path: root/ChangeLog

v1.7.5 - 2021-12-06
-------------------
Kim Woelders (83):
      Build .xz instead of .bz2 release tarball
      Drop imlib2-config (use pkg-config)
      Test: Add some minimal regression testing
      Test: Fix dist
      imlib2_load: Add option to use imlib_load_image_immediately()
      JPEG loader: Cosmetics
      JPEG loader: Parse EXIF data and handle orientation
      autofoo: Drop support for libungif
      Imlib.h: Add version macros
      imlib2_load: Add verbose option
      Fix build (Imlib2.h is now built)
      test: Add icon-64.pbm
      test: Add test_load_2
      imlib2_view: If verbose show error message on failure
      imlib2_load: Use clock_gettime() when available
      debug: Infrastructure
      debug: Add some debug related to file access and image loading
      Introduce im->fsize
      PNG loader: Cosmetics
      PNG loader: Use mmap() during signature check
      TIFF loader: Use mmap() during signature check
      TGA loader: Use im->fsize, cosmetics, debug
      WEBP loader: Cosmetics
      WEBP loader: Use mmap() for loading
      FF loader: Cosmetics
      FF loader: Use mmap() for loading
      ARGB loader: Cosmetics
      ARGB loader: Use mmap() for loading
      BMP loader: Cosmetics
      BMP loader: Use mmap() for loading
      ICO loader: Cosmetics
      ICO loader: Use mmap() for loading
      LBM loader: Cleanups
      LBM loader: Use mmap() for loading
      PNM loader: Use mmap() for loading
      XBM loader: Fix potential buffer overrun
      XBM loader: Cosmetics
      XBM loader: Use mmap() for loading
      XBM loader: Ignore comments and other stuff in header
      XPM loader: Use mmap() for loading
      test: test_load improvements
      imlib2_view: Add option to cache images
      Introduce UPDATE_FLAG()
      Introduce ARRAY_SIZE()
      Loader cosmetics
      Loader loading: Tweaks
      Loader loading: Move to __imlib_FindBestLoaderForFormat()
      Loader loading: Move __imlib_GetLoaderList()
      Loader loading: Minor loader lookup refactoring
      Loader loading: Avoid always loading all loaders
      Loader loading: Don't bother looking up load() if we have load2()
      WEBP loader: Enable loading animated images (first frame by default)
      configure.ac: Fixup after recent change
      Drop unnecessary free() NULL argument checks
      Remove some unneeded headers
      Rename X11 related files for clarity
      Move ImlibImagePixmap population to __imlib_AddImagePixmapToCache()
      test: Add X11 drawable grabbing test
      Move pixmap stuff to x11_pixmap.c/h
      Trivial changes in __imlib_Grab...() function prototypes
      Refactor imlib_create_scaled_image_from_drawable()
      Revert "Refactor imlib_create_scaled_image_from_drawable()"
      Refactor imlib_create_scaled_image_from_drawable() - take 2
      x11_grab.c: Rename source/destination variables for clarity
      Fix y-upscaling in imlib_create_scaled_image_from_drawable()
      test_grab: Cleanups, cosmetics
      test_grab: Add scale-down tests
      Only set MAINTAINERCLEANFILES in top-level Makefile.am
      imlib2_view: Enable grabbing/viewing drawables
      x11_grab.c: Cosmetics
      x11_grab.c: Introduce function to get shape mask
      Speedup in imlib_create_scaled_image_from_drawable()
      Avoid signedness warning
      Avoid "exceeds maximum object size" warning
      blend.c: Tweaks, cleanups
      Simplify build wrt. asm files
      test_save: Check images with alpha too
      test: Add some scaling/rotation tests
      Refactoring around mmx and rotate function calls
      Refactoring around mmx and scaling function calls
      Refactor condition for using assembly functions
      Update doc for imlib_load_image_fd()
      1.7.5

NRK (1):
      WEBP loader: fix key selecting last frame

v1.7.4 - 2021-09-17
-------------------
Kim Woelders (15):
      imlib2_view: Move property stuff to separate file
      imlib2_view: Cleanups
      imlib2_view: By default scale large images to fit on screen
      imlib2_view: Add some debug
      imlib2_view: Fix issue with new default scaling
      WEBP loader: Remove forgotten debug printout
      WEBP loader: Rename fd variable to be same as everywhere else
      LBM loader: Fix potential out-of-bounds memory access
      GIF, TIFF, WEBP loaders: Fix loading if filename does not have usual suffix
      Revert "GIF, TIFF, WEBP loaders: Fix loading if filename does not have usual suffix"
      GIF, TIFF, WEBP loaders: Fix loading if filename does not have usual suffix - take 2
      Add script to generate Changelog
      Update Changelog to new format
      image.c: Use the LOAD_... macros to check loader return values
      1.7.4

v1.7.3 - 2021-08-08
-------------------
Kim Woelders (2):
      autofoo: Resurrect non-pkg check for bzip2
      1.7.3

v1.7.2 - 2021-07-27
-------------------
Kim Woelders (31):
      Remove some unnecessary X_DISPLAY_MISSING stuff
      rend.c: Remove some pointless lines
      Add XBM loader
      Add imlib2_load and /build to .gitignore
      Remove obsolete and unused AC_HEADER_STDC
      Restore file:key functionality
      ICO loader: Fix (disabled) debug stuff
      ICO loader: Enable specifying ico image index by key
      Remove unused Context functions
      context.c: Fix potential segv
      LBM loader: Fix handling of missing RLE data
      Fix clang-analyzer warnings - bin (trivial)
      Fix clang-analyzer warnings - loaders (trivial)
      Fix clang-analyzer warnings - loaders (suppress bogus)
      Fix clang-analyzer warnings - lib (mostly trivial)
      Fix clang-analyzer warnings - lib (less trivial)
      autofoo: Move more to pkg-config
      TIFF loader: Drop use of libtiff defined types deprecated in libtiff-4.3.0
      TGA loader: Fix loading small images without footer
      Spec file: Add git tag to rpm file name (if built from git checkout)
      Cleanups: while->for loops (loaders list)
      Cleanups: while->for loops (context list)
      Cleanups: while->for loops (im->tags list)
      Cleanups: while->for loops (images list)
      Cleanups: while->for loops (pixmaps list)
      Drop unused Imlib_Object_List:last
      Add feature to build with ASAN (--enable-gcc-asan)
      Correct (disabled) debug printouts
      Loaders: Remove unnecessary headers
      WEBP saver: Use fopen() etc. like all other savers
      1.7.2

v1.7.1 - 2020-12-09
-------------------
Daniel Kolesa (1):
      Fix big endian build

Kim Woelders (6):
      rend.c: Fix __imlib_generic_render() when jump != 0
      grab.c: Support 30bpp display in __imlib_GrabXImageToRGBA()
      WEBP loader: Add initial signature check
      XPM loader: Get transparency right when doing header-only loading
      Silence a couple of sign-compare warnings
      1.7.1

v1.7.0 - 2020-08-01
-------------------
Alexander Volkov (3):
      GIF loader: Don't close file descriptor twice
      Introduce imlib_load_image_from_fd()
      Don't rescan loaders

Kim Woelders (49):
      XPM loader: Major speedup for cpp > 2
      imlib2_load: Properly check non-full loads (load data too)
      imlib2_load: Use getopt()
      imlib2_load: Add repeated load option
      Simplify __imlib_FileExtension()
      Refactor many __imlib_File...() functions to use common __imlib_FileStat()
      Drop the __imlib_IsRealFile() file check in __imlib_File...() functions
      image.c: Add some space for readability
      image.c: Remove some unnecessary clearing of calloc'ed structs
      image.c: Rework some obscure file name stuff in __imlib_SaveImage()
      image.c: Don't strdup() real_name when not necessary in __imlib_LoadImage()
      image.c: Use real_file to get file time
      image.c: Introduce __imlib_ErrorFromErrno()
      image.c: Use loader return value, not im->w to determine load success
      Loader cleanups
      Saver cleanups
      image.c/h: Cleanups
      image.c: Move image tag functions to separate file
      image.c: Move loader functions to separate file
      image.c: Enable non-dirty pixmap cache cleaning
      image.c: Minor refactoring of pixmap cache cleaners
      image.c: Move data_memory_func assignment to better place
      imlib2_view: Various tweaks
      Fix loader cleanup breakage (gif)
      image.c: Remove redundant pixmap unref
      image.c: Add infrastructure to simplify progress handling
      Loaders: Simplify/fix progress handling
      Savers: Simplify progress handling
      Introduce __imlib_LoadEmbedded()
      Introduce __imlib_LoaderSetFormats()
      Make ImlibLoader struct opaque
      autogen.sh: Add -n as alternative to NOCONFIGURE
      Fix enum conversion warnings (gcc10)
      JPG, PNG loaders: Avoid clobber warnings
      Add a couple of consts
      TIFF loader: Minor speedup
      ID3 loader: Some mostly cosmetic rearrangements
      GZ, BZ2 loaders: Accept more file names
      __imlib_FileExtension: Use basename if there are no dots
      Revert "JPG, PNG loaders: Avoid clobber warnings"
      JPG, PNG loaders: Avoid clobber warnings - Take N+1
      Add infrastructure for new loader entry - load2()
      Move loaders to load2()
      Reduce number of stat() calls during load
      configure.ac: Drop initial config.cache removal
      imlib2_load: Optionally use imlib_load_image_fd()
      Fix build without X11
      Remove a couple of unused includes
      1.7.0

