This is a major update (9.8 was released 2001), most significant changes
are support for new (Linux) kernel interfaces and improved drive cache
model.
pkgsrc note: I've tried to retain Darwin/FreeBSD/DragonFly support,
except in scsi_interface.c where the structure changed significantly.
Someone who can test please add it back.
changes: many bugfixes
pkgsrc changes:
-use brasero rather than nautilus-cd-burner for CD writing
(this works well for me now)
-use musicbrainz rather than -2
(audio CD reading doesn't work on NetBSD due to HAL problems,
so I couldn't test)
changes: bugfixes
pkgsrc change: enable pkg-config again, so that the .pc file gets
installed (didn't find a pkg which depends on the old shout-config
script)
from Guillaume Lasmayous per PR pkg/41468
This is a major update - from the announcement:
24/32 bit support, audio archive support (bz2/iso9660/zip), all config
entries are completely optional now, multiple and pulseaudio mixer
support, new audio playing modes (consume, etc..), sticker database,
improved tag parsing of all types and too many interesting highlights left
out of this list
pkgsrc notes:
Archive support isn't enabled yet (disabled per default upstream),
neither is pulseaudio support,
just kept the functionality of the previous version.
I didn't find out what libcue does for flac support -- album FLACs
work well for me without that, so I've left it out for now.
#mp3splt v. 2.2.5
- added -x option to not write Xing header
mostly libmp3splt (but important for mp3splt users) :
- fixed bug with -g and multiple usages of @o (libmp3splt)
- fixed a small bug with genre in -g
- Xing header is now written even if -n option is used (libmp3splt)
- fixed bug with stdin & id3v2 & xing
- fixed critical bug with overlap splitpoints not working (feature #1530782 => bug)
- fixed critical bug with mp3 stdin (& framemode)
- fixed some warning messages & added some more
- fixed dewrap error message
- fixed dewrap bug to not attempt to create output directories when '-d' option is used
- fixed minor bug #2790952 : time mode creating an empty file (at the end)
#mp3splt v. 2.2.4
- fixed issue with CUE file not being imported correctly : \r and \n
problem (bug #2778717)
- fixed bug #2691576 - CUE file seg fault if no artist
- fixed segfault when output format was empty
- fixed bug with ID3v1 and '@N' (was 1,3,5,... instead of 1,2,3,...)
- fixed time split with '0.0' was infinite looping
- fixed empty output format '' seg fault
- small fixes in the man page
- re-enabled mp3splt ogg symlink configure option
- added options @m,@M,@s,@S,@H,@h + no @h or @H in output filenames if hundreths are 0
- added default custom tags as %[@o,@N=1]
- print to stdout is now on stderr for the '-l' option, except for the filenames
- added --with-mp3splt(-libraries|-includes)?, --disable-mp3splttest configure options
- added overlapping split files with the -O option (feature request #1530782)
- fixed compilation problem, bug #2785683
- fixed critical bug with overlap splitpoints not working (feature
#1530782 => bug)
- fixed critical bug with mp3 stdin (& framemode)
- fixed bug #2786300 in plugins compilation (added -lm)
- Xing header is now written even if SPLT_NO_TAGS option is used
- added SPLT_OPT_XING option to not write Xing header
- fixed bug with custom tags and multiple usages of @o
- fixed bug with stdin & id3v2 & xing
- fixed a small bug with genre in custom tags
- fixed some warning messages & added some more
- fixed dewrap error message
- fixed dewrap bug to not attempt to create output directories when
'-d' option is used
- fixed minor bug #2790952 : time mode creating an empty file (at the end)
- fixed issue with CUE file not being imported correctly : \r and
\n problem (bug #2778717)
- fixed bug #2691576 - CUE file seg fault if no artist
- fixed bug #2520183 : wrong size in snprintf call
- fixed bug with ID3v1 and '@N' (was 1,3,5,... instead of 1,2,3,...)
- fixed time split with '0.0' was infinite looping
- fixed empty output format '' seg fault
- added options @m,@M,@s,@S,@H,@h + no @h or @H in output filenames
if hundreths are 0
- added --with-(mad|id3)(-libraries|-includes)?, --disable-[mad|id3]test
configure options
- added overlapping option of split files (feature request #1530782)
Sound Juicer "Bonnie and Clyde" 2.26.1
======================================
* Read the track artist instead of album artist in Musicbrain3
* Don't crash if the release date is unknown
* Read tracks when falling back to gvfs
Shared directories can now be created independently by the pacakges
needing them and will be removed automatically by pkg_delete when empty.
