Bug#656928: libraptor2-0: please add a Breaks against libraptor1 versions without symbol versioning
Package: libraptor2-0
Version: 2.0.6-1
Severity: important
Justification: http://bugs.debian.org/613411
Tags: patch
Hi,
Adrian Knoth wrote:
The situation is as follows:
- ardour depends on liblrdf (mind the l after lib)
- liblrdf depends on libraptor1
- ardour depends on libslv2
- libslv2 depends on librdf0 (no l this time)
- librdf0 (= 1.0.13-1) depends on libraptor2
And there you have it, clashing symbols.
The solution is to add symbol versioning to both raptor libraries, then
recompile librdf, liblrdf and afterwards libslv2 (the whole dependency
chain backwards). In the end, recompile ardour.
Is that last step (in the end, recompile ardour) needed?
I'm asking because if it is, this will break partial upgrades from
squeeze. If it isn't, it should be enough for the updated librdf0 to
gain a (direct or indirect) Breaks against versions of libraptor1 that
lack versioned symbols, to ensure the correct upgrade order.
That is, probably libraptor2-0 should Breaks: libraptor1 ( 1.4.21-3).
How about this patch?
---
debian/changelog |7 +++
debian/control |1 +
2 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/debian/changelog b/debian/changelog
index c48a8a87..37a940cc 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,10 @@
+raptor2 (2.0.6-1.1) local; urgency=low
+
+ * debian/control: add a breaks relation by libraptor2-0 against squeeze
+libraptor1 to force upgrades to a version with symbol versioning
+
+ -- Jonathan Nieder jrnie...@gmail.com Sun, 22 Jan 2012 16:39:45 -0600
+
raptor2 (2.0.6-1) unstable; urgency=low
* New upstream release
diff --git a/debian/control b/debian/control
index 8d758eeb..abe13400 100644
--- a/debian/control
+++ b/debian/control
@@ -21,6 +21,7 @@ Package: libraptor2-0
Section: libs
Architecture: any
Depends: ${misc:Depends}, ${shlibs:Depends}
+Breaks: libraptor1 ( 1.4.21-3)
Suggests: raptor2-utils
Description: Raptor 2 RDF syntax library
Raptor is a C library providing a set of parsers and serializers for
--
1.7.9.rc2
___
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-multimedia-maintainers
Bug#628805: jackd: free(): invalid pointer: 0xb74f2efc
Hi, Adrian Knoth wrote: This bug is caused by a change in binutils. (see the ticket) Presumably the famous Add SORT_BY_INIT_PRIORITY (2010-12-15), since it changes the order of constructor invocation. We (FFADO upstream) are working on a fix, there's already some code available, but it will take some weeks to get it right. In the meantime, using binutils-gold is said to fix it. You think we should temporarily add it to the build-deps? Yes, that sounds sensible to me. Thanks for working on this. Regards, Jonathan ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-multimedia-maintainers
Re: [PATCH/RFC 0/3] Re: libavcodec-dev: please mention the fate of CODEC_TYPE_AUDIO macros in NEWS or changelog
Reinhard Tartler wrote: Thanks, I've applied (most) of them mostly unmodified. Thanks. The changes all make sense to me fwiw. ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#623682: [PATCH/RFC 0/3] Re: libavcodec-dev: please mention the fate of CODEC_TYPE_AUDIO macros in NEWS or changelog
Hi, Reinhard Tartler wrote: * don't try to install non-existing documentation, fixes FTBFS on powerpc * add NEWS.Debian file * install doc/APIChanges and refer to them in NEWS.Debian (Closes: #623682) * readd deprecated avcodec_thread_init in libavformat Thanks for this. I tried the package and it seems that NEWS.Debian isn't actually installed. While working on that, I picked a few other nits: * the APIChanges should only be relevant to people using the development libraries, so I think it makes more sense to only install them to lib*-dev and libav-doc * using source format 3.0 (quilt), so git reset --hard and other habits don't produce confusing results * allowing debian/rules clean as a non-root user. In modern times, most people don't run