NetBSD Problem Report #23727

Received: (qmail 28760 invoked by uid 605); 12 Dec 2003 20:06:00 -0000
Message-Id: <20031212200559.CBAFC1115F@narn.netbsd.org>
Date: Fri, 12 Dec 2003 20:05:59 +0000 (UTC)
From: rkr@olib.org
Sender: gnats-bugs-owner@NetBSD.org
Reply-To: rkr@olib.org
To: gnats-bugs@gnats.NetBSD.org
Subject: mplayer does not configure on amd64
X-Send-Pr-Version: www-1.0

>Number:         23727
>Category:       pkg
>Synopsis:       mplayer does not configure on amd64
>Confidential:   no
>Severity:       critical
>Priority:       low
>Responsible:    pkg-manager
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Dec 12 20:07:00 +0000 2003
>Closed-Date:    Sat Dec 13 00:49:04 +0000 2003
>Last-Modified:  Sat Dec 13 00:52:00 +0000 2003
>Originator:     Richard Rauch
>Release:        NetBSD/amd64-current
>Organization:
n/a
>Environment:
NetBSD socrates 1.6ZE NetBSD 1.6ZE (GENERIC) #0: Mon Nov 10 11:15:44 CET 2003  fvdl@k2:/usr/obj/current/sys/arch/amd64/compile.amd64/GENERIC amd64

>Description:
Attempting to build fails during configure because the amd64 port's CPU is
not correctly identified.  ("uname -p" gives x86_64, which isn't
recognized at all.  "uname -m" is then tried, and "x86_64" would be
accepted, but we return "amd64", which also fails.)

>How-To-Repeat:
Attempt to build mplayer on NetBSD/amd64-current.
>Fix:
Post-"make patch", I made the following change to add "amd64" to the
valid "uname -m" output types recognized by configure.  Then the whole
thing compiles and installs.  (Not yet tested, sorry.)

--- configure.old       2003-12-12 12:54:06.000000000 -0600
+++ configure   2003-12-12 12:54:48.000000000 -0600
@@ -406,7 +406,7 @@
       case "`( uname -m ) 2>&1`" in
       i[3-9]86*|x86|x86pc|k5|k6|k6_2|k6_3|k6-2|k6-3|pentium*|athlon*|i586_i686|i586-i686) host_arch=i386 ;;
       ia64) host_arch=ia64 ;;
-      x86_64) host_arch=x86_64 ;;
+      x86_64|amd64) host_arch=x86_64 ;;
       ppc) host_arch=ppc ;;
       alpha) host_arch=alpha ;;
       sparc*) host_arch=sparc ;;

>Release-Note:
>Audit-Trail:

From: Thomas Klausner <wiz@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:  
Subject: pr/23727 CVS commit: pkgsrc/graphics
Date: Fri, 12 Dec 2003 22:09:02 +0000 (UTC)

 Module Name:	pkgsrc
 Committed By:	wiz
 Date:		Fri Dec 12 22:09:02 UTC 2003

 Modified Files:
 	pkgsrc/graphics/gmplayer: distinfo
 	pkgsrc/graphics/mencoder: Makefile distinfo
 	pkgsrc/graphics/mplayer: distinfo
 	pkgsrc/graphics/mplayer-share: Makefile Makefile.common
 	    Makefile.depends PLIST distinfo
 	pkgsrc/graphics/mplayer-share/patches: patch-aa patch-ad

 Log Message:
 Update to 1.0rc3.
 Package fixes:
 mencoder now installs a correct man page link on MANZ=1 systems.
 Added a patch so this might compile on amd64 systems, addresses
 PR 23727.

 Changes:
 MPlayer 1.0pre3: "The Real Counter"

 DOCS:

     * all MPlayer and MEncoder options documented
     * all FFmpeg/libavcodec options documented
     * Support for building incomplete XML documentation
     * Support for building all-in-one HTML docs
     * HTML documentation obsoleted
     * English, Polish and French HTML documentation removed
     * Spanish documentation translation finished
     * Macedonian runtime output translation added
     * small improvements and bug fixes all over the place

 Ports:

     * fixed compilation of Matroska on MinGW32/Cygwin
     * support for detecting non-Intel CPUs under Cygwin
     * sub autoloading under Windows
     * removed pthread dependency (compiles out of the box on a default NetBSD install)
     * more 64 bit fixes
     * a lot of Mac OS X fixes
     * prefer TOOLS/cpuinfo over linuxemu's /proc/cpuinfo under FreeBSD/x86
     * Darwin XMMS libs supported
     * support newer Darwin versions in Altivec detection code
     * support for newer Apple GCCs

 Codecs and demuxers:

