kelvie Wong <kelvie@ie...>]

Also, a quick glance at the other demuxes tells me that this isn't
checked for many of the other ones, as well, so someone may want to
look into that.

-- 
Kelvie Wong

[xine-devel] [PATCH] TTA demux: Fix the calculation of the stream length

[Kelvie Wong <kelvie@ie...>]

# HG changeset patch
# User Kelvie Wong <kelvie@...>
# Date 1260938956 28800
# Node ID 08ca808df71c5b3cf7009daebdaf5f2a973ef75d
# Parent  63a2cfa17ff38e03a34b704a08613cc4a7085e79
TTA demux: Fix the calculation of the stream length.

Its resolution was in frames (+/- ~1 second), now it is calculated from the
number of samples, as it should be.

diff -r 63a2cfa17ff3 -r 08ca808df71c src/demuxers/demux_tta.c
--- a/src/demuxers/demux_tta.c	Sun Dec 13 03:28:16 2009 -0800
+++ b/src/demuxers/demux_tta.c	Tue Dec 15 20:49:16 2009 -0800
@@ -64,7 +64,7 @@
       uint16_t channels;
       uint16_t bits_per_sample;
       uint32_t samplerate;
-      uint32_t data_length;
+      uint32_t data_length; /* Number of samples */
       uint32_t crc32;
     } XINE_PACKED tta;
     uint8_t buffer[22]; /* This is the size of the header */
@@ -130,7 +130,7 @@
     buf = this->audio_fifo->buffer_pool_alloc(this->audio_fifo);
     buf->type = BUF_AUDIO_TTA;
     buf->pts = 0;
-    buf->extra_info->total_time = (int)(FRAME_TIME * this->totalframes)*1000;
+    buf->extra_info->total_time = (int)(le2me_32(this->header.tta.data_length) * 1000.0 / le2me_32(this->header.tta.samplerate)); /* milliseconds */ 
     buf->decoder_flags = 0;
 
     /* Set normalised position */
@@ -281,7 +281,7 @@
 
 static int demux_tta_get_stream_length (demux_plugin_t *this_gen) {
   demux_tta_t *this = (demux_tta_t *) this_gen;
-  return (int)(FRAME_TIME * this->totalframes * 1000);
+  return le2me_32(this->header.tta.data_length) * 1000.0 / le2me_32(this->header.tta.samplerate); /* milliseconds */
 }
 
 static uint32_t demux_tta_get_capabilities(demux_plugin_t *this_gen) {

[xine-devel] [PATCH] TTA demux: Fix a buffer overflow

[Kelvie Wong <kelvie@ie...>]

# HG changeset patch
# User Kelvie Wong <kelvie@...>
# Date 1260703696 28800
# Node ID 63a2cfa17ff38e03a34b704a08613cc4a7085e79
# Parent  0546197b637ecd44fcea4ff4ddbb2e333f03e845
TTA demux: Fix a buffer overflow

It can be the case that the header is larger than buf->max_size (8 kilobytes),
especially for slightly larger files.  This sends them in parts so we don't
overfill the buffer.

diff -r 0546197b637e -r 63a2cfa17ff3 src/demuxers/demux_tta.c
--- a/src/demuxers/demux_tta.c	Fri Dec 11 19:28:37 2009 -0800
+++ b/src/demuxers/demux_tta.c	Sun Dec 13 03:28:16 2009 -0800
@@ -164,6 +164,12 @@
 static void demux_tta_send_headers(demux_plugin_t *this_gen) {
   demux_tta_t *this = (demux_tta_t *) this_gen;
   buf_element_t *buf;
+  xine_waveformatex wave;
+  uint32_t total_size = sizeof(xine_waveformatex) + sizeof(this->header) +
+    sizeof(uint32_t)*this->totalframes;
+  unsigned char *header;
+
+  header = malloc(total_size);
 
   this->audio_fifo  = this->stream->audio_fifo;
 
@@ -181,27 +187,43 @@
   /* send start buffers */
   _x_demux_control_start(this->stream);
 
