SBo scripts not building on current (read 1st post, pls)

kingbeowulf · 03-09-2016, 10:30 AM

Quote:

Originally Posted by mats_b_tegner

My guess is that nvidia-legacy* is broken as well. I have written an email to nVidia.

Being "legacy" I doubt Nvidia will worry about it. As I mentioned in the email to, I might drop nvidia-installer: It is only used by the Nvidia install script and thus is not useable for the SBo script and pkgtools. It's left over cruft from when I took over Nvidia maintenance.

mats_b_tegner · 03-09-2016, 02:08 PM

Quote:

Originally Posted by kingbeowulf

Being "legacy" I doubt Nvidia will worry about it. As I mentioned in the email to, I might drop nvidia-installer: It is only used by the Nvidia install script and thus is not useable for the SBo script and pkgtools. It's left over cruft from when I took over Nvidia maintenance.

Okay, lets remove nvidia-installer from the SlackBuild-script then, if it doesn't serve a purpose.

ecoslacker · 03-09-2016, 05:00 PM

It looks like the link to scipy SlackBuild is broken, isn't or is it just me?

suppy · 03-09-2016, 05:03 PM

Quote:

Originally Posted by ecoslacker

It looks like the link to scipy SlackBuild is broken, isn't or is it just me?

It is for me too. You can try grabbing it from ftp://ftp.slackware.org.uk/sbosrcarch/ which hosts a backup of many of the source files linked to from SBo.

ponce · 03-09-2016, 05:10 PM

thanks for reporting it: I fixed the scipy download link on my git branch, will go in at the next merge.

ecoslacker · 03-09-2016, 05:18 PM

Thanks that was useful. I hope the link to the source of scipy gets fixed soon.

USUARIONUEVO · 03-09-2016, 06:15 PM

afpfs-ng need patches and fix to make inside the package the include files.

see archlinux to patches
https://projects.archlinux.org/svnto...kages/afpfs-ng

and after make install

Quote:

# Copy the header files
cd include
for header in afpfs-ng/*.h; do
install -Dm644 "$header" "$PKG/usr/include/$header"
done
cd ..

willysr · 03-09-2016, 09:58 PM

I updated the afpfs-ng to include the headers, but do we need all those patches?

USUARIONUEVO · 03-09-2016, 10:12 PM

Quote:

Originally Posted by willysr

I updated the afpfs-ng to include the headers, but do we need all those patches?

Im not really sure , but archlinux patch de sources ..

reason?

i dont know.

EDIT: i go to download and test your build ... headers are required by for example medusa. (not in SBo)

EDIT2: cant test , not in git , when script are in git i test to rebuild main package wih your script and test later if medusa enble afpfs in configure time.

EDIT3: hydra ,is another want this afpfs

Checking for AFP (libafpclient.so) ...
... found

USUARIONUEVO · 03-10-2016, 04:17 AM

Sorry for repost , but i found fcrackzip overwrite a original file from unzip package from slackware.

http://www.linuxquestions.org/questi...nt-4175574483/

the slackbuild for fcrackzip, after install , need to move zipinfo to fzipinfo ,for non overwrite the original slackware executable , or break unzip.

After install DESTDIR=

Quote:

# No overwrite the original zipinfo from slackware
mv "$PKG/usr/bin/zipinfo" "$PKG/usr/bin/fzipinfo"

willysr · 03-10-2016, 06:00 AM

Thanks. I have pushed an update for fcrackzip

orbea · 03-10-2016, 08:23 AM

Since I have found myself a bit ahead of SBo, here are two patches for building with ffmpeg.3.0 which might help when the time comes. I figure they will be more useful here than on my hdd.