Tobias Stoeckmann (2):
      ICO loader: Do not crash on invalid files
      ICO loader: Handle malloc failures

v1.6.1 - 2019-12-13
-------------------
Kim Woelders (11):
      gz loader: Use FILE, not fd
      gz, bz2 loaders: Fix recent breakage when file name has more than two dots
      Quit on 'q' or 'esc' key press in all imlib2_... test utilities
      Rename imlib2_test_load to imlib2_load
      imlib2_load: Optionally write to stderr instead of stdout
      imlib2_view: Add progress debug options
      Enable specifying loader/filter paths with environment variables
      BMP loader: Remove some bogus conditions
      XPM loader: Minor optimization for cpp > 2
      LBM loader: Fix header-only loading
      1.6.1

Luiz Carlos Ramos (1):
      BMP loader: Fix size calculation when saving files

v1.6.0 - 2019-11-24
-------------------
Alexander Volkov (1):
      Allow to use custom memory management functions for loaded images

Kim Woelders (64):
      Add __imlib_LoadImageWrapper() handling all load() calls
      imlib2_conv: Report error on save failure
      Autofoo cosmetics
      Trivial cleanups in imlib2_... test programs
      Add imlib2_test_load program
      Cleanups in load() functions
      Centralize handling of im->format
      Sort loaders in Makefile.am
      Remove obsolete dmalloc stuff
      Move SWAP.. macro definitions to common.h
      Use common PIXEL_ARGB() macro to compose pixels
      Add new ICO loader
      Spec file simlifications and cleanups
      Fix memory leak in imlib_list_fonts()
      XPM loader: Refactor exit cleanup handling
      XPM loader: Fix potentially uninitialized pixel data
      XPM loader: Fixup after "Refactor exit cleanup handling"
      Revert "XPM loader: Fix potentially uninitialized pixel data"
      XPM loader: Cosmetics (reduce indent level)
      XPM loader: Fix several colormap issues
      XPM loader: Simplify pixel value handling
      XPM loader: Add missing pixels (malformed xpm)
      XPM loader: More simplifications
      JPG loader: Refactor
      JPG loader: Do proper CMYK conversion
      Add new WebP loader
      Remove pointless im->data checks in loaders
      WepP loader: Fix memory leak in error path
      JPG loader: Fix memory leaks in error paths
      Fix ABI break
      ICO loader: Add binary flag to fopen()
      JPG loader: Refactor error handling
      Rename/add byte swap macros
      BMP loader: Major makeover - numerous bug fixes and feature enhancements
      Miscellaneous imlib_test_load tweaks
      GZIP loader: Check filename before uncompress
      imlib2_test_load: Fixup after recent change
      Re-indent everything using indent-2.2.12
      TGA loader: Refactor
      Eliminate WRITE_RGBA()
      Simplify autogen.sh
      Simplify pixel color handling in api.c
      Use pixel instead of r,b,g,a in __imlib_render_str()
      Use macro for pixel color access in savers
      Eliminate READ_RGBA()
      XPM loader: Accept signature not at the very start of the file
      Simplify loader lookup functions
      imlib2_view: Enable selecting next/prev using keys too
      imlib2_view: Fix event processing bug
      imlib2_test_load: Fixup recent breakage for real
      imlib2_test_load: Check progress conditionally
      imlib2_view: Add verbose option, quit on Escape too
      TGA loader - Mostly cosmetic refactoring
      TGA loader: More mostly cosmetic changes
      TGA loader: Support horiontal flip
      TGA loader: Add simple 16 bpp handling
      TGA loader: Tweak error handling
      ICO loader: Fix non-immediate loading
      Remove __imlib_AllocateData() w,h args
      imlib2_view: Fix next/prev selection if last/first image is bad
      ICO loader: Fix memory leak in error path
      XPM loader: Correct signature check (avoid accessing unset data)
      gz, bz2 loaders: Simplify, eliminate unnecessary strdups, cosmetics
      1.6.0.

Olof-Joachim Frahm (欧雅福) (2):
      Check filename before opening archive file.
      tga loader: implement handling of palette

Ralph Siemsen (1):
      loader_tga: fix regression in RLE raw byte handling

v1.5.1 - 2018-03-17
-------------------
Kim Woelders (13):
      Fix build without HAVE_X11_SHM_FD (T6752)
      XPM loader: Fix potential use of uninitialized value (T6746)
      BMP loader: Fix infinite loop with invalid bmp images (T6749)
      PNM loader: Simplify (fixing ASCII format parsing issues T6751)
      BMP loader: Fix warnings found with -O3
      Maximum image dimension should be 32767, not 32766
      PNG loader: Correct various error handling cases
      Add missing const to imlib_apply_filter() script argument
      Warning fixes in imlib2_... programs
      imlib2_view: Limit window dimensions to 32767
      grab.c: Fix gcc8 warning
      imlib2_conv.c: Fix gcc8 warning
      1.5.1.

v1.5.0 - 2018-02-22
-------------------
Alexander Volkov (3):
      put a check for shared memory inside __imlib_ShmGetXImage()
      introduce __imlib_ShmDestroyXImage() instead of __imlib_ShmDetach()
      Add support for MIT-SHM FD-passing

Kim Woelders (19):
      XPM loader: Fix incorrect image invalidation.
      Make some more functions static.
      Introduce __imlib_LoadImageData()
      Remove redundant CAST_IMAGE()
      imlib2_grab: Always use imlib_create_scaled_image_from_drawable() to grab image
      imlib_create_scaled_image_from_drawable(): speed up 1:1 case
      imlib_create_scaled_image_from_drawable(): Drop shape handling if unshaped
      Indent
      Autofoo cosmetics
      Strip trailing whitespace, cosmetics
      Fix potential OOB memory access if border elements are negative
      Fix potential OOB memory access if border sizes exceed image dimensions
      Introduce IMLIB2_SHM_OPT to enable overriding/testing SHM modes
      Add IMLIB2_XIMAGE_CACHE_COUNT to enable testing the ximage cache
      Refactor the XImage cache
      Add imlib_get_cache_used()
      Expose XImage cache control functions
      Drop -Waggregate-return
      1.5.0.

v1.4.10 - 2017-04-15
-------------------
Kim Woelders (3):
      PNM loader: Fix reading PNM bitmaps.
      Fix missing break.
      1.4.10.

Tobias Stoeckmann (3):
      Avoid out of boundary operations while parsing xpm
      Properly release resources on error path
      Prevent OOB read with large file support on 32 bit

v1.4.9 - 2016-04-29
-------------------
Alexander Smirnov (1):
      Fix cross-endianness for masks (1-bit depth images)

Bernhard Übelacker (1):
      gif: fix oob reads w/bad colormaps

Kim Woelders (16):
      Fix "assuming signed overflow does not occur" warning.
      Fix some "variable set but not used" warnings.
      Fix some "variable might be clobbered" warnings.
      Fix off-by-one OOB read in __imlib_MergeUpdate().
      Revert "gif: fix oob reads w/bad colormaps"
      GIF loader: Fix out-of-bound reads from colormap.
      GIF loader: Remove check made redundant by previous commit.
      GIF loader: Reduce progress checks from per-pixel to per-row.
      GIF loader: Indent.
      Fix potential divide-by-zero in imlib_image_draw_ellipse().
      Make a number of functions static.
      Eliminate pImlibExternalFilter type.
      Trivial file function cleanups.
      Move __imlib_ItemInList() to file.c.
      Fix various potential OOM crashes.
      1.4.9.

Yuriy M. Kaminskiy (3):
      Fix integer overflow resulting in insufficient heap allocation
      loader_xpm: remove nonsense/impossible/broken condition
      Harden API and internals against overly large images

v1.4.8 - 2016-03-12
-------------------
Chloe Kudryavtsev (1):
      add time.h include to common.h for time_t

FRIGN (1):
      Add a Farbfeld loader

Kim Woelders (4):
      Remove redundant config.h.
      Add compile to MAINTAINERCLEANFILES.
      Indent.
      1.4.8.