debian/rules binary as true root anyway, and in the worst case debian/rules clean will error out and the user will know what to do (just a pet peave). If you have any complaints, I'd be glad to roll a new version of the patches addressing them. Thoughts? Jonathan Nieder (3): clean up NEWS.Debian and related documentation use source format 3.0 (quilt) debian/rules: allow clean as non-root Doxyfile |2 +- debian/NEWS.Debian | 18 -- debian/README.Debian |9 + debian/changelog | 10 ++ debian/control |1 - debian/libavcodec-dev.NEWS |9 + debian/rules | 16 debian/source/format |2 +- libavcodec/avcodec.h |2 +- libavcodec/utils.c |2 +- libavcodec/version.h |3 +++ 11 files changed, 43 insertions(+), 31 deletions(-) delete mode 100644 debian/NEWS.Debian create mode 100644 debian/libavcodec-dev.NEWS -- 1.7.5 ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#623682: [PATCH 1/3] clean up NEWS.Debian and related documentation
libav provides a NEWS.Debian file under debian/ but does not install it currently; rename it to libavcodec-dev.NEWS so apt-listchanges can automatically install it. Only installing in a -dev package ensures sysadmins without the development libraries installed don't have to worry about API changes that will be irrelevant to them. Just the libavcodec-dev package should be enough to make the API changes visible. While at it, move APIchanges.gz so it is just installed by the various development libraries and libav-doc. The details of the API would be irrelevant to someone just using programs that use the library. Maybe APIchanges should just be in libav-doc, but to be conservative this keeps it in the development libraries too for now. Now that NEWS is attached to a particular package and shown on upgrade, it does not seem like a good place to mention the source package renaming, which does not require any action in response to the upgrade. Move the note to README.Debian, where it is more likely to help the confused reader. Signed-off-by: Jonathan Nieder jrnie...@gmail.com --- debian/NEWS.Debian | 18 -- debian/README.Debian |9 + debian/changelog |8 debian/libavcodec-dev.NEWS |9 + debian/rules |6 +- 5 files changed, 31 insertions(+), 19 deletions(-) delete mode 100644 debian/NEWS.Debian create mode 100644 debian/libavcodec-dev.NEWS diff --git a/debian/NEWS.Debian b/debian/NEWS.Debian deleted file mode 100644 index f1f3a92..000 --- a/debian/NEWS.Debian +++ /dev/null @@ -1,18 +0,0 @@ -libav (4:0.7~b1-1) experimental; urgency=low - - * The 0.7 release series of Libav has a number of old and deprecated -APIs removed, so your packages might need some adaptations. Please see -the file /usr/share/doc/libavcodec53/APIChanges for a list of API -additions and removals. More detailed information about the new API is -provided in the shipped doxygen documentation. - - -- Reinhard Tartler siret...@tauware.de Tue, 19 Apr 2011 15:04:55 +0200 - -libav (4:0.6.2-1) unstable; urgency=low - - * Libav represents the successor of the former ffmpeg source package. It -provides all libraries and provides an upgrade path for existing -application packages. - - -- Reinhard Tartler siret...@tauware.de Sat, 19 Mar 2011 08:13:46 +0100 - diff --git a/debian/README.Debian b/debian/README.Debian index 405e591..798b209 100644 --- a/debian/README.Debian +++ b/debian/README.Debian @@ -76,6 +76,15 @@ packages are: -- Reinhard Tartler siret...@tauware.de, Sun, 26 Jul 2009 10:38:10 +0200 +Source package +== + +Libav represents the successor of the former ffmpeg source package. It +provides all libraries and provides an upgrade path for existing +application packages. + + -- Reinhard Tartler siret...@tauware.de Sat, 19 Mar 2011 08:13:46 +0100 + License of the Debian ffmpeg packages = diff --git a/debian/changelog b/debian/changelog index 9f6d545..dc47350 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,3 +1,11 @@ +libav (4:0.7~b1-2.1) local; urgency=low + + * only install doc/APIChanges in *-dev and libav-doc packages + * move note on source package lineage to README.Debian + * install NEWS.Debian in libavcodec-dev + + -- Jonathan Nieder jrnie...