+  /* create header */
+  wave.cbSize = total_size - sizeof(xine_waveformatex);
+
+  memcpy(header, &wave, sizeof(wave));
+  memcpy(header+sizeof(xine_waveformatex), this->header.buffer, sizeof(this->header));
+  memcpy(header+sizeof(xine_waveformatex)+sizeof(this->header), this->seektable, sizeof(uint32_t)*this->totalframes);
+
   /* send init info to decoders */
   if (this->audio_fifo) {
-    xine_waveformatex wave;
+    uint32_t bytes_left = total_size;
 
-    buf = this->audio_fifo->buffer_pool_alloc (this->audio_fifo);
-    buf->type = BUF_AUDIO_TTA;
-    buf->decoder_flags = BUF_FLAG_HEADER|BUF_FLAG_STDHEADER|BUF_FLAG_FRAME_END;
-    buf->decoder_info[0] = 0;
-    buf->decoder_info[1] = le2me_32(this->header.tta.samplerate);
-    buf->decoder_info[2] = le2me_16(this->header.tta.bits_per_sample);
-    buf->decoder_info[3] = le2me_16(this->header.tta.channels);
+    /* We are sending the seektable as well, and this may be larger than
+       buf->max_size */
+    while (bytes_left) {
+      buf = this->audio_fifo->buffer_pool_alloc (this->audio_fifo);
+      buf->decoder_flags = BUF_FLAG_HEADER|BUF_FLAG_STDHEADER;
+      buf->type = BUF_AUDIO_TTA;
 
-    buf->size = sizeof(xine_waveformatex) + sizeof(this->header) + sizeof(uint32_t)*this->totalframes;
-    memcpy(buf->content+sizeof(xine_waveformatex), this->header.buffer, sizeof(this->header));
-    memcpy(buf->content+sizeof(xine_waveformatex)+sizeof(this->header), this->seektable, sizeof(uint32_t)*this->totalframes);
+      /* Copy min(bytes_left, max_size) bytes */
+      buf->size = bytes_left < buf->max_size ? bytes_left : buf->max_size;
+      memcpy(buf->content, header+(total_size-bytes_left), buf->size);
 
-    wave.cbSize = buf->size - sizeof(xine_waveformatex);
-    memcpy(buf->content, &wave, sizeof(wave));
+      bytes_left -= buf->size;
 
-    this->audio_fifo->put (this->audio_fifo, buf);
+      /* The decoder information only needs the decoder information on the last
+         buffer element. */
+      if (!bytes_left) {
+        buf->decoder_flags |= BUF_FLAG_FRAME_END;
+        buf->decoder_info[0] = 0;
+        buf->decoder_info[1] = le2me_32(this->header.tta.samplerate);
+        buf->decoder_info[2] = le2me_16(this->header.tta.bits_per_sample);
+        buf->decoder_info[3] = le2me_16(this->header.tta.channels);
+      }
+      this->audio_fifo->put (this->audio_fifo, buf);
+    }
   }
+  free(header);
 }
 
 static int demux_tta_seek (demux_plugin_t *this_gen,

Re: [xine-devel] [PATCH 0 of 2] A few more TTA demux fixes.

[Kelvie Wong <kelvie@ie...>]

2009/12/12 Darren Salt <linux@...>:
> I demand that Kelvie Wong may or may not have written...
>
>> These patches fix a couple of minor thins..
>
> These patches are showing signs of 'git format-patch | hg import'. Not
> difficult to clean up, fortunately...
>

Ahh, the patch stats, right?  I'll get rid of those from commits from now on.

>> Also, when I set buf->pts (while sending chunks) to the proper value,
>> seeking stops working, is that normal?  I haven't looked into the quirks of
>> xine yet, but at least in Amarok (Phonon), it stops the seek from working.
>
> Hmm. Interesting. Not sure about that; could be a demuxer issue, but there
> are others who know more...
>
>> Also, while I fixed the lastframe problem (which caused some TTA files to
>> crash xine), there is still some noise at the end of the file, though I
>> suspect this may be Amarok trying to crossfade.
>
> Not being an Amarok user, I wouldn't know. Switch that off and see what
> happens...?
>

There's no GUI option to remove crossfading, yet, but I'll either
disable it (in code if I have to) and try again.

-- 
Kelvie Wong

Re: [xine-devel] Forbidden to access parts of Bugzilla

[Darren Salt <linux@yo...>]

I demand that lorenzodes may or may not have written...