Quentin Rameau (2):
      Fix bz2 loader filename check
      Fix zlib loader filename check

v1.4.7 - 2015-04-04
-------------------
Fabian Keil (14):
      Prevent division-by-zero crashes
      imlib_conv: Use proper buffer size to prevent invalid write of size one
      loader_gif: Don't read uninitilized memory in case of invalid input
      loader_gif(): Abort gif parsing if DGifGetLine() fails
      Fix segfault when opening input/queue/id:000007,src:000000,op:flip1,pos:51 with feh
      Make IMAGE_DIMENSIONS_OK() more restrictive
      load_pnm: Deal with fread() errors consistently
      __imlib_LoadImage(): Additionally check loader_ret to detect loader failures
      loader_tga: Abort file loading if the file obviously isn't large enough
      imlib_save_image(): Check loader return code for errors
      loader_tga.c: Properly signal if decoding uncompressed BGRA data failed
      loader_tga.c: Properly signal if decoding RLE compressed data failed
      imlib_save_image_with_error_return(): Check loader return code to prevent use of unitialized memor
      load_gif: Make sure rows isn't used partly unitialized

Heiko Becker (1):
      GIF loader: Fix for libgif version 5.1

Kim Woelders (20):
      GIF loader: Fix for libgif version 5.
      Remove OS/2 support.
      Remove empty format.c/h.
      Header file cleanups.
      GIF loader: Simplify error handling.
      GIF loader: Fix segv on images without colormap.
      Revert "__imlib_LoadImage(): Additionally check loader_ret to detect loader failures"
      Revert "loader_gif(): Abort gif parsing if DGifGetLine() fails"
      JPEG loader: Fix load() return code when only reading header.
      BMP loader: Simplify pixel fetch.
      Autofoo update (AC_PROG_LIBTOOL -> LT_INIT).
      Autofoo macro cleanups (ac_->ec_).
      Change a number of internal function names.
      Indent.
      Don't include Imlib2.h indirectly via image.h.
      Remove __hidden.
      Enable visibility hiding by default.
      Indent the remaining unindented files.
      Generate a ChangeLog using git shortlog between existing tags.
      1.4.7.

Michał Górny (1):
      fix -I flags to support building out-of-source

Mike Frysinger (5):
      simplify --enable-visibility-hiding handling
      imlib2-config: delete old reference to @my_libs@
      check return value of fread/write funcs
      fix X_DISPLAY_MISSING redefined warnings when X is disabled
      do not link with X libs when X is disabled

v1.4.6 - 2013-12-21
-------------------
Cedric BAIL (2):
      autotools: move to AC_CONFIG_HEADERS.
      trunk: remove use of AM_PROG_CC_STDC as AC_PROG_CC does it.

Kim Woelders (20):
      TGA loader: Fix for corrupt RLE format.
      Change INCLUDES to AM_CPPFLAGS.
      Fix drawing of closed polygons in certain situations (ticket 2309).
      Cosmetics for readability.
      Minor optimisation (avoid bogus valgrind complaint).
      Update configure.ac (mostly suggestions by autoupdate).
      Update .gitignore.
      Set warning options when using gcc.
      Silence compiler warnings (lib).
      Silence compiler warnings (loaders).
      Silence compiler warnings (filters).
      More warning fixes (64 bit).
      Don't set -std=gnu99.
      Autofoo updates.
      Indent.
      Deuglification.
      Cosmetics (reduce deep indentation level).
      Simplify some more and fix certain case of cpp > 2.
      Remove ChangeLog and obsolete rule to generate it.
      1.4.6.

v1.4.5 - 2011-08-15
-------------------
Carsten Haitzler (1):
      no more debian dir in our svn - debian now handles this tehmselves.

Kim Woelders (13):
      imlib2_view stuff.
      Avoid some duplicated code.
      Fix reading tiff images with orientation other than ORIENTATION_TOPLEFT (ticket 563).
      Fix imlib_render_image_part_on_drawable_at_size() ...
      Fix pnm image loading in certain situations (ticket 721).
      Better image cache time stamp test.
      Indent (most) .c files.
      Fix loading of .pbm's (P4) when width is not a multiple of 8.
      PNM loader: Refactor around progress and exits.
      Fix program linking (ticket 725).
      Avoid referencing /usr/X11R6 when x_dir is not set.
      1.4.5.
      Update ChangeLog.

Lucas De Marchi (7):
      Remove unneeded code with notnull.cocci script
      FORMATTING
      Apply memset-calloc.cocci
      Convert (hopefully) all comparisons to NULL
      Revert and re-apply badnull patch
      Fix common misspellings
      Fix typos

Mike Frysinger (2):
      png loader: use png_jmpbuf() macro rather than hitting jmpbuf directly
      png loader: do not hit "interlaced" member of the png ptr struct

v1.4.4 - 2010-05-05
-------------------
Carsten Haitzler (1):
      minor buglet - w * w - duh!

Kim Woelders (6):
      No need for libtoolize twice?
      Remove unused __imlib_FileField functions.
      Add rule to generate ChangeLog.
      New ChangeLog.
      1.4.4.
      Update ChangeLog.

v1.4.3 - 2010-03-14
-------------------
Carsten Haitzler (4):
      crash fix.
      fix the copying license to
      fix bufferoverflow in id3 loader.
      handle modified dir date of 0

Christopher Michael (3):
      Fix blank line after trailing space warning.
      Add m4 macro dir as suggested by autoconf.
      Enable silent rules in building imlib2

Daniel Kolesa (11):
      Removed debian things like rasterman did in some others from Makefile.am and configure.ac/in(and modifed AUTHORS, autogen.sh and README in E-MODULES-EXTRA) in
      Removed debian subfolders - prepared for a new set of debian subdirs.
      Added new set of debian subdirs.
      prepare debian rules files for new chmod permissions.
      Added debian/rules files with a new permissions set(755).
      Updated debian stuff everywhere.
      forgot to add --prefix=/usr
      i found it is useless and cdbs is setting up prefix automatically.
      Updated debian stuff(grew Standards-Version, updated python-ecore rules file)
      Updated Standards-Version
      Fixed debian stuff everywhere.

Gustavo Sverzut Barbieri (1):
      there is no m4 dir anymore.

Kim Woelders (29):
      Remove obsolete libltdl dir.
      No need for empty NEWS.
      No need for empty NEWS ... and remove it!.
      Remove imlib2-config (use pkgconfig).
      Revert imlib2-config removal. There still may be systems without pkgconfig out there.
      Remove redundant BUILD_X11 tests, indent.
      Don't build static libs for modules.
      Refactor shm stuff.
      Fix alignment error on amd64 (patch from Erik Boettcher).
      Fix big endian bug in bmp loader (jogness, ticket 195).
      Fix imlib_font_query_size width calculation when there are undefined  codepoints (ticket 230).
      Relax 8192 pixel dimension limit (ticket 361).
      Relax 8192 pixel dimension limit (ticket 361).
      Remove unimplemented imlib_clip_line prototype and documentation (ticket 379).
      Fix build for x86 on x86_64.
      Correct image dimension check (pixels are four bytes).
      Use PNG_CFLAGS when compiling png loader (ticket 449).
      Fix compiling .S files when using older automake (ticket 449).
      Remove incorrect test for X11/X.h.
      Fix for libpng-1.4 (png_check_sig->png_sig_cmp).
      Add .gitignore.
      Fix excessive checking for shm extension in __imlib_GrabDrawableToRGBA().
      Only check XShmAttach once in __imlib_ShmGetXImage().
      Avoid cast.
      Oops - forgot extern.
      No need to touch README.
      Remove ancient ChangeLog.
      Remove .cvsignores.
      1.4.3.

Michael Jennings (1):
      Fix build.

Mike Frysinger (1):
      imlib2: bumpmap: link against -lm since we use sin()/cos() functions

Vincent Torri (1):
      aclocal flag could be needed. I prefer being polite and saying nothing about libtool...

v1.4.2 - 2008-10-22
-------------------
Carsten Haitzler (1):
      ok. i think most binary files are now fixed.

Kim Woelders (12):
      Various loader fixes (Marcus Meissner, bug 494).
      Oops - correction.
      Loader fixes based on patch from Hans de Goede/Fedora. Fix off by one error in check (tga loader).
      Introduce imlib_context_disconnect_display().
      Return value is not a pointer.
      Ignore+-.
      pnm loader fix (Marcus Meissner, ticket 25).
      xpm loader fix (Marcus Meissner, ticket 28).
      Fix incorrect event loop causing 100% load until an X event is received.
      Indent.
      Lets make it at least 1.4.2 next time.
      Fix loading of jpeg files with 4 color components (kntriant, ticket 84).