     * better support for (buggy) MEncoder created streams in libmpeg2
     * MPEG-TS demuxer updates
     * fixed the MPEG muxer
     * fixed tons of bugs in MMS streaming
     * user settable http user-agent field
     * improved (now working) seeking with support for growing files in Realmedia format
     * saveable and loadable index files (workaround for formats not supporting files over 2GB)
     * Ultimotion VfW decoder
     * support for MMS streams with UTF urls
     * big DVB demuxer update
     * fixed MOV demuxer to always read ImageDesc (fixes ffsvq3 bugs)
     * fixed an old bug in AVI/Waveformatex size calculation
     * XviD API-4 (1.0 beta2) support
     * improved FLAC-in-Ogg support
     * Matroska: better AC3 detection
     * Matroska: VobSubs, MP2, FLAC, AAC and HE-AAC support
     * improved detection of Real RTSP through URL analysis
     * fixed stereo 16 bit TWOS
     * fixed MPNG when using BGR24 images
     * control functions in OGG demuxer
     * http cookies support
     * LML-M4 MPEG4 capture card raw stream format support

 Filters:

     * some fixes in the delogo filter
     * optimized eq2 filter
     * weighted gamma support in the eq2 filter
     * new filters: hue, spp, fil, yuvcsp
     * alternative postprocessing filter (spp)
     * big scale filter updates
     * zrmjpeg filter, a kick-ass mjpeg encoder using FFmpeg/libavcodec
     * filmdint: new inverse telecine filter, heavily mmx and 3dnow optimized

 FFmpeg/libavcodec:

     * Lagrange multipliers instead of qscale, encoding quality much improved
     * fixed an old bug in MPlayer's support code
     * support wider range of VOLs in H263
     * direct rendering support in MJPEG
     * massive MLib optimizations
     * more VQA files supported
     * Theora support
     * Sunplus JPEG (SP5X) support
     * H.263 GOB fixes
     * massive cleanups
     * reduced memory footprint (!)
     * fixed decoding if aspect ratio changes
     * MMX2 optimizations in HuffYUV
     * lots of bugs squashed in HuffYUV
     * long standing decoding errors after 'first frame is not keyframe' fixed
     * CRI ADX support
     * XA ADPCM support
     * fixed files with odd dimensions on PPC
     * optimized MPEG2 bitstream parsing
     * better interlacing framework
     * support for IBM's XLC compiler
     * support for SAR (sample aspect ratio)
     * floating point AAN DCT
     * dynamically alloc big data fields (lower memory usage in most cases!)
     * fixed stereo IMA-ADPCM encoding
     * inline vs always_inline - some speedups
     * MMX2 optimized FDCT
     * accurate 2-4-8 DCT
     * much faster DV encoding (it beats libdv)
     * support Cinepak files with funky (not divisible by 4) resolutions
     * Altivec optimized FDCT
     * XvMC speedups
     * initial ZyGoVideo decoding
     * RGB support in FFV1
     * new Palette API, all codecs updated
     * Noise reduction of DCT coefficients
     * "av_log" logging API
     * support for buffer hints from codecs
     * BGR24, RGB555 and PAL8 image formats
     * 8BPS, MS RLE, MS Video1, QT RPZA, QT SMC, FLIC and TrueMotion1 decoders
     * 16x8 MV visualisation support
     * H.263 overlapped block motion compensation (OBMC), 4MV support
     * H.263 alternative inter vlc support
     * H.263 deblocking filter (MMX optimized)
     * H.263 modified quantization support
     * H.263 slice structured mode support
     * Real RV20 decoder (with B-frame support)

 Drivers:

     * VIDIX equalizer support in fbdev, svga and vesa
     * VIDIX colorkeying support in fbdev, svga and vesa
     * obsoleted option: -fb
     * new SiS 650/651/740/etc VIDIX driver
     * safer vo_directx direct rendering
     * massive nvidia_vid updates
     * gl2 updates
     * ggi now supports non-directbuffer rendering and slices
     * tdfxfb fixed on gcc 3.x
     * dfbmga updated, G200 support
     * colorkeying can be disabled
     * Radeon 9800 (R350) support in VIDIX
     * proper fix for buffer size workarounds in vo_zr
     * new vo_zr2 driver, which uses the advantages of the filter layer
     * fixed vesa when used together with pthread
     * runtime stay-on-top functionality (in almost all output drivers)
     * ALSA 1.x support

 SWScaler:

     * updated MLib (Sun VIS) support
     * more accurate filters (rounding fixes)
     * MMX optimized UYVY output
     * eliminated chroma scaling bugs

 Others:

     * fixed aspect and geometry
     * fixed compilation when libavcodec is missing (although it's a bad idea)
     * MTRR detection added to TOOLS/cpuinfo
     * FAAD detection in configure fixed/improved
     * improved FLAC configure detection
     * GUI file selector updated for newer formats
     * Fontconfig support
     * proper VobSub seeking support
     * removed oldskool libcss support (nobody uses it and has drawbacks)