@gmail.com Mon, 02 May 2011 19:10:14 -0500 + libav (4:0.7~b1-2) experimental; urgency=low * don't try to install non-existing documentation, fixes FTBFS on powerpc diff --git a/debian/libavcodec-dev.NEWS b/debian/libavcodec-dev.NEWS new file mode 100644 index 000..5250b95 --- /dev/null +++ b/debian/libavcodec-dev.NEWS @@ -0,0 +1,9 @@ +libav (4:0.7~b1-1) experimental; urgency=low + + The 0.7 release series of Libav has a number of old and deprecated + APIs removed, so your packages might need some adaptations. Please see + the file APIChanges.gz for a list of API additions and removals. More + detailed information about the new API is provided in the shipped + doxygen documentation. + + -- Reinhard Tartler siret...@tauware.de Tue, 19 Apr 2011 15:04:55 +0200 diff --git a/debian/rules b/debian/rules index 1f30d79..67db8aa 100755 --- a/debian/rules +++ b/debian/rules @@ -116,10 +116,14 @@ binary-arch: build install codecs.txt formats.txt dh_testroot dh_installman -pffmpeg debian/qt-faststart.1 dh_installdocs doc/optimization.txt - dh_installdocs -A CREDITS doc/TODO doc/APIchanges + dh_installdocs -A CREDITS doc/TODO dh_installdocs -A debian/README.Debian dh_installdocs -p libavcodec52 codecs.txt dh_installdocs -p libavformat52 formats.txt + dh_installdocs -p libavcodec-dev -plibavdevice-dev \ + -plibavformat-dev -plibavfilter-dev -plibpostproc-dev \ + -plibswscale-dev -plibavutil-dev -plibav-doc \ + doc/APIchanges dh_installexamples -pffmpeg doc/ffserver.conf
Bug#623682: [PATCH 2/3] use source format 3.0 (quilt)
Allow dpkg-source to apply patches when unpacking the source. We no
longer need to keep track of whether the source is patched or not, and
quilt is no longer required at build time.
This works very nicely with dpkg-dev version 1.15.5.4 (2009-12-07) or
1.14.28 (2010-01-03) and later. In particular, since the modified
files on't need to be touched at build time, it avoids unnecessary
trouble when using make, git diff, git reset --hard, git
merge, and other tools.
Signed-off-by: Jonathan Nieder jrnie...@gmail.com
---
Doxyfile |2 +-
debian/changelog |1 +
debian/control |1 -
debian/rules |9 +++--
debian/source/format |2 +-
libavcodec/avcodec.h |2 +-
libavcodec/utils.c |2 +-
libavcodec/version.h |3 +++
8 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/Doxyfile b/Doxyfile
index 5d80a2b..d7a550e 100644
--- a/Doxyfile
+++ b/Doxyfile
@@ -563,7 +563,7 @@ RECURSIVE = YES
# excluded from the INPUT source files. This way you can easily exclude a
# subdirectory from a directory tree whose root is specified with the INPUT
tag.
-EXCLUDE=
+EXCLUDE= debian debian-shared debian-static debian-cmov .pc
.git
# The EXCLUDE_SYMLINKS tag can be used select whether or not files or
# directories that are symbolic links (a Unix filesystem feature) are excluded
diff --git a/debian/changelog b/debian/changelog
index dc47350..0a70434 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -3,6 +3,7 @@ libav (4:0.7~b1-2.1) local; urgency=low
* only install doc/APIChanges in *-dev and libav-doc packages
* move note on source package lineage to README.Debian
* install NEWS.Debian in libavcodec-dev
+ * use dpkg source format 3.0 (quilt)
-- Jonathan Nieder jrnie...@gmail.com Mon, 02 May 2011 19:10:14 -0500
diff --git a/debian/control b/debian/control
index ffe607c..59ba400 100644
--- a/debian/control
+++ b/debian/control
@@ -38,7 +38,6 @@ Build-Depends: debhelper (= 7),
libxext-dev,
libxfixes-dev,
libxvmc-dev,
- quilt,
texi2html,
yasm [amd64 i386 kfreebsd-amd64 kfreebsd-i386 mips mipsel s390 sparc powerpc],
zlib1g-dev
diff --git a/debian/rules b/debian/rules
index 67db8aa..d883601 100755
--- a/debian/rules
+++ b/debian/rules
@@ -1,7 +1,5 @@
#!/usr/bin/make -f
-include /usr/share/quilt/quilt.make
-
EPOCH=4:
DEB_SOURCE := $(shell dpkg-parsechangelog | sed -n 's/^Source: //p')
DEB_VERSION := $(shell dpkg-parsechangelog | sed -n 's/^Version: //p')