Michael Jennings (1):
      Tue Oct 21 21:28:27 2008                        Michael Jennings (mej)

Mike Frysinger (1):
      fix from Hans de Goede to look for /usr/share/X11/rgb.txt as most modern systems are using now

Peter Wehrfritz (1):
      configure.in -> configure.ac

v1.4.1 - 2008-06-10
-------------------
Carsten Haitzler (12):
      fontset patch from winfred
      patches for imlib2 and e.
      ASPARAGUS!
      1. fix a lot of things so they pass make distcheck - so many things have broken. guys - need to be more careful! 2. asparagus 3. some extra docs/comments for evas
      various patches from the devel mailing list in - and fixed where needed.
      font chaining patch
      round as a #define - and xpm loader has extra rgb.txt sourc
      different png loading to fix png greyscale loads
      asparagus!
      push initial ctxt too
      asparagus! also pass distcheck and have common autofoo init that is consistent for package, version etc.
      2 possible security vulns fixed. should probably release new version with these.

Davide Andreoli (1):
      Update doxy style

Falko Schmidt (3):
      remove duplicate line. clean install files for test package.
      fix test package and clean up some install files.
      generalize imlib2 library install files

Kim Woelders (19):
      Back out fontset patch.
      Make code indentable.
      Add indent profile.
      Indent C files.
      Indent, rewrap long comment lines.
      Mark some (new) functions deprecated as they are likely to be removed (see bug 118). Change some names to match coding style a bit more.
      Push initial ctxt - continued, wasn't done everywhere.
      Handle some out-of-memory situations without crashing.
      Oops - missed one (malloc check).
      Indent.
      Remove restriction to 8 bits per sample (suggested by David A. Gatwood, bug 374). Set stopOnError (seems like the proper thing to do?).
      Fix destination image loading in imlib_image_copy_alpha_to_image() (Victor Paesa - bug 474).
      Remove incorrect test in __imlib_copy_alpha_data() (Victor Paesa - bug 475).
      Fix destination image loading in __imlib_BlendImageToImageSkewed() (Victor Paesa - bug 479).
      Fix destination image loading in __imlib_BlendImageToImage() (Victor Paesa - bug 480).
      Fix HSV color conversion so it matches the API documentation (Dariusz Knocinski).
      Map after resize to avoid initial placement silliness.
      MAINTAINERCLEANFILES: aclocalm4->aclocal.m4
      Add support for TrueType Collections (suggested by Arne Goetje, bug 487).

Michael Jennings (2):
      Mon Mar 10 22:38:16 2008                        Michael Jennings (mej)
      Mon Jun  9 22:46:01 2008                        Michael Jennings (mej)

Mike Frysinger (3):
      fix running with libtool-2.2+
      if png is disabled, set png_ok so the summary display is nice
      add X status to the summary display

Sebastian Dransfeld (2):
      Fix signed warning.
      Use pkg-config to check for png

Vincent Torri (1):
      * improve autotools stuff  * move libtool versioning from src/bin/Makefile.am to configure.in  * formatting

v1.4.0 - 2007-05-06
-------------------
Carsten Haitzler (6):
      fix possible overflow in tga loader
      fix width and height checks in case of buffer overflow.
      fix clip?
      line patch for imlib2 from john williams.
      asparagus - pass distcheck.
      up to 1.4.0 ...

Kim Woelders (3):
      Fix major memory leak in xpm loader.
      In imlib_render_pixmaps_for_whole_image() and imlib_render_pixmaps_for_whole_image_at_size() don't complain about NULL mask_return. A NULL mask_return is handled appropriately down the line and simply suppresses rendering of a mask.
      Add option to build with visibility=hidden + associated fixups.

Michael Jennings (1):
      Sat Dec  2 22:59:35 2006                        Michael Jennings (mej)

Mike Frysinger (4):
      need AM_PROG_AS as pointed out by automake-1.10/Marc-Andre Landry
      use -std=gnu99 in CPPFLAGS if compiler supports it
      need to call AC_PATH_X to make sure have_x is set early enough and then we need to not clobber it when checking for X11/X.h
      cleanup and simplify ... this should also fix the preprocessor paste error seen on x86

Nathan Ingersoll (1):
      Protect against segfaults if XImage allocation fails. Return usable status to the API caller so it can handle the error condition.

Sebastian Dransfeld (2):
      No longer needed.
      Remove unused files.

Tilman Kuepper (1):
      don't _require_ freetype2

v1.3.0 - 2006-09-30
-------------------
Ben Rockwood (1):
      Solaris workarounds.

Carsten Haitzler (9):
      fix bmp loader advances
      bmp fixes
      cvs is back up.. time for some asparagus!
      if u run out of memory - actually free stuff
      1. autofoot patches. 2. fix maximize to work again. :)
      fix blah-config includes
      remove openembedde pkg info - old and dead
      fix x detect
      asparagus - forgot to commit

David Walter Seikel (1):
      .cvsignore++

Falko Schmidt (1):
      Fix some dependency issues regarding xlibs-dev.

Horms (3):
      Save and restore autogenerated changelog when debian/rule's clean target runs. Otherwise the following breaks because make distclean removes autogenerated files, but debian/rules expects the changelog to always be there:
      Need not depend on libc6-dev | libc-dev as it is in build-essential
      The section of a library's -dev package is generally libdevel

Kim Woelders (10):
      libImlib2.so minor number should have been bumped. Bad raster :)
      Patch from Dmitry Antipov: - Visibility hiding - Move common asm macros to asm.h - Fix some typos.
      Trivial warning fixes.
      imlib_copy_drawable_to_image() and imlib_create_image_from_drawable(): - When mask is set to (Pixmap)1 (and the context drawable is a window)   the window shape is used for image alpha.
      Add some options:  -id <drawable> to grab other than root window.  -w/width       set output image width.  -h/height      set output image height.  -noshape       do not use window shape.  -help          show usage.  -v             show info about the grabbed drawable.
      Remove effectively unused actual_depth variable.
      Enable grabbing of ARGB drawables.
      Fix pixmap and gc caching when rendering to drawables with different depths.
      Enable setting alpha threshold used when rendering masks (was fixed 128).
      Set Release like most other places in the e17 tree. Remove XFree86-devel requirement causing trouble when using xorg.

Sebastian Dransfeld (2):
      Remove unused variables.
      EAPI

v1.2.2 - 2006-03-18
-------------------
Azundris (1):
      * update specs

Ben Rockwood (1):
      ID3 support is reported as MP3 support, which is confusing and perhaps misleading, updated reporting to be more accurate.

Carsten Haitzler (31):
      it's been a while, so it was time for some ASPARAGUS on our plates
      1. id3 album cover loader patches 2. i reduced list note memory usage by 20% - shoudl work better with malloc as ti is now a power of 2 as well 3. optimised evas internals to make use of event freezes to make e17'sw menu popups a LOT snappier 4. fixed using last member of list nodes - bad - shoudl use api as this is private stuff really 5. added config profile stuff to e17  u can literally maintain multiple config profiles and choose which one at any time etc.
      fix digikam crash
      dont modify alpha if img has no alpha
      this SHOULD fix cross-endianness issues (serve and client not same endianess) ...
      expand tmp image
      apps/e/enlightenment.spec CVS: apps/entice/configure.in apps/entrance/configure.in CVS: libs/ecore/configure.in libs/edb/configure.in libs/embryo/configure.in CVS: libs/emotion/configure.in libs/epeg/configure.in CVS: libs/epsilon/configure.in libs/esmart/configure.in CVS: libs/etox/configure.in libs/evas/configure.in libs/imlib2_loaders/configure.in CVS: ---------------------------------------------------------------------- ��������� :)
      2nd asapargus for the weekend :)
      ramkumar's id3 updates
      amd64 alignment fix
      id3 .spec additions
      include math! and stuff.
      a bit of asparagus action for shits & giggles
      asparagus!
      after some quiet on the western front - asparagus.
      asparagus. and make distcheck passes again.
      asparagus - again. lots fo leak fixes and other fixes have been happening, so i think an asparagus is a good idea - sorry package config people. :)
      imlib2 cross-endianess fix from Geoffrey Giesemann
      another amd64 name
      movdqa -> movdqu where appropriate
      bmp loader in cvs
      bmp loader in cvs
      tiff loader fix
      oops - typo. fix. works now.
      already in AUTHORS :)
      tiff patch - simon
      big fat asparagus!
      fix tiff off-by-1 pixel
      asparagus!!!
      asparagus!!!
      1.2.2 of imlib2 - for kwo :)

Horms (1):
      fix typos

Kim Woelders (7):
      Be quiet if the file is rejected because it doesn't have a .mp3 extension.
      Quiet.
      Bad fix - Revert.
      1) Quit silently if file doesn't exist. 2) Don't close if open failed (fixes segv).
      Avoid useless graphics exposure events from imlib_create_scaled_image_from_drawable().
      Another attempt to fix rendering of certain(?) fonts.
      Fix colormap when grabbing 8 bit depth pixmaps.