> Getting "Forbidden You don't have permission to access /attachment.cgi on
> this server." while trying to upload a patch :(

Seeing the same here. Apparently, I've been using an automated program to
explore the site and I've been sending possibly-bad MIME multipart data (when
trying to add an attachment) and I've been doing SQL injection attacks (when
trying to get a bug list).

Not guilty on all counts. mod_security, as currently configured, is being far
too aggressive; I'm inclined to call it mod_DoS.

-- 
| Darren Salt            | linux at youmustbejoking | nr. Ashington, | Doon
| using Debian GNU/Linux | or ds    ,demon,co,uk    | Northumberland | Army
| + http://www.youmustbejoking.demon.co.uk/ & http://tartarus.org/ds/

I love treason but hate a traitor.

Re: [xine-devel] [PATCH 0 of 2] A few more TTA demux fixes.

[Darren Salt <linux@yo...>]

I demand that Kelvie Wong may or may not have written...

> These patches fix a couple of minor thins..

These patches are showing signs of 'git format-patch | hg import'. Not
difficult to clean up, fortunately...

> Also, when I set buf->pts (while sending chunks) to the proper value,
> seeking stops working, is that normal?  I haven't looked into the quirks of
> xine yet, but at least in Amarok (Phonon), it stops the seek from working.

Hmm. Interesting. Not sure about that; could be a demuxer issue, but there
are others who know more...

> Also, while I fixed the lastframe problem (which caused some TTA files to
> crash xine), there is still some noise at the end of the file, though I
> suspect this may be Amarok trying to crossfade.

Not being an Amarok user, I wouldn't know. Switch that off and see what
happens...?

-- 
| Darren Salt            | linux at youmustbejoking | nr. Ashington, | Doon
| using Debian GNU/Linux | or ds    ,demon,co,uk    | Northumberland | Army
| + http://www.youmustbejoking.demon.co.uk/ & http://tartarus.org/ds/

Never invest in anything that eats.

[xine-devel] [PATCH 2 of 2] [PATCH 1/2] TTA demux: don't read past the last frame

[Kelvie Wong <kelvie@ie...>]

# HG changeset patch
# User Kelvie Wong <kelvie@...>
# Date 1260588519 28800
# Node ID b3156e3df0e4b5645aba1f0c9d536457b5ee639a
# Parent  ef5c9c0cbedba224540bcdfa922a57e61adcb0a5
[PATCH 1/2] TTA demux: don't read past the last frame.

>From b08a19a886e30fd62fb7c174242af5ba7e05ba2d Mon Sep 17 00:00:00 2001
Date: Thu, 10 Dec 2009 21:13:29 -0800
I don't think whoever wrote this played a TTA file all the way to the end.
---
 src/demuxers/demux_tta.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff -r ef5c9c0cbedb -r b3156e3df0e4 src/demuxers/demux_tta.c
--- a/src/demuxers/demux_tta.c	Fri Dec 11 19:28:37 2009 -0800
+++ b/src/demuxers/demux_tta.c	Fri Dec 11 19:28:39 2009 -0800
@@ -115,7 +115,7 @@
   demux_tta_t *this = (demux_tta_t *) this_gen;
   uint32_t bytes_to_read;
 
-  if ( this->currentframe > this->totalframes ) {
+  if ( this->currentframe >= this->totalframes ) {
     this->status = DEMUX_FINISHED;
     return this->status;
   }

[xine-devel] [PATCH 1 of 2] [PATCH 2/2] TTA demux: fix buffer->extra_info->total_time

[Kelvie Wong <kelvie@ie...>]

# HG changeset patch
# User Kelvie Wong <kelvie@...>
# Date 1260588517 28800
# Node ID ef5c9c0cbedba224540bcdfa922a57e61adcb0a5
# Parent  4082bc062c361c73f84bc630c84def77e4e45f67
[PATCH 2/2] TTA demux: fix buffer->extra_info->total_time