Packages needing empty directories can use the @pkgdir command in PLIST.
Discussed and ok'd in thread starting at
http://mail-index.netbsd.org/tech-pkg/2009/06/30/msg003546.html
Two of the patches were from upstream CVS, the other two are not needed
any longer because the configure script was improved.
libvorbis 1.2.3 (2009-07-09) -- "Xiph.Org libVorbis I 20090709"
* correct a vorbisfile bug that prevented proper playback of
Vorbis files where all audio in a logical stream is in a
single page
* Additional decode setup hardening against malicious streams
* Add 'OV_EXCLUDE_STATIC_CALLBACKS' define for developers who
wish to avoid avoid unused symbol warnings from the static
callbacks defined in vorbisfile.h
libvorbis 1.2.2 (2009-06-24) -- "Xiph.Org libVorbis I 20090624"
* define VENDOR and ENCODER strings
* seek correctly in files bigger than 2 GB (Windows)
* fix regression from CVE-2008-1420; 1.0b1 files work again
* mark all tables as constant to reduce memory occupation
* additional decoder hardening against malicious streams
* substantially reduce amount of seeking performed by Vorbisfile
* Multichannel decode bugfix
* build system updates
* minor specification clarifications/fixes
libvorbis 1.2.1 (unreleased) -- "Xiph.Org libVorbis I 20080501"
* Improved robustness with corrupt streams.
* New ov_read_filter() vorbisfile call allows filtering decoded
audio as floats before converting to integer samples.
* Fix an encoder bug with multichannel streams.
* Replaced RTP payload format draft with RFC 5215.
* Bare bones self test under 'make check'.
* Fix a problem encoding some streams between 14 and 28 kHz.
* Fix a numerical instability in the edge extrapolation filter.
* Build system improvements.
* Specification correction.
- Set LICENSE=modified-bsd
- Give the package to pkgsrc-users
WavPack Library Source Code - 4.50.1
------------------------------------
fixed: alignment fault when manipulating APEv2 tags (non-x86 only)
fixed: build on UNIX via elimination of non-standard strnlen()
wavpack.exe (command-line encoder) - 4.50.1
wvunpack.exe (command-line decoder) - 4.50.1
--------------------------------------------
fixed: checking return value of iconv_open() prevents core dump on Solaris
WavPack Library Source Code - 4.50
----------------------------------
added: dynamic noise shaping for improved hybrid quality
added: option to merge blocks of similar redundancy
added: ability to store and retrieve extra mode level
fixed: alignment fault on some big-endian machines
fixed: compiling with enable-mmx on gcc 4.3.x (thanks Joachim)
improved: allow bitrate to be calculated for files down to 1/10 second
improved: decoding of corrupt files (prevents heap overrun crashes)
wavpack.exe (command-line encoder) - 4.50
-----------------------------------------
added: dynamic noise shaping for improved hybrid quality
added: --channel-order option to reorder nonconforming multichannel files
added: --merge-blocks option to optimize storage of LossyWAV output files
added: ignore -o on Windows for compatibility with Linux version
fixed: alignment fault on some big-endian machines
improved: reformatted and expanded --help display
wvunpack.exe (command-line decoder) - 4.50
------------------------------------------
fixed: don't ignore fractions of seconds in --skip option
added: show extra level and dns status for newer files (-s command)
added: ignore -o on Windows for compatibility with Linux version
improved: decoding of corrupt files (prevents heap overrun crashes)
improved: display bitrate for files down to 1/10 second
version 3.0.10 (2009-03-30)
+ cat mode: fixed bug that prevented use of the -d option
+ cat mode [win32]: properly write WAVE data to the terminal
+ conv mode [win32]: properly read WAVE data from the terminal
version 3.0.9 (2009-03-18)
+ allow fix mode to fix (essentially postpad) just one file, which is useful
when using the -c option to check for boundary issues
+ added new global -F option to specify a file containing input filenames
(most modes will now accept input filenames from a single source, using the
following order of precedence: file specified by -F option, otherwise
filenames on the command line, otherwise filenames read from the terminal)
version 3.0.8 (2008-08-27)
+ added new -a option to show default encoder/decoder arguments
+ fixed bug when prompting for overwrite action when stdin was
already at EOF on certain platforms (e.g. OS X)
+ split mode: handle CUE sheets that contain Unicode BOMs
+ wv format: support self-extracting WavPack files
version 3.0.7 (2008-03-02)
+ support for TAK decoding (requires TAK 1.0.4+)