Alephone:
https://github.com/Aleph-One-Maratho...f41833e9bfc048

Code:

diff --git a/Source_Files/FFmpeg/Movie.cpp b/Source_Files/FFmpeg/Movie.cpp
index 3b59853..a0074a7 100644
--- a/Source_Files/FFmpeg/Movie.cpp
+++ b/Source_Files/FFmpeg/Movie.cpp
@@ -420,7 +420,7 @@ bool Movie::Setup()
         video_stream->codec->width = view_rect.w;
         video_stream->codec->height = view_rect.h;
         video_stream->codec->time_base = (AVRational){1, TICKS_PER_SECOND};
-        video_stream->codec->pix_fmt = PIX_FMT_YUV420P;
+        video_stream->codec->pix_fmt = AV_PIX_FMT_YUV420P;
         video_stream->codec->flags |= CODEC_FLAG_CLOSED_GOP;
         video_stream->codec->thread_count = get_cpu_count();
         
@@ -539,7 +539,7 @@ bool Movie::Setup()
     // initialize conversion context
     if (success)
     {
-        av->sws_ctx = sws_getContext(temp_surface->w, temp_surface->h, PIX_FMT_RGB32,
+        av->sws_ctx = sws_getContext(temp_surface->w, temp_surface->h, AV_PIX_FMT_RGB32,
                                      video_stream->codec->width,
                                      video_stream->codec->height,
                                      video_stream->codec->pix_fmt,
diff --git a/Source_Files/FFmpeg/SDL_ffmpeg.c b/Source_Files/FFmpeg/SDL_ffmpeg.c
index 609c908..203b049 100644
--- a/Source_Files/FFmpeg/SDL_ffmpeg.c
+++ b/Source_Files/FFmpeg/SDL_ffmpeg.c
@@ -86,7 +86,7 @@ extern int convert_audio(int in_samples, int in_channels, int in_stride,
  *  Provide a fast way to get the correct context.
  *  \returns The context matching the input values.
  */
-struct SwsContext* getContext( SDL_ffmpegConversionContext **context, int inWidth, int inHeight, enum PixelFormat inFormat, int outWidth, int outHeight, enum PixelFormat outFormat )
+struct SwsContext* getContext( SDL_ffmpegConversionContext **context, int inWidth, int inHeight, enum AVPixelFormat inFormat, int outWidth, int outHeight, enum AVPixelFormat outFormat )
 {
     SDL_ffmpegConversionContext *ctx = *context;
 
@@ -621,7 +621,7 @@ int SDL_ffmpegAddVideoFrame( SDL_ffmpegFile *file, SDL_Surface *frame )
     {
         case 24:
             sws_scale( getContext( &file->videoStream->conversionContext,
-                                   frame->w, frame->h, PIX_FMT_RGB24,
+                                   frame->w, frame->h, AV_PIX_FMT_RGB24,
                                    file->videoStream->_ffmpeg->codec->width,
                                    file->videoStream->_ffmpeg->codec->height,
                                    file->videoStream->_ffmpeg->codec->pix_fmt ),
@@ -634,7 +634,7 @@ int SDL_ffmpegAddVideoFrame( SDL_ffmpegFile *file, SDL_Surface *frame )
             break;
         case 32:
             sws_scale( getContext( &file->videoStream->conversionContext,
-                                   frame->w, frame->h, PIX_FMT_BGR32,
+                                   frame->w, frame->h, AV_PIX_FMT_BGR32,
                                    file->videoStream->_ffmpeg->codec->width,
                                    file->videoStream->_ffmpeg->codec->height,
                                    file->videoStream->_ffmpeg->codec->pix_fmt ),
@@ -1605,7 +1605,7 @@ SDL_ffmpegStream* SDL_ffmpegAddVideoStream( SDL_ffmpegFile *file, SDL_ffmpegCode
     stream->codec->gop_size = 12;
 
     /* set pixel format */
-    stream->codec->pix_fmt = PIX_FMT_YUV420P;
+    stream->codec->pix_fmt = AV_PIX_FMT_YUV420P;
 