Michael Jennings (2):
      Thu Sep  1 16:53:13 2005                        Michael Jennings (mej)
      Thu Sep  8 17:12:14 2005                        Michael Jennings (mej)

Mike Frysinger (23):
      add error checking to all autogen scripts
      fix whitespace
      fixes from the PaX guys to make sure we dont have executable stacks
      make sure the masks are in the .data section like they should be
      merge PIC-happy code by PaX/Kevin Quinn/me
      ignore amd64 objects
      fix typo in IMMQ cleanup count as pointed out by Peter Beutner in Gentoo Bug 102519
      fix for cygwin building (and anyone else who doesnt define RTLD_LOCAL)
      cleanup x86/amd64 autofoo output
      allow users to control whether jpeg/png support
      sneak in a hack to remove CXX/F77 checks to improved configure speed
      allow user to control tiff/zlib/bzip2/id3 support
      touchup amd64/x86 asm handling, unify all the autodetection warnings, make gif support configurable, and default to giflib instead of old libungif
      fix by Tres Melton to address 64bit errors: dont cast pointers as ints, cast them as longs
      add a new helper macro by Tres Melton: IS_ALIGNED_128
      asm_loadimmq.S is included by other files, it isnt supposed to be compiled by itself
      only use GNU stack markings when generating ELF objects
      as pointed out by Quan, we need asm_loadimmq.S in EXTRA_DIST
      make sure people know the mmx support is 32bit only so it isnt for amd64
      touchup help output
      fix whitespace
      only declare do_mmx when it is needed
      move imlib_hash_size up in the code so we dont have to declare a prototype for internal usage

R.Ramkumar (2):
      Added documentation for tag id3-link-url Made the section on performance issues a bit clearer
      Removed some compiler warnings issued by gcc-4.0.1 on issues of signedness in comparison.

Ryan Little (1):
      make dist pkgs build again

Sebastian Dransfeld (13):
      Add asm_loadimmq.S to dist.
      sssh
      Silence
      Silence.
      If the version from config has something after x.y.z, drop it.
      * Add X headers if needed * Formatting
      Check for .dll extension on cygwin
      MIN and MAX is defined in common.h
      Add paranthesis to clean up.
      Remove unused variables.
      Remove unused variables. Remove signedness warning. Print pointers with %p
      Remove excessive strlen usage.
      Build fix.


v1.2.1 - 2006-03-18 (from dawn of time)
---------------------------------------
Azundris (7):
      various fu for changed evas-API
      nominal fix for memory leak in font.c (freetype1 font handling), just so we're in a defined state before switching to freetype2.  by azundris and atmos.
      * assorted fixes for RPM-building
      * spec-file (for RPM)
      * add token entry for LBM loader
      Grrr! : )
      * wonky versioning.

