ChangeLog

current:
  * fix v4l config and build variations (bug #3348888)
    - thanks to jarekczek for reporting this!
    - NB v4l1 removed from kernel as of 2.6.38
  * fix missing python thread initialization (bug #3349199)
    - thanks to jarekczek for reporting this problem!
  * fix missing configure check for Python.h (bug #3092663)
    - thanks to Zoltan Kovacs for reporting this problem!
  * fix C++ wrapper missing Symbol.quality (bug #3076524)
    - thanks to Rudy Melli for pointing this out!
  * fix C++ wrapper bugs (bug #3396068)
    - thanks to anotheruser1 for reporting this!
    - add new C++ wrapper test
  * fix avoid compositing ISBN10 data
  * add support for GraphicsMagick as ImageMagick alternative
  * mention xmlto in HACKING (patch #3297039)
    - thanks to jarekczek for the patch!
  * disable addons by default until synchronization with main symbol is working
  * fix image scanner min quality filtering
  * fix i25 buffer overrun corner case
  * fix EAN addon enables
  * fix zbarimg to accept autodetected stdin (lone dash)
  * fix Qt 4.6.3 compilation error (patch #3178244)
    - thanks to hrhristov for the patch!
  * add Python Processor support for request_size interface
  * fix Python Processor support for GIL, allowing asynchronous scanning
  * fix jpeg decoder skip handling
    - thanks to jarekczek for the patch!
  * rename dprintf macro to avoid conflicts (patch #3128538)
    - thanks to maurochehab for the patch!
  * add support for EAN/UPC 2 and 5 digit add-on symbols
    - deprecate original, unfinished add-on APIs
    - add self-checking to test_decode
  * fix support for GS1 AIs
    - thanks to jockusch for a patch!
    - add decoder/symbol "modifier" flags and config read access
    - set flags or emit GS appropriately for Code 128 FNC1
    - add iphone, java, perl, python bindings for modifiers and configs
  * add support for Code 93 symbology
  * add video size request to (Py)GTK widget (req #3034522)
    - thanks to Jerome Charaoui for the patch!
  * add support for GS1 DataBar Expanded (RSS Expanded) symbology
  * add language bindings for DataBar
  * add preliminary support for GS1 DataBar (RSS-14) symbology
  * enhance decoder reliability (EAN, Code 39, Code 128)
    - enhance decoder test
  * fix documentation of command exit codes (bug #3017322)
  * fix C++ video callback bug
    - add C and C++ processor examples
  * add per-symbology cache consistency
    - reliable symbologies decode immediately by default
    - no more need to disable cache with video
    - also fix crop bugs w/video scanning
  * add support for scanning partial images (cropping)
    - update c++, python, perl, java bindings and tests
  * fix couple of leaks
  * remove doc hints about GFDL (bug #3004411)
    - apply LGPL to API docs
  * relax Code 39 width ratio checks
  * core library portability improvements
    - remove some unnecessary C99 syntax
    - add configure checks for errno
    - enhance C++ Symbol interface
  * adjust Python module README (add examples and note DLL in path)
  * fix QR Code missing from man pages (bug #2969857)
  * cleanup decoder assertions and locking (bug #2966916)
  * add Java interface to library via JNI
    - add Java tools and JNI build tests to configure
    - fix compiler warnings from binary output changes
  * fix output of binary data in zbarimg and zbarcam
    - thanks to fukuchi for a patch!
    - add base64 format for binary xml output
  * add coarse symbol orientation information (patch #2913094)
    - thanks to Anssi for a patch!
    - add decode direction feedback to decoder
    - update C++, Python, Perl and ObjC wrappers
    - add orientation to test suites
  * fix inconsistent fourcc endian handling bugs (bug #2918994)
    - thanks to jdccdevel for a patch!
    - add fourcc conversion and parse interfaces to zbar.h
  * report QR immediately for video (no inter-frame consistency check)
  * add python distutils build infrastructure