     /* set mpeg2 codec parameters */
     if ( stream->codec->codec_id == AV_CODEC_ID_MPEG2VIDEO )
@@ -2223,7 +2223,7 @@ int SDL_ffmpegDecodeVideoFrame( SDL_ffmpegFile* file, AVPacket *pack, SDL_ffmpeg
                                    file->videoStream->_ffmpeg->codec->height,
                                    file->videoStream->_ffmpeg->codec->pix_fmt,
                                    frame->overlay->w, frame->overlay->h,
-                                   PIX_FMT_YUYV422 ),
+                                   AV_PIX_FMT_YUYV422 ),
                        ( const uint8_t* const* )file->videoStream->decodeFrame->data,
                        file->videoStream->decodeFrame->linesize,
                        0,
@@ -2245,7 +2245,7 @@ int SDL_ffmpegDecodeVideoFrame( SDL_ffmpegFile* file, AVPacket *pack, SDL_ffmpeg
                                            file->videoStream->_ffmpeg->codec->height,
                                            file->videoStream->_ffmpeg->codec->pix_fmt,
                                            frame->surface->w, frame->surface->h,
-                                           PIX_FMT_RGB32 ),
+                                           AV_PIX_FMT_RGB32 ),
                                ( const uint8_t* const* )file->videoStream->decodeFrame->data,
                                file->videoStream->decodeFrame->linesize,
                                0,
@@ -2259,7 +2259,7 @@ int SDL_ffmpegDecodeVideoFrame( SDL_ffmpegFile* file, AVPacket *pack, SDL_ffmpeg
                                            file->videoStream->_ffmpeg->codec->height,
                                            file->videoStream->_ffmpeg->codec->pix_fmt,
                                            frame->surface->w, frame->surface->h,
-                                           PIX_FMT_RGB24 ),
+                                           AV_PIX_FMT_RGB24 ),
                                ( const uint8_t* const* )file->videoStream->decodeFrame->data,
                                file->videoStream->decodeFrame->linesize,
                                0,

OpenSceneGraph (From arch):

Code:

Description: Replace deprecated FFmpeg API
Author: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Last-Update: <2015-11-02>

--- openscenegraph-3.2.1.orig/OpenSceneGraph/src/osgPlugins/ffmpeg/FFmpegDecoderVideo.cpp
+++ openscenegraph-3.2.1/OpenSceneGraph/src/osgPlugins/ffmpeg/FFmpegDecoderVideo.cpp
@@ -71,7 +71,7 @@ void FFmpegDecoderVideo::open(AVStream *
     findAspectRatio();
 
     // Find out whether we support Alpha channel
-    m_alpha_channel = (m_context->pix_fmt == PIX_FMT_YUVA420P);
+    m_alpha_channel = (m_context->pix_fmt == AV_PIX_FMT_YUVA420P);
 
     // Find out the framerate
     m_frame_rate = av_q2d(stream->avg_frame_rate);
@@ -91,20 +91,19 @@ void FFmpegDecoderVideo::open(AVStream *
         throw std::runtime_error("avcodec_open() failed");
 
     // Allocate video frame
-    m_frame.reset(avcodec_alloc_frame());
+    m_frame.reset(av_frame_alloc());
 
     // Allocate converted RGB frame
-    m_frame_rgba.reset(avcodec_alloc_frame());
-    m_buffer_rgba[0].resize(avpicture_get_size(PIX_FMT_RGB24, width(), height()));
+    m_frame_rgba.reset(av_frame_alloc());
+    m_buffer_rgba[0].resize(avpicture_get_size(AV_PIX_FMT_RGB24, width(), height()));
     m_buffer_rgba[1].resize(m_buffer_rgba[0].size());
 
     // Assign appropriate parts of the buffer to image planes in m_frame_rgba
-    avpicture_fill((AVPicture *) (m_frame_rgba).get(), &(m_buffer_rgba[0])[0], PIX_FMT_RGB24, width(), height());
+    avpicture_fill((AVPicture *) (m_frame_rgba).get(), &(m_buffer_rgba[0])[0], AV_PIX_FMT_RGB24, width(), height());
 