Carsten Haitzler (486):
      adding imlib2 code in.. NOT a lib yet... :) but playable code and loader system
      fixed minor bug in png loader.... added copying file :)
      updated loader api to include progress callback stuff.... :)
      adding the start of an actual aip layer... if you have any comments about this api - speak up now - because once it's final - that's it - thats the final api for imlib - anythig api.c calls etdc. can be changed - unless its the loader/saver api. that can't be changed either once its all final. :)
      new api bits :)
      ok - fix that to compile :)
      added Gary V. Vaughan's patches for libtool loader stuff and now its all automaked... :)
      add libtool libltdl form Gary...
      buugger me blummy :)
      all i have to say is.... OH YEAH! animated alpha blends on my root window... got a 640x480 image blending WIHT its alpha channel on my root window... drawing at... 20 frames per second... now if that dont make me happy.. i dont know what will :)
      flim! :)
      add some more stuff :)
      adding color modifier api backend stuff.... :)
      remember to not free images made form external data if it wasnt copied.. and free colors from color cubes once the context is invalid.. :)
      lots more work on mr imlib2 :)
      and more updates :) wheeeeeeeee
      more work on imlib2.. :)
      ooh is imlib2 ever workign fast now baybeee.. blending one image onto another .. with clipping, scaling, anti-aliasing and more.. need to add a bit to the api, and move the stuff nowin api.c off into imlib backend sinc ethat stuff doesnt belong in api.c
      jpeg loader added that does everything RIGHT - needto mapk the png loader do the same. :)
      and now the png loader does full progress callbacks and multi-phase loading correctly... WHEEEEEEEEEEe :)
      ooh now imlib2 has a sexy demo for you people :) mmmmm watch the alpha blending... mmmmmmmmm
      more playing with imlib2... :)
      add some more images just to show off :)
      get rid of printfs i dont need no more :)
      add operation type to blend ops.. :)
      oooh more blending and operation code :)
      better autogen.sh comments to help you build imlib2
      some mroe echos...
      added updates work.. well starting on it.. :)
      more wokr on updates
      lots of new image manipulation functions and minor fix in loader module code.
      add some files
      rewmove files i didnt mean tot add
      cleaned up code a bit... :) minor speedup for sparse (lost of transparent bits in images) for alpha blending :)
      get rid of extra space
      ok- fix depth retireval code :)
      try make png laoder work on big endian... :)
      and one mroe fix for big endian boxes for imlib png loader
      and add soem comments
      fix 15/16bpp depth problems
      let autoconf figure out our endianess
      make install isnot system loader dirs
      update autogen.sh
      fix main.c
      make imlib2 demo event based - test rect combining code in handling exposures and stuff - works it seesm - need to expand api though... expose handlign works fine as does rect mergeing and stuff.. must more efficient updating method now for demo.. template for stuff to be used by apps later :)
      add some files...
      more font stuff
      more code for font stugff being added.. more to come...
      why did i have a Makefile in cvs ?
      add some test truetype fonts - just for testing... and truetype font rendering code... :)
      bad bad font.......
      actualyl chekc if the font laod works and remove another bad font
      get rid of soem useless fonts...
      get rid of silyl fonts with silyl names... i hate those names... :)
      this font segfaults freetype.. ooh nice freetype :)
      nuke some more unusable fonts :)
      another useless font
      remove some more useless fonts
      some fixes to font code... :)
      fix raster map over-allocation problem for fonts.. :)
      add prototypes and cleanup unused vars
      you mightnt guess it - but rotated text all works now.. :)
      fix ups some toehr stuff...
      some fixes to get the output nextx and nexty right... :)
      add some of the font api to the api :)
      add speculative fotn cache ability - just like we have for images and pixmaps and ximages.
      add actual api.h calls to the font caching stuff...
      we have... anti-aliased line drawing code now... :) (and funnily enough - UNLIKE gimp it actually CAN draw a straight line for shit with anti-aliasing)
      color modifiers in imlib2 now done.. cleaned up soem code...
      again........ :)
      work work work...
      more flim
      more flim code..... >8)
      memcpy :)
      nice FAST gradient drawing code.... :) eat my dast... MUHAHAHAHAHHA! :)
      pixel query call.... need this one
      and the flim goes on..........
      dont chose visuals > 24bit :)
      LOTS of checkign in the api now to make sure the calling program can't stuff things up too badly...
      get rid of images i'm nto using...
      better api.h
      oooooooooops - thanks hans! :)
      dont need that fixme..
      speed testing code back.. just testing...
      try this..
      optmiseeeeeeeeeeeeeeeeeee. :)
      eeek math error at 255 (becomes 254) not surprising i didnt notice.. i looked at the results rsather than numerically evaluating...
      and handle ABGR ordering in 24/32bpp
      added ability to attach integert vlue and data poitner tags to images by string keys (with destructors optional) - wil be used for saving of images (savers will look for these keys to gleen parameters for saaving)
      fix some minro roundoff problems as before...
      udless &'s
      add TODO...
      structure for savign all done - now just need to fill in the save() functions in the loaders (yes laoder are also savers - loader and saver are interchangeable).
      we have a jpeg saver and the saver code works
      whee more robus tagging...
      and now it all works...
      flim
      and now thats all better.
      fix that........
      fix that bitchift..
      re-structure......
      restructure the direcotry a bit.......
      fix the version
      update README
      add ignores......
      more in ignore
      handle progress callback for saving in png loader..
      stop testing saving.. it works..
      no printf
      starting on pnm loader (ppm, pgm pbm, pam) - will finish later...
      oops makefile......
      hmm that didnt compile.. ooh fun :)
      pnm loader handles binary formats allright... :)
      binary png loaders done..plus speculating on the P8 format... dont like it much... i think ineed a FAST trivial to load ARGB format.
      pnm loader can save now...
      argb format loader & saver. my own format just so i can load and save raw ARGB data blindingly fast for imlib2 :)
      get rid of saver func
      oops - fix that filled rect drawing code
      add ignores.......
      fix a little of the rend code - never testyed that bit... andf the imlib2_view works nicely iwth zooming too :)
      primitive timeout.. its not even that good.. :)
      now that works better
      Makefile NOT Makefil ! :)
      more correct makefile.am in base........
      add soem stuff and new blend.c from ryan :)
      again..........
      mising 2 important calls inthe font code... :)
      ooooooooops :)
      added AUTHORS file.. fixed copyting....
      oopsie in blend.c
      oop s- clipping problme wiht lines.. fixed :)
      oh oops - image blending whilst scaling want quite right in the api.. :)
      ok - gradients now dont overflow the precision buffer as badly.. :)
      oops saver does rescan loader - so unless you laoded an image no laoders will be around... and it wont get rescanned on save.. :)
      lets break the Imlib2 api and chnage it... now its context based.. :)
      just up the versions to show i did something... :)
      spec file too...
      test program back to normal.. nwo works with api changes...
      fixe view to compile & work
      imlib2_view works again...
      oops :) fixed :)
      blum
      more blum - bloody freetype - why does debian have to go move the headre to a different location to where it always was?
      compile damnit...
      include config.h
      fix fix fix fix......................... :)
      jpeg loader stays quiet - png loader handles grayscale + alpha images correctly
      fix dat.........
      create .a's :)
      fix loaders......
      for acceleration to work i nee to add a parameter to put_back_data
      oops typo :)
      fix missing case in scaling for blending objects...
      get clipping right...
      add loader flush call and fix gif loader to be able to load when theres no progress set :)
      oops - expand indexed images...
      allow full paths for font names too..
      search path for font mroe sanely
      off by one in string alloc! bugger! :)
      no more dmalloc now :)
      dont be so anal abotu ewncodings... if no apple or windows encoding is there just use encoding charmap 0 :)
      i cant beleieve i missed wrappign the pixmap free function....
      add to header.....
      add dither mask pixmap rendering contexts...
      up version... add c++ usability..
      updates and fixes.. versioning etc...
      oops - forgort to remove param from imlib_free_color_rangex
      put that back...
      2 more checks in save calls for image data...
      that was silly! fix fix fix - thanks alan :)
      ooops - fix :)
      endinaness for masks broken onf sparc.. fix...
      oops typo
      get enmdianess roight for sparc (and ppc) for masks...
      oopsie - problem with non extension format images :)
      oooops - image and pixmap cache baddies.. :(
      const char *
      ooops - big eng9ian bug! :)
      rotattion code added... :)
      authors.. BTW - anyone watching commtis list please check AUTHORS... if your'e nto listed plese tell me to add you... I never do well maintaining it.
      add files...
      dont add that1
      flum..........
      flum
      better
      poatch main.c - but rottest doesnt work.. must fix later
      get context patch from tom......
      I'm back....... :)
      fix cmod.......
      optmize.. fix endianess stuff... :)
      oops - missed modfifying colros there.. :)
      rotate speedups - rend bugfix... wheeeeeee
      fix endianes problems..... works now on sparc solaris nicely.. :)
      ummmm fix dat.....
      BGR56r & BGR555 support.......... please test if u have a display like this :)
      nicer including of config.h
      oopsa typo
      damn willem! you love playing with imlib2 don;t you? :-) good show :)
      speedup scaling down....... but i cant seem to get any speedup for up scaling
      optimize scalign down routine for RGBA as well as RGB...
      um ooops - how did that happen?
      no more of that thanks
      need new updates call.....
      updates..... actually clip if only 1!
      faster scaling up.......... :)
      fix dither mask generation.. works again now.. use for icons to dnd
      dont need that code no more
      now that was bad! fix update appending :)
      add asm for blending.... this will break imlib2 right now for all platforms that arent xz86 intel 9unless you rmove the asm form the makefile and blend.c
      check for i686 artch and only then compile the mmx asm (i586 isnt guaranteed to have mmx - NB libs built for mmx will NOt work on non mmx boxes right now need to do a runtime chekc for that)
      dont compile mmx data struct in if no mmx asm is used
      fix spec file - dont buidl demos package
      optimize mmx blending more.. uswed to do 15 million pixels/sec... now does 25 million per sec.. compared to the C (9million per sec) thats pretty good now
      include updated comments
      fix some blending cases
      fix corner case for clippign where integer math rounds source widht to 0 where it shoudl be 1.
      blum blum blum
      full fix of logic in blending rgb->rgb functions in C NB: the mmx asm needs to be chnaged to reflect this
      changelog..... NOOOOOOOOO cant be! :)
      oops =- add
      foudn evil mmx code overwriting memeory! thanks mej. back to the C code for you!
      no - DONT put dmalloc in!
      get rid of printf
      FIX FIX! evil mmx code! missing decls! thanks dragan - mej ::)
      man.. more mmx asm for scaling.. thanks willem... you love this dont you? :)
      wow willem.. scalign down mmx code too.. :)
      unpatch scaling down code - there seems to be a segfault in it somewhere :)
      rgba code for plain 16bpp using mmx... :) and blend rounding asm error fixed thanks willem :)
      add the bugger :)
      mmx scaling back in - but forcibly disabled. new C scalign for scalign down.. works now.. :)
      disable damnit!
      scalign code back to old scaling... new scalign code has bugs... even the C code has segv's.. somewhere... :(
      sorry - needed to unpatch code for old scaling to work.........
      slight api changes..... problem was we have a useless paramin the pixmap gen calls - it shoudl have used the context... :)
      ok.. mmx asm for routines again.. and this time... they seem to not segv :)
      add
      better asm detection - there's an --enable-mmx now too if you want to force or disable the feature by force... it will try autodetect under linux but only on the build machine...
      nicer help message
      better configure check for freetype- hopefulyl people wont keep askign dumb questions anymroe about freetype.h
      cleaner......
      beter freetype_h stuff back
      handle infinite loop for tile if scalign down to 1x1)
      add -help patch and also fill in some options
      write text at any angle............... :) patches form willem again :)
      allow for flipping whilst scaling and rendering... :)
      aha! trying to free null pointers? NO NO NO - bad boy!
      asm for colormod ops......... :)
      build dither table for masks alwasy... even in depths > 16
      fixzed C code for ALPHA destination.. ok NOW its got it right... havent done asm code though... MAY need to optimize C code math for alpha dest.
      get rid of unused flsuh func, cleanup rects properly for case of 1 rect
      the RIGHT math for RGBA->RGBA ops... :) not optmized at all tho :)
      blend.c - RGBA destination works -and its optimized.. just a lookup..
      up to 0.0.5
      spec file up...
      use willems math... :)
      gawwwwwwwwwd - oops th = h not th = w; :)
      add tga loader - thanks dan :)
      OS2 pacthes make imlib2 build on os/2 apparently :)
      oops - forgot to commit that :)
      got rid of X calls in loader - no XParseColor
      oosp parse better :)
      add headre checking.....
      fix tga loader....... :)
      bowis's filter stuff... :)
      turn off bump mappign for now..
      disable others.......
      loader that loads and saves images from a dbm database.... :)
      shoudl in theory handle locks better....
      IT WORKS! :)
      oooooooh look at that.. it now supports compressed image data in the db
      fix endianess problem with loader
      patch to fix loader to handle non line-feed header pnm's :) and ascii too.
      apparently clone doesnt lone EVERYTHING.. now it clones all of it except attached data tags...
      fix mem leak in lisitng fonts
      add 1 more font routine for getting geometry - useful. you'll need to update imlib2 too to get evas to compile & work - it uses this routine
      fix static gc for multipel servers
      imlib2-config added
      get rid of that replacement..
      try that
      api call was silly - changed it :)
      noticed there was a set filter but no get.. addded
      possible crash fixed
      foudn bug in mmx asm blending.. 1 line hihg blends get skipped.. fix! :) (ugly fix tho)
      get rid of comment
      db loader/saver needs edb now - much better! :)
      no debugging printf
      add docs to cvs
      dont need -ldb anymore
      wooo! found bug in filename:key splitting.... fix fix fix... :)
      fix big endian code :)
      simple commented demo
      doc looking much better
      added generic slow-path rendering code
      and enable the fast path again
      add willems docs to the docs ........... :)
      only build loaders if headres/libs are found - chheck for libs and note them minimum requirements are jpeg, png and db loaders.
      up version number.. tentative for a 1.0 release... revamp rpm packages completely. split loaders into their own packages (more logical units). main imlib2 requires the jpeg, png and db laoders at a  minimum
      fix requires to be more accurate
      oops - fix that
      docs get built......... :)
      add makefile for docs
      and make them build.......... PROPERLY!
      we can add them back in... conditions put elzewhere
      ewwwwwwwwwwwwps! scaling blending buggy! fix fix fix fix :) GOTCHA!
      dont NEED those dependancies
      missing some load data checks for some routines... fix fix fix
      ok- lyly can have his changelog back.. ugly - eats space with nothing useful :) - thats what cvs is for... :)
      add api call to get text string advances........
      add docs..... and prefix is /usr for rpms damnit! :)
      add call to get text inset for string
      fix visual picker :)
      666 colorcube rendering works....... :)
      handle lower depths :)
      no prointf!
      unrolled span rendering a little more (switch stamement is now just once per span 0 thats good enough.... :) )
      fix that! that aint static! (__imlib_dynamic_filters_init that is)
      fix pixmap caching...... :)
      free image? dirty the pixmaps that belong to it and set image pointer to null
      fix masks over network
      work under bigendian again
      paparnoid lseeks & fseeks in tiff loader due to bsd bugginess :)
      add dat
      oops fix that
      extar -> extra
      cflags -> cppflags
      polygons now........ don't anti-alias anymore.. will be fixed.. BUT
      clip rect fix
      no more ellipse segv's :)
      fix a small segv problem with pollies! :)
      --without-x patch from steve
      filter docs :)
      off by one may have been causing segv's ? :)
      pnm loader more paranoid about pnm format checks
      patch for grabbing form ximage's - plug dont segv when u set the format to NULL.
      more paranoid abotu allocating memory with realloc
      bigendianess patch from nathan
      os/2 fixes :)
      masa's internationalization + x font support for imlib2 :)
      oops - fix soem ascent & descent problems
      err commit?
      errr - clean clean.. fix color pixle caclfor 8bpp and 8bpp non dithered had a.. er... bug :)
      attempt to fix xfd font transparency....... ??? :)
      x fonts blend again now.... and colro correctly too :)
      * up version to 1.0.1 - will do minor release real soon. * fix build so loaders build if u have no imlib2 installed * fix requires and buildrequires in spec file for freetype & edb to be correct
      fix minor segv in gradient rendering if your color range has err... no colors
      patch from matt
      minor fixup if last char in string has 0 boundingbox width :)
      use 128x128 dither mask for rgb666 (in 8bpp) rendering
      how the HELL did that 8 get there and things still compile?
      possible infinite loop in cache code.. fix.
      on a stick!
      errrr?
      ewwwww - fix infinte loop bug... :)
      fix blend mode for alpha dest when drawing text
      oops - hard coded op - fixed that :)
      and then he found some minor bugs in file examination... and then they were fixed. :)
      found it!!!! :)
      adam's patches... :)
      add adam to AUTHORS
      sorry - debain dir breaks build. removed form configure and makefile... also up to 1.0.2
      errr oops - aleak.. fix fix fix :)
      oops and a leak in the png loader.. and fix a potential leak in the jpeg saver for when things go awry.
      carsten's context stack patch... with a default context entry too.. :)
      oops - widht & height 1 grad get div by 0 .. fix fix fix :)
      oooooooooooooooops fix fix fix bug bug :)
      up to 1.0.3
      imconvert.c added to tarball
      franz's patches to support other color spaces :)
      add then :)
      add :)
      in cvs :) clear function :)
      add a color clearer too
      bmtext dithers.... filters work.. :)
      add mark's patch
      and add author...
      beat me silly. i forgot to allocate the memory... :) yay.. fixed :)
      virtualize real file and key splitting nicely into image data struct. now we escpae literal colons with double colons. it's documented too. (this was easier to do as i also have to do it for ssving files and you cant stat to see if a file exisit fi you havent saved it yet)
      oops - chekc for keys and real files first
      alright - comprimise. theres a imlib_context_set_filename_raw_mode() call now - if you want to deal with filenames and not have them interpreted use this and set it to 1.
      double up modified date checks... incase of colon
      oops developer debug info wrong.. fix
      oops.. free pointers that might be null.. checdk for that first!
      ooooh that would leak if we added the same path all the time... which shoudl be ignored... :)
      db loader out of imlib2...
      aha! oops :)
      hmm - fix segv with x font support
      hmmm now why did this break? hmmmm.... err.,.. hmmm
      oooops... fix fix :)
      working offline... :)
      blum! :)
      break out if no footer
      can handle "comment" tag :)
      ok ok - pass make disctcheck
      todo list.. and even how to do it! :)
      tiff compression patch :)
      better loader... ignore trailing garbage xv puts on the end of an xpm...
      vrsion -> 1.1.0
      make packages
      update
      dont need x11 includes
      -lm
      build without x.........
      handle error stuff right
      tillman's png interlacing support patch! yum!
      bugsie! fixed!
      no savies! :) <-tilman
      tga loader fixed. now it works! :)
      Michel Briand <michelbriand@free.fr> mmap tga loader
      kwo imlib2 patch
      actually use cache in 32bpp/24bpp
      kwo's patch
      1.1.1
      just .in now - the whole transition is over.
      fix distcheck
      distipoos
      oops filters wrong spot
      imlib2_loaders...
      roatate from buffer patch
      kim patchies.. and in authors!@ :)
      patches... :)
      ltdl be gone fromt he src tree! :)
      dont return values if void return defined!
      nuke libltdl subdir
      pc fix
      --without-x works again
      handle files too small to be a valid tga!
      try using advance metrics
      jose's AA rendering patches are in! :)
      mr gonzales's latest "final" code.
      autofool cleanups... argh! i hate autofools!
      build things THIS way... :)
      sorry - mej - you seem to have broken the build on other systems and imlib2-config wasnt being installed - also it was deciding it had to cross-compile and build i686-gnu-linux-imlib2-config etc. files... had to move autogen.sh to this... :(
      revert... :( breakies
      more bmp fixes
      pallet + transp fix
      up to 1.1.2 - security fixes, some other fixed, ilbm loader
      fix brian
      bits per smaple! :)
      buildie cleanies
      remooov!
      and new much cleaner tree.
      get the file list right.
      ldflags by bye
      fix install
      no asflags recurse
      oosp accidentally put this in. damn!
      openembedded build files... this makes life so easy to build efl for embedded... :)
      cleaner configure.in
      oopsies. fixies
      bart patch for debian package stuff
      we dont need no steenking x headers in the loaders
      no x headers.
      ditch x headers from those loaders
      progname... :)
      no more segv/buffer overflow
      bmp patch
      make hsv reversible
      FILL IN @REQUIREMENTS@ IN PC.IN
      url...
      dont double guess unicode.. just let it be raw
      link modules back to imlib2 in case they are used in a python extension that dlopens imlib2.so....
      change versions. some went DOWN - because they arent releases and i'm trying to remove the _pre ascii from the version. i added a .001 (a release number) so we can automatically or easiyl generate releases... sorry guys. but it's kind of "for the good of the code". :)
      lround -> round
      DISPLAY_MISSING define from spanky
      imlib2 configure.in patch
      1. e17's init icons get put into an e_box for arrangement nad the init splash determines the location...
      if there is NO x dir.... still link anyway
      remove files so dirs go away
      auto-package imlib2...
      fix more
      touchies!
      add data filea and test progs to -devel
      asparagus time for some of the core efl bits...
      bz2 loader fix. thanks julia!
      no dot!
      bettter detect for mmx (same as evas now)
      asparagus!
      amd64 asm patches
      John Slaten's amd64 mmx patch
      pass make distcheck
      make e17 pass make distcheck.... and... ASPARAGUS time...   http://enlightenment.freedesktop.org for tarballs
      asparagus!
      full asparagus
      xpm segv fix
      asparagus... make e17 distcheck.
      aspara!