@@ -61,12 +59,11 @@ build: build-stamp
build-stamp: $(addprefix build-stamp-, $(FLAVORS))
debian-shared/tools/qt-faststart
touch $@
-clean: clean-real unpatch
-clean-real:
+clean:
dh_testdir
dh_testroot
rm -f build-stamp $(addprefix build-stamp-, $(FLAVORS)) \
- $(addprefix configure-stamp-, $(FLAVORS)) patch-stamp \
+ $(addprefix configure-stamp-, $(FLAVORS)) \
build-doxy-stamp
rm -rf $(addprefix debian-, $(FLAVORS)) doxy
dh_clean
@@ -153,7 +150,7 @@ binary-arch: build install codecs.txt formats.txt
binary: binary-indep binary-arch
.PHONY: build $(addprefix build-, $(FLAVORS)) build-doxy \
- clean clean-real \
+ clean \
configure $(addprefix configure-, $(FLAVORS)) \
binary binary-indep binary-arch \
install \
diff --git a/debian/source/format b/debian/source/format
index d3827e7..163aaf8 100644
--- a/debian/source/format
+++ b/debian/source/format
@@ -1 +1 @@
-1.0
+3.0 (quilt)
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index df4e617..9d85f28 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -3684,7 +3684,7 @@ int avcodec_check_dimensions(void *av_log_ctx, unsigned
int w, unsigned int h);
enum PixelFormat avcodec_default_get_format(struct AVCodecContext *s, const
enum PixelFormat * fmt);
-#if LIBAVCODEC_VERSION_MAJOR 53
+#if FF_API_THREAD_INIT
/**
* @deprecated Set s-thread_count before calling avcodec_open() instead of
calling this.
*/
diff --git a/libavcodec/utils.c b/libavcodec/utils.c
index 6d31254..49ffa70 100644
--- a/libavcodec/utils.c
+++ b/libavcodec/utils.c
@@ -1287,7 +1287,7 @@ void ff_thread_await_progress(AVFrame *f, int progress,
int field)
#endif
-#if LIBAVCODEC_VERSION_MAJOR 53
+#if FF_API_THREAD_INIT
int avcodec_thread_init(AVCodecContext *s, int thread_count)
{
diff --git a/libavcodec/version.h b/libavcodec/version.h
index 7e53474..1ddee2c 100644
--- a/libavcodec/version.h
+++ b/libavcodec/version.h
@@ -86,5 +86,8 @@
#ifndef FF_API_REQUEST_CHANNELS
#define FF_API_REQUEST_CHANNELS (LIBAVCODEC_VERSION_MAJOR 54)
#endif
+#ifndef FF_API_THREAD_INIT
+#define FF_API_THREAD_INIT (LIBAVCODEC_VERSION_MAJOR 54)
+#endif
#endif /* AVCODEC_VERSION_H */
--
1.7.5
___
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http
Bug#623682: [PATCH 3/3] debian/rules: allow clean as non-root
After running debian/rules build as an unprivileged user, there is no reason to require (fake) root privileges to clean up. Even when cleaning up after sudo debian/rules install, at worst the clean target would error out because unable to remove a file in a directory owned by root and the operator will know to run debian/rules clean as root to fix it. Signed-off-by: Jonathan Nieder jrnie...@gmail.com --- debian/changelog |1 + debian/rules |1 - 2 files changed, 1 insertions(+), 1 deletions(-) diff --git a/debian/changelog b/debian/changelog index 0a70434..cb6a9ae 100644 --- a/debian/changelog +++ b/debian/changelog @@ -4,6 +4,7 @@ libav (4:0.7~b1-2.1) local; urgency=low * move note on source package lineage to README.Debian * install NEWS.Debian in libavcodec-dev * use dpkg source format 3.0 (quilt) + * allow debian/rules clean as unprivileged user -- Jonathan Nieder jrnie...@gmail.com Mon, 02 May 2011 19:10:14 -0500 diff --git a/debian/rules b/debian/rules index d883601..fe4a6ea 100755 --- a/debian/rules +++ b/debian/rules @@ -61,7 +61,6 @@ build-stamp: $(addprefix build-stamp-, $(FLAVORS)) debian-shared/tools/qt-fastst clean: dh_testdir - dh_testroot rm -f build-stamp $(addprefix build-stamp-, $(FLAVORS)) \ $(addprefix configure-stamp-, $(FLAVORS)) \ build-doxy-stamp -- 1.7.5 ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#623797: idjc: ftbfs with libavcodec-dev from experimental
Source: idjc
Version: 0.8.5-1
Severity: important
Justification: future ftbfs
Tags: upstream patch
Hi Alessio et al,
Trying to build idjc against libavcodec-dev 4:0.7~b1-1 (from
experimental), I get compile-time errors due to the missing
CODEC_TYPE_AUDIO symbol. The patch below seems to fix it.