     // Override get_buffer()/release_buffer() from codec context in order to retrieve the PTS of each frame.
     m_context->opaque = this;
-    m_context->get_buffer = getBuffer;
-    m_context->release_buffer = releaseBuffer;
+    m_context->get_buffer2 = getBuffer;
 }
 
 
@@ -263,8 +262,8 @@ int FFmpegDecoderVideo::convert(AVPictur
 #ifdef USE_SWSCALE
     if (m_swscale_ctx==0)
     {
-        m_swscale_ctx = sws_getContext(src_width, src_height, (PixelFormat) src_pix_fmt,
-                                      src_width, src_height, (PixelFormat) dst_pix_fmt,
+        m_swscale_ctx = sws_getContext(src_width, src_height, (AVPixelFormat) src_pix_fmt,
+                                      src_width, src_height, (AVPixelFormat) dst_pix_fmt,
                                       /*SWS_BILINEAR*/ SWS_BICUBIC, NULL, NULL, NULL);
     }
 
@@ -311,14 +310,14 @@ void FFmpegDecoderVideo::publishFrame(co
     AVPicture * const dst = (AVPicture *) m_frame_rgba.get();
 
     // Assign appropriate parts of the buffer to image planes in m_frame_rgba
-    avpicture_fill((AVPicture *) (m_frame_rgba).get(), &(m_buffer_rgba[m_writeBuffer])[0], PIX_FMT_RGB24, width(), height());
+    avpicture_fill((AVPicture *) (m_frame_rgba).get(), &(m_buffer_rgba[m_writeBuffer])[0], AV_PIX_FMT_RGB24, width(), height());
 
     // Convert YUVA420p (i.e. YUV420p plus alpha channel) using our own routine
 
-    if (m_context->pix_fmt == PIX_FMT_YUVA420P)
+    if (m_context->pix_fmt == AV_PIX_FMT_YUVA420P)
         yuva420pToRgba(dst, src, width(), height());
     else
-        convert(dst, PIX_FMT_RGB24, src, m_context->pix_fmt, width(), height());
+        convert(dst, AV_PIX_FMT_RGB24, src, m_context->pix_fmt, width(), height());
 
     // Wait 'delay' seconds before publishing the picture.
     int i_delay = static_cast<int>(delay * 1000000 + 0.5);
@@ -345,7 +344,7 @@ void FFmpegDecoderVideo::publishFrame(co
 
 void FFmpegDecoderVideo::yuva420pToRgba(AVPicture * const dst, AVPicture * const src, int width, int height)
 {
-    convert(dst, PIX_FMT_RGB24, src, m_context->pix_fmt, width, height);
+    convert(dst, AV_PIX_FMT_RGB24, src, m_context->pix_fmt, width, height);
 
     const size_t bpp = 4;
 
@@ -363,31 +362,28 @@ void FFmpegDecoderVideo::yuva420pToRgba(
     }
 }
 
-
-
-int FFmpegDecoderVideo::getBuffer(AVCodecContext * const context, AVFrame * const picture)
+int FFmpegDecoderVideo::getBuffer(AVCodecContext * const context, AVFrame * const picture, int flags)
 {
+    AVBufferRef *ref;
     const FFmpegDecoderVideo * const this_ = reinterpret_cast<const FFmpegDecoderVideo*>(context->opaque);
 
-    const int result = avcodec_default_get_buffer(context, picture);
+    const int result = avcodec_default_get_buffer2(context, picture, flags);
     int64_t * p_pts = reinterpret_cast<int64_t*>( av_malloc(sizeof(int64_t)) );
 
     *p_pts = this_->m_packet_pts;
     picture->opaque = p_pts;
 
+    ref = av_buffer_create((uint8_t *)picture->opaque, sizeof(int64_t), FFmpegDecoderVideo::freeBuffer, picture->buf[0], flags);
+    picture->buf[0] = ref;
+
     return result;
 }
 