version 0.10:
  * hotfix add MinGW import lib to Windows installer
  * attempt fix for Qt4 < 4.4
  * release updates
    - fix perl Processor init without video
  * fix window closed notification during events
    - add read_one example, fix xs compiler warnings, update perl docs
  * add result query interfaces to image scanner and processor
    - move result filtering into image scanner (from user)
    - abort output blockers when window closed
  * Windows updates
    - update installer and README for distribution of dependency libraries
    - fix applications to use binary output for xml and raw modes
  * add regression tests to makefile
  * cleanup warnings from newer gcc versions
  * fix excessive i25 QZ checks
  * add regression script
    - add zbarimg xml output for every source (even empty)
    - add edge detection to svg debug overlay
  * image scanner cleanup and minor performance enhancements
  * bug hunt and stability improvements
    - fix broken processor locks
    - fix X connection polling, revert previous separate thread workaround
    - refuse to resize windows larger than screen size
    - fix window output scaling - preserve image aspect ratio, adjust overlay
    - fix window redraw
    - fix crash on Xv image creation failure (still need XImage fallback)
    - clean up zbarimg exit cases (last image window close, missing decodes)
  * always use separate video thread when threads enabled (even v4l2)
  * add configure check for features.h
  * overlay enhancements
    - add fps to overlay
    - add overlay control to processor
    - add windows overlay drawing
  * tweak linear code position info
  * trim deep qrcode hierarchy
  * fix zero length symbol data
  * fix QR structured append result handling
  * cleanup SVG debug dump (partial)
    - some QR integration API cleanup
  * extract explicit result container, separate from image
    - remove (broken/tmp) converted image result sharing
    - add explicit symbol recycle API, update processor/widgets to use
    - cleanup and update C++ API
    - update Python/Perl APIs
    - fix broken and add new Python/Perl tests
  * cleanup QR result integration
    - add hierarchy to symbol results
    - extract symbols during text decode, preserving position and structure
    - outline QR symbols in window overlay
    - tmp fix converted image result propagation
  * factor image scanner cache and apply to QR
    - fix image scanner handler called once per-image (vs every decode)
  * QZ and clustering fixes to QR integration
    - remove qr_finder QZ checks
    - decrease center clustering threshold from 1/3 to 1/5 of length
    - add img_scanner svg debug output
    - manually add config.rpath to workaround broken autofoo
  * finish initial integration of QR Code reader from maemo-barcode project
  * zbar-side updates for QR Code integration
    - add linear scanner position interface
    - add QR finder position feedback
    - integrate QR Code reader with img_scanner
    - refactor some symbol/image interaction
    - change default scanner density to 1
    - add iconv to build infrastructure
  * initial hooks for QR Code, first pass at finder
  * fix broken builds with --disable-pthread

version 0.9:
  * hotfix configure check for Wand as well as MagickWand (bug #2848437)
  * hotfix trim extraneous MagickWand calls (bug #2848445)
  * release updates
  * fix uninitialized member in Qt widget (bug #2844846)
  * move image conversion out of image scanner up to higher levels
    (preparation for library split)
  * add symbol quality metric and image scanner position config
    - update python, perl and docs
  * compatibility fixes
    - work around ImageMagick API breakages
    - fix some OS X compile issues
  * Qt widget cleanup
    - handle video device set before window attached
    - prevent exceptions from escaping event handlers
  * more Qt window creation hook fixes
    - NB may still cause problems if video is opened before window is visible
  * finish fix for Qt init ordering (bug #2844846)
  * potential fix for display init ordering (bug #2844846)
    - new workaround for filtering bad EAN results in the image scanner
  * more testing, fixes and cleanup
    - fix v4l1
    - fix/add null processor interface
  * change default image scanner density to 1x1
    - random cleanup, vfw hang, quit key
    - fix scan_image example MSVC project
    - windows installer tweaks
  * add zbarcam to windows installer
  * major restructuring to improve platform abstraction
    - add lock, thread and timer abstractions
    - migrate specific notification locks to platform independent layer
    - fixes to vfw capture interface
    - fix window format conversion issues
    - fix some broken configure checks
    - zbarcam working in windows!
  * fix symbol leaks (bug #2820658)
    - add symbol reference counting
  * add support for binary symbol data
  * initial VFW video support
    - mostly working with uvc driver, broken for others
    - factor out common video buffer handling
    - fix processor to destroy window *before* video (ref TODO)
    - use /dev/video* VFW pseudo-devices
    - windows configure skip non-windows checks
    - prep for platform refactoring
  * fix zbarimg b&w format handling
  * fix scan (image) boundary QZ handling (bug #2807538)
    - add linear scanner manual flush API
    - linear scanner always starts/ends w/a space of width 0
    - remove artificial image scanner border
    - decoders special case 0 width space for QZ checks
    - add missing Code 128 leading QZ check
  * fix Code39 max ICS checks (bug #2807447)
    - add decoder lock owner tracking (debug)
    - update dbg_scan to match img_scanner
  * first pass installer
    - add version and icon resources for libzbar, zbarimg
  * zbarimg working in windows
    - switch to StretchDIBits over DrawDib
    - refactor some window drawing code to remove redundancies
    - make refcounts thread safe
    - clean up alloc/free bugs
  * convert zbarimg to C (cross compiled C++ cannot run w/native libraries)
    - fix DrawDib image width granularity
    - fix window resize dimensions to include decorations
    - images still inverted, otherwise zbarimg now "working" in windows
  * refactor processor implementation to support cross-platform
    - first pass windows processor (before debugging)
    - make processor locks reentrant (call APIs from callback)
  * initial Windows support for window interface
    - currently supports VFW DrawDib interface for image display
      (DirectDraw and others TBD)
    - also basic processor window creation/manipulation
    - Windows configure tests