http://bugs.debian.org/623682 has some related details.
Thanks for keeping this package maintained well.
Regards,
Jonathan
c/avcodecdecode.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/c/avcodecdecode.c b/c/avcodecdecode.c
index 36090ba..51dffae 100644
--- a/c/avcodecdecode.c
+++ b/c/avcodecdecode.c
@@ -232,7 +232,7 @@ int avcodecdecode_reg(struct xlplayer *xlplayer)
for(self-stream = 0; self-stream self-ic-nb_streams; self-stream++)
{
self-c = self-ic-streams[self-stream]-codec;
- if(self-c-codec_type == CODEC_TYPE_AUDIO)
+ if(self-c-codec_type == AVMEDIA_TYPE_AUDIO)
break;
}
--
1.7.5.rc3
___
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#623799: idjc: ftbfs with CONFIG_SHELL=/bin/dash
Source: idjc Version: 0.8.5-1 Severity: important Justification: future ftbfs Tags: upstream patch Hi again, With dash in experimental as /bin/sh, this package fails to build from source. (Please forgive my laziness in omitting the corresponding error message.) The problem can also be experienced with dash from sid by overriding autoconf's choice of shell: CONFIG_SHELL=/bin/dash debian/rules build See http://bugs.debian.org/582952 for details. The cause seems to be use of the ksh-invented operator in configure.ac. Shells like dash interpret it as a request to run in the background () and redirect stdout (); to redirect both stdout and stderr one has to use the more precise /dev/null 21, as in the patch below. With the changes below applied, the package builds again. If there is additional information that would be useful, please feel free to ask. Regards, Jonathan configure.ac |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/configure.ac b/configure.ac index d681377..828c019 100644 --- a/configure.ac +++ b/configure.ac @@ -50,7 +50,7 @@ if test $makeffmpeg != no ; then AC_SUBST(HAVE_AVCODEC, 1) AC_MSG_CHECKING([for avcodec.h in legacy ffmpeg directory]) -gcc `pkg-config --cflags libavcodec` ffmpeg_avcodec.c -o /dev/null /dev/null +gcc `pkg-config --cflags libavcodec` ffmpeg_avcodec.c -o /dev/null /dev/null 21 if test $? -eq 0 ; then AC_MSG_RESULT([yes]) AC_DEFINE(FFMPEG_AVCODEC, 1, [if set use legacy include file location]) @@ -108,7 +108,7 @@ fi PKG_CHECK_MODULES([PYGTK], [pygtk-2.0 = 2.6.0]) AC_MSG_CHECKING([for mutagen]) -python -c import mutagen /dev/null +python -c import mutagen /dev/null 21 if test $? -ne 0 ; then AC_MSG_RESULT([no]) AC_MSG_ERROR([required dependency (mutagen / python-mutagen) is missing]) -- 1.7.5.rc3 ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#623682: libavcodec-dev: please mention the fate of CODEC_TYPE_AUDIO macros in NEWS or changelog
Package: libavcodec-dev Version: 4:0.7~b1-1 Hi, idjc fails to build for me: | gcc -DHAVE_CONFIG_H -I. -I..-Wall -std=gnu99 -O2-I/usr/include/FLAC -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -g -O2 -c -o idjcmixer-avcodecdecode.o `test -f 'avcodecdecode.c' || echo './'`avcodecdecode.c | avcodecdecode.c: In function ‘avcodecdecode_reg’: | avcodecdecode.c:235:33: error: ‘CODEC_TYPE_AUDIO’ undeclared (first use in this function) | avcodecdecode.c:235:33: note: each undeclared identifier is reported only once for each function it appears in | make[4]: *** [idjcmixer-avcodecdecode.o] Error 1 | make[4]: Leaving directory `/tmp/idjc/c' Presumably this is due to a small planned API change, based on libavcodec/avcodec.h: | #if LIBAVCODEC_VERSION_MAJOR 53 | #define CodecType AVMediaType | | #define CODEC_TYPE_UNKNOWNAVMEDIA_TYPE_UNKNOWN | #define CODEC_TYPE_VIDEO AVMEDIA_TYPE_VIDEO | #define CODEC_TYPE_AUDIO AVMEDIA_TYPE_AUDIO [...] Please consider documenting this in the changelog or NEWS.Debian, with a few words of advice for callers (probably that they should use s/CODEC_/AVMEDIA_/ and the latter name has worked since version x.y.z for some x, y, and z, I suppose). Thanks for keeping the ffmpeg libs working nicely. Regards, Jonathan ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#623682: libavcodec-dev: please mention the fate of CODEC_TYPE_AUDIO macros in NEWS or changelog