-
-
-void FFmpegDecoderVideo::releaseBuffer(AVCodecContext * const context, AVFrame * const picture)
+void FFmpegDecoderVideo::freeBuffer(void *opaque, uint8_t *data)
 {
-    if (picture != 0)
-        av_freep(&picture->opaque);
-
-    avcodec_default_release_buffer(context, picture);
+    AVBufferRef *ref = (AVBufferRef *)opaque;
+    av_buffer_unref(&ref);
+    av_free(data);
 }
 
-
-
 } // namespace osgFFmpeg
--- openscenegraph-3.2.1.orig/OpenSceneGraph/src/osgPlugins/ffmpeg/FFmpegDecoderVideo.hpp
+++ openscenegraph-3.2.1/OpenSceneGraph/src/osgPlugins/ffmpeg/FFmpegDecoderVideo.hpp
@@ -94,8 +94,8 @@ private:
                 int src_pix_fmt, int src_width, int src_height);
 
 
-    static int getBuffer(AVCodecContext * context, AVFrame * picture);
-    static void releaseBuffer(AVCodecContext * context, AVFrame * picture);
+    static int getBuffer(AVCodecContext * context, AVFrame * picture, int flags);
+    static void freeBuffer(void * opaque, uint8_t *data);
 
     PacketQueue &           m_packets;
     FFmpegClocks &          m_clocks;
--- openscenegraph-3.2.1.orig/OpenSceneGraph/src/osgPlugins/ffmpeg/FFmpegParameters.cpp
+++ openscenegraph-3.2.1/OpenSceneGraph/src/osgPlugins/ffmpeg/FFmpegParameters.cpp
@@ -19,7 +19,7 @@ extern "C"
     #include <libavutil/pixdesc.h>
 }
 
-inline PixelFormat osg_av_get_pix_fmt(const char *name) { return av_get_pix_fmt(name); }
+inline AVPixelFormat osg_av_get_pix_fmt(const char *name) { return av_get_pix_fmt(name); }
 
 
 namespace osgFFmpeg {
--- src/OpenSceneGraph-3.4.0/src/osgPlugins/ffmpeg/FFmpegDecoderAudio.cpp.orig	2016-02-18 21:25:39.627923629 +0000
+++ src/OpenSceneGraph-3.4.0/src/osgPlugins/ffmpeg/FFmpegDecoderAudio.cpp	2016-02-18 21:26:17.071140100 +0000
@@ -227,8 +227,7 @@
         if (avcodec_open2(m_context, p_codec, NULL) < 0)
             throw std::runtime_error("avcodec_open() failed");
 
-        m_context->get_buffer = avcodec_default_get_buffer;
-        m_context->release_buffer = avcodec_default_release_buffer;
+        m_context->get_buffer2 = avcodec_default_get_buffer2;
 
     }

Bindestreck · 03-10-2016, 12:26 PM

volti 0.2.3 not working anymore due to new version of python-dbus:

Code:

Traceback (most recent call last):
  File "/usr/bin/volti", line 27, in <module>
    assert dbus.version >= (0, 80, 0)
AssertionError

I notified the maintainer, but if you cannot wait you can change line 27 on python-file /usr/bin/volti:

Code:

assert dbus.version >= (0, 80, 0)

to

Code:

assert dbus.version >= [0, 80, 0]

It is a very small change, instead of tuples it uses a list.

Cheers!

ecoslacker · 03-10-2016, 02:57 PM

I installed scipy 0.14.0 from SlackBuilds but it appears to be some bugs in this version (like this: https://bugzilla.redhat.com/show_bug.cgi?id=1286349) so I decided to use the same SlackBuild for building the latest version of scipy 1.16.1.

The build process tries to find the file "INSTALL.txt" but in the tarball the file its named "INSTALL.rst.txt", so I had to rename this file in a new tarball. Beside this the package builds fine and appears to work as expected.

Maybe this package should be patched or upgraded, in order to work in Slackware-current (14.2).

orbea · 03-10-2016, 03:22 PM

Or you could just change the slackbuild so it copies INSTALL.rst.txt instead...