version 0.8:
  * release updates
  * add "raw" output option (without symbology prefix) to apps (req #2671863)
  * fix Code 39 min length and document min/max configs (bug #2787925)
  * fix zbar_image_free_data windows recursion loop (bug #2796277)
  * fix uninitialized decoder configs (bug #2800163)
  * switchover from subversion to mercurial

version 0.7:
  * fix perl MANIFEST
  * release updates (version, NEWS, packaging)
  * adjust [py]gtk distributed files
  * draw new logo (rough, no Xrender yet)
  * fix Makefile.am syntax error
  * fixup some perl distribution details
  * project name change: everything "zebra" becomes "zbar"
  * remove old logo
  * add first pass python bindings!
  * fix perl mortality bug
  * add new cfg constants to perl
  * fix perl doc handler ref
  * fix processor set_active error propagation
  * add wiki examples (scan_image.*, processor.*)
  * add missing trailing quiet zone checks for ean and code39
  * add min/max code length cfg/check for i25,code128,code39,pdf417
  * add image scan density API/option
  * tweak option parser to be more strict/correct about abbreviations
  * add API to force specific video io modes (req #2293955)
  * apply patches for more broken driver workarounds (req #2293955)
  * fix(?) C++ string to fourcc conversion
  * add missing C++ wrappers
  * add additional examples to man pages (symbology enable/disable)
  * add missing options to man page synopsis
  * add missing --xml option to man pages

version 0.6:
  * hotfix broken perl install (name change)
  * add missing files to distribution
  * release updates (version, NEWS, pacakging)
  * rename perl module to Barcode::Zebra (fit with existing cpan namespace)
  * add perl documentation and example
  * add v4l version debug/test override
  * add docs for new zebracam prescale option
  * add video pre-scaling API/option to video/processor/zebracam (req #2277340)
  * add few missing APIs to perl xs wrapper
  * fix missing libjpeg #ifdef in convert
  * initial support for decoding jpeg images using libjpeg!
  * workaround broken v4l2 drivers in USERPTR mode
  * have configure double check Magick++ install (bug #2582232)
  * update README dependency list
  * fix C++ warnings in Processor
  * fixes for building DLLs with libtool under MinGW
  * automatically remove "processor" layer if poll.h is unavailable
  * test_decode portability workarounds
  * add config/compile time symbology enables
  * add low-level PDF417 decode stage - converts widths to codewords
  * add XML output option to zebracam and zebraimg
  * add sequence number image attribute, set to frame number by video
  * change v4l2 interlaced only drivers to warning instead of hard fail
  * workaround broken drivers that return error for V4L2_FIELD_NONE request
  * add some initial PDF417 hooks
  * first pass perl bindings for Processor, Scanner and Decoder
  * fix error propagation double free bug
  * add missing APIs: processor userdata and image data length
  * fix configure check for v4l2 - thanks to Simon Matter for the patch!
  * finish support for UPC-E
  * fix zebraimg to scan all pages/frames of a multi-frame input
  * fix debian packaging dependencies (bug #2070164)
  * *remove* debian directory from distribution (bug #2070164)
  * fix inappropriately installed headers and man pages (bug #2055133)
  * fix pygtk multiple definition link errors on darwin (bug #2052681)
  * fixes to configure for detecting python settings (bug #2052663)
  * remove zebrapygtk module link against libpython (bug #2052663)
  * add drag and drop support for *images* to Qt widget...unfortunately not
    very useful; every application i tried drops uri-refs rather than images
  * minor reference documentation updates