Reinhard Tartler wrote: Please note that upstream does not support an direct migration path from 0.5-0.6, i.e., only APIs that are missing in the last stable release are considered critical. This particular issue therefore wouldn't qualify, but there might be more. Makes sense. I was wondering about when the API changed mainly to see whether it is safe to use the new names in sid. Of course others might have other uses for the information. Would installing the file doc/APIChanges [1] and refering to that in a NEWS.Debian help you? [1] http://git.debian.org/?p=pkg-multimedia/libav.git;a=blob;f=doc/APIchanges Yes, that looks excellent. In the motivating case, grepping for CODEC_TYPE_ yields no hits, but scanning through, it is not too hard to find the discussion of enum CodecType. Thanks, Jonathan ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#582356: csound: please rebuild against tcl 8.5 in experimental
Hi Felipe, Felipe Sateler wrote: Tcl is already at 8.5 by default, so this bug is not necesary anymore. Ah, good to hear. The important work (making csound build against arbitrary tcl versions) had already been done before this all started. Thanks and sorry for the noise. ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Re: unblock: tap-plugins/0.7.1-1
Hi, Alessio Treglia wrote: Please unblock tap-plugins 0.7.1-1 Here are the obvious reactions. :) - new upstream point release. previous version is from 5 years ago, before the PTS learned about testing migration. - the proposed version has been in sid for about seven months - no open bugs - popcon vote: 473. seems like a sizable, growing user base. - the new upstream version: + addresses some bugs: . 64-bit cleanliness . floating-point underflow handling . uninitialized values . dangerous typo in Makefile + adds a new and very simple stereophonic-to-binaural plugin + I haven't read the diff in detail, but it seems small and reviewable - the packaging changes are bigger than ideal for this moment in the freeze --- in particular, a debian/rules rewrite. Do they fix a bug? Maybe a more minimal update is possible? Of course I am not a member of the release team, so if that seems like a lot of trouble, it is probably best to ping them first. Regards, Jonathan ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#582356: csound: please rebuild against tcl 8.5 in experimental
Felipe Sateler wrote: Hmm, I'm confused. Why would the experimental buildd use the experimental version if there is a suitable version in unstable? Agh, let me start over. The default Tcl/Tk version in unstable (as set by tcltk-defaults) is 8.4, whereas the version in experimental is 8.5. Before updating the version in unstable, we would like to see what fallout that would have, by rebuilding all relevant packages against the version in experimental. Now as you noted, experimental buildds use packages from sid by default, so this requires sourceful uploads to change build-deps to explicitly request the version from experimental: Build-Depends: tcl-dev (= 8.5), tk-dev (= 8.5) Since tk-dev 8.5 depends on tk (= 8.5.0-1), this should pull in the tcl and tk packages from experimental, too. Previous versions of sbuild did not cope with that, so in my original message I suggested Build-Depends: tcl-dev (= 8.5), tk-dev (= 8.5), tcl (= 8.5), tk (= 8.5) even though that has been unnecessary for the last couple of months or so. So a simple binNMU would insufficient and the patch I sent would be more than sufficient. Sorry for the ramble. Hopefully this is clearer. Jonathan ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
Bug#564828: subgraph starting at ... timed out after about five seconds of audio playback in mplayer