Chris Ross (20):
      * Changed some of the methods to stop furutre name conflicts * Added Willems patch for bump mapping -very vey cool, check out   test/imllib2 * Died due to excess excitement over bump mapping
      * Stuff from term, fixes some rpm build issues with imlib_view
      One fricken character. Freetype 1.2, not 1.1
      * modifed the script engine, instead of three passes i've nobbled it down to   one, this means the bump_mapped pr0n will now render a coupla degree's faster  (gilbertt this is for you, and those pictures of pabs' mom) * Update Imlib2.h and api.c to reflect changes
      Gah, turn off debug mode, and comment out necessary blurb...
      Added Willem patch for the bump map filter, now does proper bump mapping from an infinet light source. Needs to be optimsed further - lookup tables or some such..... thats for another day.
      * rewrite of the script parser, basically you can now parser a filter as a   variable to another filter as willem requested the other day.   eg.     filter( var=anotherfilter( var=13,var=30 ), var=blum );
      * forgot to mention that i've added an option to test/imlib2    -bmp2pt add this too it's command line and it'll bump map to where the            cursor is.
      Clean up of code, all macros for filter_param -> real varaible are put in script.h. Changed filters to reflect this change, and actually plan on writing some more macors and filters soon.
      Ok, some more clean ups to the filter stuff, should have some new filters to play with soon - want to get the stuff correct before I commit some more stuff. dox is the start of dox2 the document viewer based on imlib2. Designed so that the style of the docs is seperate from the contents. Will evolve rapdily over the next week.
      More changes. Still doesn't do anything.
      blu7m.
      blum. fixed "error" on first install type bug in the dox tree.
      blum.
      Start of a filter test app, and applied Willem's patch for imlib2. thanks Willem.
      New pic. Got bored with the last one =).
      Moving over to the new home in the efm module.
      Be quiet.
      Argh.
      You know what I do to fools? I pity them :)