version 0.5:
  * release updates (version, NEWS, packaging)
  * add pkg-config files
  * update to latest libtool and new autoconf macros
  * cleanup library symbol exports
    * remove test programs using internal hooks
  * improve portability of some format specifiers
  * fix missing stub for --without-x - thanks to Simon Schmeisser for a patch!
  * fix --disable-pthread compile errors and warnings
  * fix XImage size mismatch background artifacts
  * fix new generated file distribution errors
  * switch Qt headers to old-style names (investigate possible Qt3 support?)
  * add independent ABI versioning for GTK and Qt widget libraries
  * reimplement widget internals to support image scanning and
    improve locking efficiency
  * add image scanning to widgets. including builtin conversions from
    toolkit image types: GtkPixbuf and QImage
  * add video opened hooks to widgets (improved use model)
  * add logo, used when there is nothing better to draw
  * add userdata to image object
  * fix image reuse cleanup bug
  * fix format specifiers in some error messages
  * enhance widget tests to support enable/disable and scan from image
  * fix broken deallocation assumptions in test_qt
  * widget API documentation (still need to hookup gtkdoc, and PyGtk docs)
  * API documentation toplevel overview
  * update configure summary for new features
  * replace all decoder assertions w/non-fatal debug spew (bug #1986478)
  * fix glib-genmarshal check
  * add first pass of Qt widget!
    - test/example in test/test_qt.cpp
    - factor video device scan to share among tests
  * more C++ integration fixes
    - additional Image ref tweaks
    - add Video.close() and Window.clear() APIs
  * fix missing image scanner handler call
  * add dereference operator to C++ Symbol
  * add count attribute to C++ Symbol
  * fix broken C++ header definitions
  * fix broken C++ Image references
  * expose internal reference counting interface
  * fix window locking bug
  * cleanup some minor memory leaks
  * convert Code 128 assertions to non-fatal warning spew
  * fix single buffer video hang (bug #1984543)
  * replace inferred video size assertion with warning message (bug #1984543)
  * add first pass of GTK widget!
  * add PyGTK widget wrapper
  * API change: modify window draw/redraw interface to improve interoperability
    with toolkits
    - add window locking for thread safety
    - zebra_window_draw() no longer actually "draws" anything
    => use window.draw to update the image from video thread
       then schedule window.redraw from toolkit GUI thread
  * fix missing C++ std lib dependencies
  * fix uninitialized handler/userdata bug in decoder
  * fix broken Code 128 checksum assertion
  * fix video destructor segfault
  * fix window destructor Xvideo errors (XID is unsigned...)
  * switch configure to use pkg-config to find most dependencies
  * API documentation updates

version 0.4:
  * release updates (version, NEWS, packaging, examples)
  * couple of portability tweaks
  * finish format conversion resize cases
  * add missing conversions
  * fix some broken conversions
  * fix some broken redraw and Xv checks
  * add decoder configuration API
    - only boolean configs currently implemented
    - integrate config option parsing w/zebracam and zebraimg
    - add config to enable/disable each symbology type
    - add optional conversions from EAN-13 to UPC-A, ISBN-10 and ISBN-13
      (disabled by default)
    - add config to emit/suppress check digit
      NB behavior change!  check digit is now emitted by default
  * related documentation updates
    - split common options to a separate entity
  * fallback to gettimeofday when POSIX timers are not available
  * image format conversion fixes
    - fix format size roundoff (NB now rounds *down*)
    - add convert and resize API to pad/crop image to specific size
      (eg, to handle Xv min/max image size)
      NB this is still not implemented for many conversions
  * fix window deletion visibility inconsistency
  * add couple processor commands
    - 'q' to delete window
    - 'd' to dump displayed image for debug
  * remove problematic includes used for v4l2 configure test
  * address compiler portability concerns w/debug print vararg macro
  * workaround v4l1_set_format() failed consistency check for broken drivers
    - change from error to warning w/more useful message
    - calc expected image size using expected value instead of driver value
  * add missing example scripts to distribution
  * add missing files for Interleaved 2 of 5
  * add support for Interleaved 2 of 5 symbology!
    - again no check digit option yet
  * increase decode window from 8 to 16 bars
    - remove Code 39 "extra bar" hack
    - add Code 39 quiet zone check
    - facilitate Interleaved 2 of 5
  * optimize character width calculations for all symbologies
  * fix image scanner bug w/lost symbols at end of scan passes
  * fix EAN-8 buffer overrun case
  * add API doc footer
  * add API documentation generated by Doxygen
    - markup, cleanup and finish writing header comments
    - setup Doxygen config file
  * add/fix window GC
  * add base support for Code 39 (no check digit or full ASCII options yet)
  * cleanup decoder locking
  * add support for EAN-8!