>From e5c3063bc95aa92076b9f971d49c5fafd0c79e80 Mon Sep 17 00:00:00 2001
Date: Thu, 10 Dec 2009 21:14:26 -0800
Supposed to be in milliseconds; totalframes is only approx. total seconds.
---
 src/demuxers/demux_tta.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff -r 4082bc062c36 -r ef5c9c0cbedb src/demuxers/demux_tta.c
--- a/src/demuxers/demux_tta.c	Fri Dec 11 23:03:42 2009 +0000
+++ b/src/demuxers/demux_tta.c	Fri Dec 11 19:28:37 2009 -0800
@@ -130,7 +130,7 @@
     buf = this->audio_fifo->buffer_pool_alloc(this->audio_fifo);
     buf->type = BUF_AUDIO_TTA;
     buf->pts = 0;
-    buf->extra_info->total_time = this->totalframes;
+    buf->extra_info->total_time = (int)(FRAME_TIME * this->totalframes)*1000;
     buf->decoder_flags = 0;
 
     /* Set normalised position */

[xine-devel] [PATCH 0 of 2] A few more TTA demux fixes.

[Kelvie Wong <kelvie@ie...>]

These patches fix a couple of minor thins..

Also, when I set buf->pts (while sending chunks) to the proper value, seeking
stops working, is that normal?  I haven't looked into the quirks of xine yet,
but at least in Amarok (Phonon), it stops the seek from working.

Also, while I fixed the lastframe problem (which caused some TTA files to crash
xine), there is still some noise at the end of the file, though I suspect this
may be Amarok trying to crossfade.

Re: [xine-devel] [PATCH]es A small memory leak fix and finish implementing the TTA demuxer.

[Darren Salt <linux@yo...>]

I demand that Kelvie Wong may or may not have written...

> 2009/12/6 Kelvie Wong <kelvie@...>:
>> I'm not sure if this is all right (as I understand it, pts is in secs
>> * 90000?).

Yes.

>> Seeking, and returning proper stream lengths and the current time is
>> implemented; I've tested it using Phonon (via the minimal Dragon media
>> player, and in Amarok), and it seeks perfectly, but for some reason,
>> using the xine GUI, it doesn't seek very well.

>> I've noticed that FLAC files don't really seek well in xine-ui either,
>> this may or may not be related.

I should hope not, since it's not the FLAC files which should be doing the
seeking... :-)

>> Also, the first patch in this series fixes a couple of memory leaks (I
>> think) as I glanced over it.  I didn't check it.

Applied, but "hg email" is preferred.

[snip]
> Just as an update (if anyone cares), I have a few other TTA fixes on the
> way... unless I'm horribly mistaken, this demuxer was written quite
> half-assed (but better than nothing, I guess).

Well... send the patches and they'll at least be looked at and, quite
probably, applied too...

-- 
| Darren Salt            | linux at youmustbejoking | nr. Ashington, | Doon
| using Debian GNU/Linux | or ds    ,demon,co,uk    | Northumberland | Army
| + http://www.youmustbejoking.demon.co.uk/ & http://tartarus.org/ds/

I may not always be perfect, but I'm always me.

Re: [xine-devel] [PATCH]es A small memory leak fix and finish implementing the TTA demuxer.

[Kelvie Wong <kelvie@ie...>]

2009/12/6 Kelvie Wong <kelvie@...>:
> I'm not sure if this is all right (as I under stand it, pts is in secs
> * 90000?).
>
> Seeking, and returning proper stream lengths and the current time is
> implemented; I've tested it using Phonon (via the minimal Dragon media
> player, and in Amarok), and it seeks perfectly, but for some reason,
> using the xine GUI, it doesn't seek very well.
>
> I've noticed that FLAC files don't really seek well in xine-ui either,
> this may or may not be related.
>
> Also, the first patch in this series fixes a couple of memory leaks (I
> think) as I glanced over it.  I didn't check it.
>
> Any comments/criticism would be great.  It's pretty late at night
> here, so I'm sure I did something stupid.
>
> --
> Kelvie Wong
>

Just as an update (if anyone cares), I have a few other TTA fixes on
the way... unless I'm horribly mistaken, this demuxer was written
quite half-assed (but better than nothing, I guess).