 MEncoder:

     * FFmpeg/libavcodec audio encoding support
     * exit if there's no video stream (avoids unexpected segfaults)
     * huge bug fixed (missing InitTimer call), screwed up timers in some cases
     * rawyuv (i420) 'encoder'


 To generate a diff of this commit:
 cvs rdiff -r1.44 -r1.45 pkgsrc/graphics/gmplayer/distinfo
 cvs rdiff -r1.20 -r1.21 pkgsrc/graphics/mencoder/Makefile
 cvs rdiff -r1.27 -r1.28 pkgsrc/graphics/mencoder/distinfo
 cvs rdiff -r1.41 -r1.42 pkgsrc/graphics/mplayer/distinfo
 cvs rdiff -r1.12 -r1.13 pkgsrc/graphics/mplayer-share/Makefile
 cvs rdiff -r1.47 -r1.48 pkgsrc/graphics/mplayer-share/Makefile.common
 cvs rdiff -r1.23 -r1.24 pkgsrc/graphics/mplayer-share/Makefile.depends
 cvs rdiff -r1.14 -r1.15 pkgsrc/graphics/mplayer-share/PLIST
 cvs rdiff -r1.31 -r1.32 pkgsrc/graphics/mplayer-share/distinfo
 cvs rdiff -r1.7 -r1.8 pkgsrc/graphics/mplayer-share/patches/patch-aa \
     pkgsrc/graphics/mplayer-share/patches/patch-ad

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

State-Changed-From-To: open->feedback 
State-Changed-By: wiz 
State-Changed-When: Fri Dec 12 22:10:36 UTC 2003 
State-Changed-Why:  
During the update to 1.0rc3 I added a patch to the configure 
script that should make it configure correctly on amd64 systems 
(based on your patch, but an additional chunk) -- please verify 
that this works for you. 

From: Richard Rauch <rkr@olib.org>
To: gnats-bugs@gnats.netbsd.org
Cc: wiz@netbsd.org, pkg-manager@netbsd.org
Subject: Re: pkg/23727
Date: Fri, 12 Dec 2003 16:38:51 -0600

 On Fri, Dec 12, 2003 at 10:11:32PM -0000, wiz@netbsd.org wrote:
 > Synopsis: mplayer does not configure on amd64
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: wiz
 > State-Changed-When: Fri Dec 12 22:10:36 UTC 2003
 > State-Changed-Why: 
 > During the update to 1.0rc3 I added a patch to the configure
 > script that should make it configure correctly on amd64 systems
 > (based on your patch, but an additional chunk) -- please verify
 > that this works for you.

 I just did a cvs update -PAd a minute ago, and now I get this when I
 try to "make":

 ===> Checking for vulnerabilities in mplayer-1.0rc1
 *** WARNING - remote-code-execution vulnerability in mplayer-1.0rc1 - see http://www.mplayerhq.hu/homepage/news.html#vuln01 for more information ***
 or define ALLOW_VULNERABLE_PACKAGES if this package is absolutely essential
 *** Error code 1

 Stop.
 make: stopped in /usr/pkgsrc/graphics/mplayer
 *** Error code 1

 Stop.
 make: stopped in /usr/pkgsrc/graphics/mplayer


 ...the prior version simply built, after I patched the configure script.


 -- 
   "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/

From: Thomas Klausner <wiz@netbsd.org>
To: Richard Rauch <rkr@olib.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: pkg/23727
Date: Fri, 12 Dec 2003 23:42:47 +0100

 On Fri, Dec 12, 2003 at 04:38:51PM -0600, Richard Rauch wrote:
 > I just did a cvs update -PAd a minute ago, and now I get this when I
 > try to "make":
 > 
 > ===> Checking for vulnerabilities in mplayer-1.0rc1

 I don't know what server you used, but 1.0rc1 is _old_.
 I just updated the package to 1.0rc3, please try again with
 that version.

  Thomas

From: Richard Rauch <rkr@olib.org>
To: Thomas Klausner <wiz@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: pkg/23727
Date: Fri, 12 Dec 2003 17:33:01 -0600

 On Fri, Dec 12, 2003 at 11:42:47PM +0100, Thomas Klausner wrote:
 > On Fri, Dec 12, 2003 at 04:38:51PM -0600, Richard Rauch wrote:
 > > I just did a cvs update -PAd a minute ago, and now I get this when I
 > > try to "make":
 > > 
 > > ===> Checking for vulnerabilities in mplayer-1.0rc1
 > 
 > I don't know what server you used, but 1.0rc1 is _old_.