version 0.3:
  * add interface stub files
  * fix wait timeouts
  * fix XImage format filtering
  * fix several error handling corner cases
  * fix c++ error handling
  * add missing Window.h
  * add better hierarchy to library sources
  * build configuration fixes
  * portability fixes
  * packaging updates
  * fix zebracam beeps
  * fix some RGB component ordering and XImage format problems
  * fix window resize and redraw state problems
  * fix EAN testcase in test_decode - thanks to Steffen Kube for the patch!
  * add APIs and (hidden) zebracam options to force specific formats for debug
  * add example scripts
  * documentation updates
  * remove implementation of deprecated img_walker
  * add XImage formats
    - basic support for 16-bit depths
  * add some missing rgb format conversions
  * add basic overlay
    - currently only markers at detected scan locations (needs improved)
  * fix memory leak for converted images w/new cleanup semantics
  * migrate inter-frame consistency check from old zebracam into image_scanner
    - add API to enable/disable/flush result cache
    - add API to retrieve cache status from returned symbol
  * cleanup user_wait blocking semantics for various state combinations
  * fix bug w/v4l1 not unlinking dequeued image
  * major restructuring for improved modularity
    NB not all changes are are backward compatible!
    - modular support for v4l2 falling back to v4l1
    - flexible support for more varied image formats (incomplete)
    - added reusable abstractions for: an "image" object and associated
      metadata, a "video" input object and a "window" output object
    - added new "processor" API to wrap everything together, simplifying
      trivial application usage (including zebracam and zebraimg)
    - removed deprecated "img_walker" interface, completely replaced by
      more integrated "image_scanner" (moving toward more image processing)
    - updated/added c++ APIs, including improved iterator interfaces
  * removed SDL dependency due to various usage incompatibilities
    (most notably the inability to embed video window into a widget)
  * cleaned up zebracam and zebraimg command line processing (bug #1838535)
  * many useful enhancements thanks to proposal by mhirsch45 (req #1759923)
    including:
    - v4l2 support
    - support for UYVY image format
    - zebracam cursor hiding
    - zebracam audio mute
    - command line video device specification,
  * significant error handling and debug improvements
  * some associated documentation updates
  * various new test programs (slowly working toward more formal test suite)
  * add missing xlink namespace to dbg_scan generated output (bug #1764188)
  * qualify char types as [un]signed to avoid non-portable C ambiguity
    - thanks to mhirsch45@users.sf.net and Colin Leroy for the patches!
      (bug #1759914)
  * add integrated 2D image scanning API
    - automatically ties to internal scanner and decoder
    - results reported via new symbol API
    - deprecated previous, cumbersome img_walker interface
    - uses new simpler/faster and denser grid pattern (FIXME add 45 and 135)
    - first step toward more sophisticated image processing
  * updated zebraimg to use new ImageScanner API
    - add initial decode location tracking
  * updated zebracam to use new img_scanner API
    - extended cache to track multiple symbols per-image
    - add initial decode location tracking
    - removed scan grid overlay
  * add configure check for ImageMagick version >= 6.3.0
  * configure dumps ImageMagick, SDL and firefox versions for debug
  * add NPAPI plugin stub and build infrastructure
  * flush zebracam output after each scanned symbol
  * integrate RPM packaging spec file - thanks to V�t Hrachov� for the draft!
    (patch #1723536)
  * finally add HACKING to distribution/install (bug #1698202)
  * add extra documentation files to install (README NEWS INSTALL, etc)
  * Debian package patch for 0.2 - thanks to V�t Hrachov�:
    - add libsdl1.2-dev as a build dependency
    - update automake (>= 1:1.10) as a build dependency
    - new package