found 564828 jackd/0.118+svn3796-1 reassign 564828 libasound2-plugins 1.0.21-3 thanks Hi ALSA maintainers, I reported this bug earlier today. Basically, mplayer locks up after some seconds of playback using the ALSA jack plugin, and even ^C will not kill it. | $ mplayer -ao alsa 69550__dataka__mu_eca_piano.wav | MPlayer SVN-r29142-4.3.3 (C) 2000-2009 MPlayer Team | mplayer: could not connect to socket | mplayer: No such file or directory | Failed to open LIRC support. You will not be able to use your remote control. | | Playing /home/jrn/web/69550__dataka__mu_eca_piano.wav. | Audio only file format detected. | == | Opening audio decoder: [pcm] Uncompressed PCM audio decoder | AUDIO: 44100 Hz, 2 ch, s16le, 1411.2 kbit/100.00% (ratio: 176400-176400) | Selected audio codec: [pcm] afm: pcm (Uncompressed PCM) | == | AO: [alsa] 44100Hz 2ch s16le (2 bytes per sample) | Video: no video | Starting playback... | A: 25.8 (25.8) of 42.0 (42.0) 0.4% | | MPlayer interrupted by signal 2 in module: play_audio | | MPlayer interrupted by signal 2 in module: play_audio From the original report (though the above transcript is from a sid/experimental hybrid, not lenny): | Stock lenny system. jackd started with jackd -d alsa -d hw:0 -r | 44100 ALSA set up to use jack output as in | http://alsa.opensrc.org/index.php/Jack_(plugin). mplayer defaults | to using ALSA to output. After a few seconds of playback (of a WAV | file, though it does the same with FLAC and MP3 files), it hangs (does | not respond to 'q' to quit), and jackd produces the following message | on its terminal: | | subgraph starting at alsa-jack.jackP.8292.0 timed out (subgraph_wait_fd=7, status = 0, state = Running) | | With 'mplayer -ao jack', on the other hand, everything works fine. After a little investigation, it looks like this is a known problem in ALSA's jack backend. Any pointers for fixing this? Perhaps it shouldn't be fixed, since the pulseaudio plugin takes care of most use cases already (with pulseaudio's jack backend as an added step of indirection). In that case, we should probably add notices to the package description and NEWS.Debian to discourage use of the jack plugin. Tested with recent packages this time. | $ COLUMNS=120 dpkg -l jackd libasound2-plugins mplayer | ||/ Name Version Description | +++---== | ii jackd0.118+svn3796-1 JACK Audio Connect | ii libasound2-plugins 1.0.21-3 ALSA library addit | ii mplayer 1.0~rc3+svn20090405-1movie player for U A little context on the use case is included below. Adrian Knoth wrote: With 'mplayer -ao jack', on the other hand, everything works fine. Ideas? Yeah, don't use mplayer with ao alsa when you really want ao jack. jackd is for pro-audio, think of recording studios, and the alsa-jackd-bridge is just a little hack to enable non-jackified apps to output in these environments. In general, update jackd, run it with realtime prios, update the alsa-jackd bridge or replace it by pulseaudio. (I see exactly the same problems with the alsa-jackd bridge, that's why I switched to pulseaudio) My friend has been working on getting some real-time audio processing done, investigating various APIs for doing that. JACK looks like a nice one. But unfortunately, if jackd is running, he could not play movies through Flash. Adobe’s Flash plugin uses ALSA, so the alsa-jackd bridge seemed like a natural fix. mplayer was just a demo program to make sure the bridge was working. Pulseaudio has plenty of advantages anyway, so I do think I’ll switch to that. In current Debian unstable, we have all these things, jackd operates the pro-audio gear and pulseaudio-module-jack bridges between consumer apps and the pro-audio jackd. Maybe jackd could Suggests: pulseaudio-module-jack for this. (But if you'd like to discuss this more, we should probably take it to pkg-multimedia-maintainers or a new bug report.) To quote jackd author Paul Davis: jackd is not intended to be a generic desktop sound server. This was clear from the home page already, but maybe the package description should mention it too. Thanks, Jonathan ___ pkg-multimedia-maintainers mailing list pkg-multimedia-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/pkg-multimedia-maintainers