-- 
Kelvie Wong

Re: [xine-devel] Forbidden to access parts of Bugzilla

[lorenzodes <lorenzodes@fa...>]

Diego Elio “Flameeyes” Pettenò ha scritto:
> Il giorno Wed, 18/11/2009 alle 22.49 +0000, Darren Salt ha scritto:
>>
>>>> Forbidden
>>>> You don't have permission to access /buglist.cgi on this server.
>>>> ----------------------------------------------------------------
>>>> Server at bugs.xine-project.org Port 80
>> Looks like '&' no longer works for parameter separation. ';' works
>> fine,
>> though...
>>
> Sorry, a stupid mod_security rule I forgot to disable :( fixed now!
> 


Getting "Forbidden You don't have permission to access /attachment.cgi 
on this server." while trying to upload a patch :(

[xine-devel] switching audio channels in xine-ui

[Uwe Bugla <uwe.bugla@gm...>]

Hi,

when xine-ui addresses a Video DVD containing more than one audio tracks
in different languages you can easily switch between several audio
tracks.

My idea / deapest wish would be to integrate a similar audio track
switching possibility for xine-ui addressing DVB TV.

As you all know there are several digital TV stations being spread
that are bilingual or even trilingual.
At the same time xine-ui is definitely capable of addressing a
channels.conf in vdr format (as opposed to mplayer which is NOT capable
of doing that).

But if you address a channels.conf in vdr format which owes more than
one audio PID separated by a comma using xine-ui, only the first audio
PID is automatically taken.
No chance to change / switch the audio channel of the broadcast!

Somewhere in the source code I read that the audio PID isn't even needed
to correctly address a channels.conf with xine-ui.
That complicates my desire enormously, so I do not have any idea where I
could at least try to imply something like a patchwork.

I would appreciate someone reading that being technically more capable
than I am to offer us a patch for making this audio channel switching
function available.
Please do not offer me nonsense answers like "use kaffeine please!" or
similar. I hate KDE and that's why I want to avoid things like kaffeine
and look DVB TV with xine-ui instead without losing functionalities.

Above from that a command line switch for producing mpeg PES files (i.
e. recording function) would also be a very good idea for xine-ui.
At the current state xine can record unusable mpeg TS files which no
cutting software can make use of.
So there should be mpeg PS streams as output instead of that, shouldn't
there? Best with command line switch.

Can someone help and offer patches please?

Regards

Uwe

[xine-devel] switching audio channels in xine-ui

[Uwe Bugla <uwe.bugla@gm...>]

Hi,

when xine-ui addresses a Video DVD containing more than one audio tracks
in different languages you can easily switch between several audio
tracks.

My idea / deapest wish would be to integrate a similar audio track
switching possibility for xine-ui addressing DVB TV.

As you all know there are several digital TV stations being spread
that are bilingual or even trilingual.
At the same time xine-ui is definitely capable of addressing a
channels.conf in vdr format (as opposed to mplayer which is NOT capable
of doing that).

But if you address a channels.conf in vdr format which owes more than
one audio PID separated by a comma using xine-ui, only the first audio
PID is automatically taken.
No chance to change / switch the audio channel of the broadcast!

Somewhere in the source code I read that the audio PID isn't even needed
to correctly address a channels.conf with xine-ui.
That complicates my desire enormously, so I do not have any idea where I
could at least try to imply something like a patchwork.

I would appreciate someone reading that being technically more capable
than I am to offer us a patch for making this audio channel switching
function available.
Please do not offer me nonsense answers like "use kaffeine please!" or
similar. I hate KDE and that's why I want to avoid things like kaffeine
and look DVB TV with xine-ui instead without losing functionalities.

Above from that a command line switch for producing mpeg PES files (i.
e. recording function) would also be a very good idea for xine-ui.
At the current state xine can record unusable mpeg TS files which no
cutting software can make use of.
So there should be mpeg PS streams as output instead of that, shouldn't
there? Best with command line switch.

Can someone help and offer patches please?

Regards

Uwe

[xine-devel] [PATCH]es A small memory leak fix and finish implementing the TTA demuxer.

[Kelvie Wong <kelvie@ie...>]

I'm not sure if this is all right (as I under stand it, pts is in secs
* 90000?).

Seeking, and returning proper stream lengths and the current time is
implemented; I've tested it using Phonon (via the minimal Dragon media
player, and in Amarok), and it seeks perfectly, but for some reason,
using the xine GUI, it doesn't seek very well.

I've noticed that FLAC files don't really seek well in xine-ui either,
this may or may not be related.

Also, the first patch in this series fixes a couple of memory leaks (I
think) as I glanced over it.  I didn't check it.

Any comments/criticism would be great.  It's pretty late at night
here, so I'm sure I did something stupid.

-- 
Kelvie Wong