version 0.2:
  * update distribution to include debian packaging
  * add consistency checking to zebracam
  * add redundant output suppression to zebraimg
  * fix couple of Code 128 decode table bugs
  * fix reversed Code 128 decode
  * add outstanding scanner edge flush to new_scan()
    - API change: scanner reset/new_scan() now return scan/decode status
    - update zebracam and zebraimg to call this between each walker pass
      (interface still needs improvement...)
    => improves in scan results for many cases
  * fix dbg_scan filename generation so results go in local directory
  * continue Code 128 refinement
    - finish character set C decode/expansion
    - add per-character validation
    - resolve scan direction in separate postprocessing pass before handling
      ASCII conversion
    - add several img_walker passes parallel to major axis (for long symbols)
    - add simple character set C example to test_decode
  * promote zebraimg images to TrueColor so colored scan pattern always shows
  * more dbg_scan tweaks
  * significant scanner improvements
    - changed "classic" [-1 0 1] first differential kernel to [-1 1]
      to improve minimum edge resolution to single pixel elements
      => still need to do some more research and validate assumptions
    - adaptive thresholding fixes
    - adjusted filtering for better edge detection
    - separate constants out to defines (FIXME add config API?)
  * fix EAN-13 secondary determinant decoding
  * dbg_scan tweaks to make annotations smaller/more usable
  * add get_color() interface to decoder
  * annotated zebraimg scan pattern for marginally useful high-level debug
  * random include cleanup
  * cleanup 64-bit compile warnings in zebraimg (bug #1712504)
  * add first-pass partial support for Code 128
    - separate out more EAN state from shared state
    - internal interface changes
    - finish dynamic buffer management
    - add shared resource locking
    - add Code 128 to test_decode program
    => still needs additional functionality and plenty of debug
    => reading both Code 128 *and* EAN still destabilized
  * add diagnostic program test_video to dump more verbose video debug
  * incorporate documentation into Debian package
    - thanks to V�t Hrachov� for the patch!
  * fix VPATH doc builds (requires automake-1.10 and autoconf-2.61)
  * build and dist fixes
    - suppress documentation rebuilds in distributed sources
  * add Debian packaging sources - thanks to V�t Hrachov� for the patch!
  * add DocBook template and build infrastructure
  * add manpages for zebracam and zebraimg
  * add GNU standard options to zebracam and zebraimg
  * internal decoder restructuring to support additional symbologies
    - separated out 1-D decoder infrastructure into generic internal API
    - moved EAN/UPC specific decoding into it's own module
  * fix confusing configure behavior which quietly avoided building
    targets with missing dependencies(?!)
    configure will now fail with a descriptive error message if you 
    do not have Magick++ and fail to specify --without-imagemagick or
    do not have SDL and fail to specify --without-sdl
  * add configure summary describing what will be built (req #1698196)
  * fix parity encoding in test_decode and add decoded symbol output
  * introduce Code 128 symbol type
  * increase width of zebra_symbol_type_t to 16 bits
  * add HACKING (bug #1698202)

version 0.1:
  * add NEWS and ChangeLog
  * fix some config/build issues found on other boxes
  * add missing ImageWalker install
  * fix scanner runaway threshold calculation bug
  * fix zebracam/zebraimg bugs overwriting currently scanning image w/scan
    pattern
  * add c++ interface to img_walker
  * apply ImageWalker to zebraimg
  * add decoder soft reset on partial mismatch
  * finish basic decoder symbol assembly/reporting
  * add decoder symbol checksum verification
  * add callback API option to decoder for "async" symbol processing
  * add "image walker" library API to trace scan pattern over 2D images
  * apply image walker to zebracam (C++/zebraimg scan pattern still TBD)
  * add audio feedback to zebracam (still has long latency)
  * add zebracam key cmd to dump frame to file (for debugging)
  * fixes for decoder/scanner reset/new_scan
  * fixes to scanner initialization and algorithm tweaks
  * made decoder less sensitive to violated quiet-zone
  * apply zebraimg workaround for imagemagick image display bug
  * add string names for symbol types to library and API to access them
  * add dbg_scan test program for visually debugging image scanner (and
    decoder)
  * add test_walk for basic image walker sanity/debug
  * removed recursive makes in favor of monolithic build
  * renamed some makefiles accordingly
  * finished some final symbol data construction
  * added result callbacks to decoder APIs for data reporting
  * zebraimg hooks into callback
  * zebracam still seems to "hang" in undecodeable state?
  * populate svn with current sources.
  * most of the basic functionality is included and working.
  * still need to combine final decode data, finish addons, etc (see TODO).
  * c++ wrappers are included and tested, but API may need tweaked.
  * zebraimg and zebracam basically working but need cleanup/polish.
  * need to create some basic documentation...
  * initial repository layout