+ updated MPEG-4 ALS to version RM20, and added output support
+ removed detection of unsupported formats RKAU and WMA
version 3.0.6 (2007-12-16)
+ support for TAK encoding (requires TAK 1.0.3+)
version 3.0.5 (2007-10-22)
+ hash mode: show progress indicators
+ cmp mode: new -c option to specify size of byte-shift comparison buffer
version 3.0.4 (2007-09-07)
+ no longer ignore raw WavPack files (wvunpack creates WAVE header for us)
+ support for TTA 3.4.x as well as TTA encoding
+ fixed display bug in joined mode when input files were not CD-quality
+ split mode: new -m option to allow character manipulation of CUE filenames
version 3.0.3 (2007-06-01)
+ improved use of exit codes
+ mkw format: removed deprecated input kluge
+ split mode: override path separation characters in CUE-generated files
version 3.0.2 (2007-02-18)
+ split mode: -c option works again
+ split mode: support wider range of CUE sheets
+ aiff format: support for sox 13.x
version 3.0.1 (2007-01-24)
+ fixed bug with WavPack format detection on 64-bit systems
+ split mode: fixed bug parsing CUE sheets whose last line lacks a newline
+ added support for MKW files
+ added support for MPEG-4 ALS files (input only)
+ added support for detection of TAK files (no input or output yet)
+ added support for detection of WMA files (no input or output yet)
+ added support for detection of RKAU files (no input or output yet)
version 3.0.0 (2007-01-01)
[platform support]
+ shntool is now a native windows executable (Cygwin is no longer needed)
[file formats]
+ added support for WavPack 4.x files
+ added support for m4a (Apple Lossless Audio Codec) files (input only)
+ added support for la (Lossless Audio) files (input only)
+ added support for tta files (input only)
+ added support for bonk files
+ added support for kxs (Kexis) files (input only)
+ added new format 'term' to send output to a terminal (standard output)
+ cust format: correct output filenames now shown (instead of "file.custom")
+ cust format: partially-written output files are now removed on error
[modes]
+ added new mode 'gen' to generate arbitrary-sized files containing silence
+ added new mode 'trim' to trim silence from front and/or back of files
+ renamed mode 'md5' to 'hash', since it now calculates MD5 and SHA1 hashes
+ hash mode: added option to reorder filenames - may be useful with -c
+ split mode: new -n option to control output count format (01, 001, ...)
+ split mode: support for wider range of CUE sheets
+ split mode: new -t option to name files based on keywords in CUE sheets
+ split mode: fixed bug in which -l option failed with large input files
+ split mode: new -e/-u options for specifying lead-in/lead-out
+ split mode: new -x option to extract specific tracks
+ len mode: show compression ratio for individual files as well as overall
+ len mode: show input file format for each file
+ len mode: new -c/-t options to suppress column names/totals line
+ len mode: configurable units for files and totals
+ conv mode: new -t option to convert WAVE data read from the terminal
+ fix, join and hash modes: filenames can now be sorted automatically using
a natural sorting algorithm, which orders filenames containing numbers the
the same way a human would, e.g. "t1.wav, t2.wav, ... t10.wav" instead of
"t1.wav, t10.wav, t2.wav, ...". as this was taken from the sources of GNU
ls (see 'ls -v'), it is considered stable enough to be the default.
if desired, sorting can still be done interactively, or not at all.
[global behavior]
+ user-configureable progress types (percentage, dots, spin, faces or none)
+ default overwrite option is now 'ask' - can be overridden with -O option
+ allow on-the-fly renaming when files exist and overwrite option is 'ask'
+ new -H option to show h:mm:ss.{ff,nnn} in output, instead of m:ss.{ff,nnn}
+ new -a and -z options to override default file name prefix and postfix
+ new options -w to suppress warnings and -q to suppress non-critical output
+ converted to getopt - this means long options have become single characters
+ alternate input decoder: -i "fmt decoder [arg1 ... argN]"
+ alternate output ext/encoder: -o "fmt [ext=abc] [encoder [arg1 ... argN]]"
+ in addition to the new -i/-o options above, default encoders/decoders
and/or their arguments, as well as file extensions, can be overridden
using any combination of appropriately named environment variables:
ST_<format>_DEC="decoder [arg1 ... argN]"
ST_<format>_ENC="[ext=abc] [encoder [arg1 ... argN]]"
e.g.:
% ST_SHN_DEC="shorten-2.3b" shntool len *.shn
% ST_FLAC_ENC="flake - %f" shntool conv -o flac *.wav
% ST_AIFF_ENC="ext=aif" shntool gen -o aiff -l 1:00