 I didn't have mplayer-share updated, sorry.  I have several packages
 that are skewed relative to pkgsrc CVS, and last did a "cvs update"
 a couple of days ago.  Rather than create potential problems with
 those other packages, I just updated the "mplayer" dir.

 Updating mplayer-share, too, let it build for me.


 > I just updated the package to 1.0rc3, please try again with
 > that version.

 Better and worse:

  * Better: It compiles without local tweaks.

  * Worse: Audio is completely gone.

    The patch that I sent earlier today let it build, run, and play
    audio.

    I tested xmms after trying the new mplayer to be sure that it
    wasn't just a flakey sound system cutting out on me.  xmms could
    play music fine, so the new mplayer seems to be the definite culprit.

 It may be worth noting that I turned *off* "arts" support for mplayer
 when I first tried to build mplayer due to problems with that package.
 (I don't remember exactly what, now, but enough things were not working
 that I think that I lost track of it and never filed a PR.  However, arts
 should not be required, as far as I know, so that shouldn't be the problem.)


 -- 
   "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/

From: Thomas Klausner <wiz@netbsd.org>
To: Richard Rauch <rkr@olib.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: pkg/23727
Date: Sat, 13 Dec 2003 00:44:21 +0100

 On Fri, Dec 12, 2003 at 05:33:01PM -0600, Richard Rauch wrote:
 >  * Worse: Audio is completely gone.
 > 
 >    The patch that I sent earlier today let it build, run, and play
 >    audio.

 The only additional change is the following part of patch-aa:
 @@ -819,7 +819,7 @@ EOF
      _optimizing=''
      ;;

 -  x86_64)
 +  x86_64|amd64)
      _def_arch='#define ARCH_X86_64 1'
      _target_arch='TARGET_ARCH_X86_64 = yes'
      iproc='x86_64'

 Perhaps you want to try removing that part, then it's just 1.0pre3
 with your patch; if audio then still doesn't work, it's a separate
 problem new in 1.0pre3; if it does work, I'll just remove the chunk.

 Cheers,
  Thomas

From: Richard Rauch <rkr@olib.org>
To: Thomas Klausner <wiz@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: pkg/23727
Date: Fri, 12 Dec 2003 18:41:24 -0600

 On Sat, Dec 13, 2003 at 12:44:21AM +0100, Thomas Klausner wrote:
 > On Fri, Dec 12, 2003 at 05:33:01PM -0600, Richard Rauch wrote:
 > >  * Worse: Audio is completely gone.
 > > 
 > >    The patch that I sent earlier today let it build, run, and play
 > >    audio.
 > 
 > The only additional change is the following part of patch-aa:
 > @@ -819,7 +819,7 @@ EOF

 Hm.  Chalk this up to operator error.

 I have several scripts that I use for caching options to mplayer,
 since I can never remember which ones it wants (and it has close to
 zero intelligence for figuring things out---unlike ogle).

 The latest version of the script (#5) is using sdl/esd for audio,
 and that apparently failed (maybe there's a silent dependancy, there?).
 #4 uses "-ao oss" and that works exactly as you have the patch in
 the repository.

 I probably typoed a "4" for a "5" when I got it to play on this
 machine earlier today.  That worked.  I no longer remember why I
 switched from oss to esd/sdl sound.  It's almost enough to make
 me switch to ogle, despite the better quality of playback that I
 get with mplayer.  (^&


 Anyway.  Sorry; mea culpa.  I think that you can close this PR.


 -- 
   "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/
State-Changed-From-To: feedback->closed 
State-Changed-By: wiz 
State-Changed-When: Sat Dec 13 00:48:07 UTC 2003 
State-Changed-Why:  
Submitter reports problem is fixed. Thanks! 

From: Thomas Klausner <wiz@netbsd.org>
To: Richard Rauch <rkr@olib.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: pkg/23727
Date: Sat, 13 Dec 2003 01:50:56 +0100

 On Fri, Dec 12, 2003 at 06:41:24PM -0600, Richard Rauch wrote:
 > Hm.  Chalk this up to operator error.
 > 
 > I have several scripts that I use for caching options to mplayer,
 > since I can never remember which ones it wants (and it has close to
 > zero intelligence for figuring things out---unlike ogle).

 You can write lots of stuff into your .mplayer/config file.
 Only one pattern of options, but that's enough for me, really.

 I really prefer mplayer's ability to fast forward 10s/1m/10m at
 at a time.

 > The latest version of the script (#5) is using sdl/esd for audio,
 > and that apparently failed (maybe there's a silent dependancy, there?).
 > #4 uses "-ao oss" and that works exactly as you have the patch in
 > the repository.

 You might want to check your MPLAYER_DISABLE_DRIVERS setting in
 /etc/mk.conf, perhaps you disabled esd and sdl yourself? :)

 Cheers,
  Thomas
>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.