Christian Kreibich (11):
      Raster,
      4:31 am. Oh my.
      An XCF loader. Currently it can handle layers, layer offsets, layer opacity, layer masks, and merging layers in the default mode (simply "looking" through all the layers). The other layer modes are missing right now (I hardly ever use anything other than "Normal" anyway, but that's just me of course).
      I don't know if the loader has endianness issues (I guess it does), but this should definitely work better.
      Ahem. Of course the load fails when you can't open the file :o)
      Bye bye XCF loader. Apparently I stepped on some Gimp people's feet with it, because of licensing issues. I guess I'll be talking to Raster next week what we'll do with the loader. Hope this makes us friends with the Gimp developers again.
      And another autoconf update ...
      Lots of sssshhh here ...
      Well then let's delete the full thing?
      Same changes for HEAD ...
      Okay, don't use AM_PROG_AS but the workaround, to fix automake issues.

Christopher Rosendahl (2):
      bad bad!
      bad bad =)

Corey Donohoe (1):
      segfault fix from Dave Weston <dtweston@student.math.uwaterloo.ca>

Dan Sinclair (1):
      - documentation fix: the angle is in radians not degrees

Franz Marini (6):
      added nick. nothing important. :)
      ok, just wrote this little function to do pixel drawing with blending. In fact, I wrote it just for the Bezier drawing function, but I thought it could be useful in other cases too. Have fun, Lightman :)
      Alright, changed index.html to document imlib_image_draw_pixel.
      Ok, just tried to compile Imlib2 under Roswell (RH 7.2 beta) and I discovered that it installs freetype 2.0.3 , and so freetype.h (for ftype1) is under /usr/include/freetype1/freetype/freetype.h . As to not have plp complaining imlib2 can't find freetype under rh 7.2 , I patched configure to work with roswell. Lightman
      Ok, reverting changes for the draw_pixel function. Now Imlib_image_draw_pixel uses Tom's macro (__imlib_draw_set_point and *_clipped) so it's faster and it handles clipping, too. Btw Tom, I choosed not to change the name of the wrap function so that : 1) I don't have to change it in doc/index.html ;) 2) I think it's a little more in line with Imlib_image_draw_line ...
      Ok, SirDibos modified the html so as to be more readable, namely, fixed <pre> tags, removed &nbsp; ... just some cleanup ...

Horms (4):
      rpm -ta now works on a tarball produced by "make distcheck" and friends
      Debian packages may now be built from output of make dist and friends
      Debian packages may now be built from output of make dist and friends
      flum

Ibukun Olumuyiwa (2):
      Patch for pkg-config support from Tilman Sauerbeck <tilman@code-monkey.de>.
      Warnings suck

John Bickers (2):
      Fixed start-of-line HAM problem. Added SHAM and CTBL load. Added greyscale load. Added IMLIB2_LBM_NOMASK check to disable masking. Added colour gun scaling, e.g. 4-bit 0x0f scales to 8-bit 0xff, not 0xf0. Changed RLE decompression by scanline instead of by byte. Removed empty save() function altogether as per some other loaders.
      Added entries for loader_lbm.c.

Kevin Brosius (2):
      Minor README update - edb is used by imlib2_loaders.
      .spec version update

Kim Woelders (3):
      Fix grabbing when source x or y is < 0.
      Fix imlib_create_scaled_image_from_drawable().
      Fix imlib_create_scaled_image_from_drawable for source_y != 0.

Laurence J. Lane (15):
      more preliminary assimilation
      clean target
      stuff
      removed dh_testversion
      stuff
      added patch by David N. Welton to cleanup configure handling Thanks, David.
      synch with current woody packages
      update rules for cvs builds (autogen.sh)
      minor update
      cleanup source names
      clean up for build
      various stuff
      Disable MMX routines. They were already (well, allegedly most of the time) disabled in the official packages for obvious reasons. I'd rather leave them enabled for CVS, but binutils headaches abound.
      another missing dependecy imlib2 suggests imlib2-loaders e17 depends on imlib2-loaders
      debhelper 3.0 changeover

Maher Awamy (1):
      Made png.so and jpeg.so link against Imlib2 when building, this makes the perl bindings problem with undefined symbols for __imlib_GetTag disappear since those two loaders call that function to determine some image flags when saving. Kick me in the butt if I am not supposed to do that but KainX said I should. The differance in .so size is minimal, 30 bytes for png.so and 20 bytes for jpeg.so.

Mandrake (11):
      Fri Oct 22 10:53:26 PDT 1999 (Mandrake)
      trying to "fix" imlib2's cvs tree
      more "hush yo mouf cvs" changes
      Sun Oct 31 20:21:13 PST 1999 (Mandrake)
      removed a warning
      fixing freetype detection stuff, maybe?
      hmm
      heh.  not actually using configure option for mmx disabling correctly
      this doesn't work at all. libtool: link: `-L../src' cannot specify a relative directory
      Er, this was just blatantly and obviously wrong. fixed.
      hush cvs

Mark Bainter (1):
      Changed --with-freetype to --with-ttf to make it consistant with the other modules.

Michael Jennings (67):
      *sigh*
      Hush.
      A GIF loader.  There is no save function yet, and you'll need libgif to use it.
      Whoops.  Forgot to call the progress callback one last time.
      Don't ask me how this got out of sync....
      *grumble*
      Ummm...
      Put those back.  I hate warnings.
      TIFF loader from Eric Dorland <dorland@lords.com>.
      BMP loader from Isaac Richards <ijr@po.cwru.edu>.  It currently has issues with progressive loading, so don't use it with feh. :-)
      Keep raster happy.
      BMP loader fix for progressive loading from Chutt.
      Murple.
      Added a function to retrieve the image filename, if it has one.  It returns a pointer to an internal string, so if you want to alter the filename, you MUST strdup() it.
      Several miscellaneous bugfixes I did while converting Eterm to use Imlib2.
      "Hi.  My name is raster, and I smoke crack.  I think I'll dereference this pointer I just freed.  Sound like a good idea?  Yes, I thought so too." :-P
      Fixed a possible divide by 0.
      Never mind.  I fixed the bugger.
      Wed Apr 26 19:58:05 PDT 2000 (KainX)
      Don't cache partially-loaded images.
      Hush up on the warning.
      Nuked some autogenerated files and added acconfig.h.
      Hush CVS.
      It's always a good idea to "make distcheck" before you commit when you add or remove files. :-)
      Some silly goose decided that these files should #include Imlib2.h.  NONE of them should include Imlib2.h.  In fact, nothing in the Imlib2 code should, but if it's absolutely necessary, make sure the local one is found before the system-wide one.
      Wed Jul 12 22:20:53 PDT 2000 (KainX)
      Miscellaneous fixes I ran across while doing the colormod stuff.  One of which fixes a seg fault bug.
      Son of a raster!  I think I just found the memory leak.  I shall now hide in shame.  Perhaps giblet should take over Eterm development.
      Imlib2 now benefits from the same MMX goop that Eterm has. :)
      Fix make distcheck.
      So fix it, don't just get rid of it. =P
      Fix error building with dmalloc support.
      Fri Aug 10 13:33:13 PDT 2001 (KainX)
      Mon Oct  8 10:00:19 2001                        Michael Jennings (mej)
      Spec file fixes here too.
      Minor portability nit to appease the Texan.
      Oops, forgot to fix configure.ac.
      Tue Jan 15 15:22:06 EST 2002 (KainX)
      Thu Mar 14 19:18:07 2002                        Michael Jennings (mej)
      Mon Apr  8 17:47:55 2002                        Michael Jennings (mej)
      Wed May 29 09:22:42 2002                        Michael Jennings (mej)
      Wed May 29 11:58:32 2002                        Michael Jennings (mej)
      Tue Jun  4 23:00:30 2002                        Michael Jennings (mej)
      Tue Jun  4 23:29:36 2002                        Michael Jennings (mej)
      Mon Mar 31 15:20:43 EST 2003 (KainX)
      Thu Apr  3 14:06:53 EST 2003 (KainX)
      Thu Apr  3 20:48:27 EST 2003 (KainX)
      Sat Jul 12 21:06:14 EDT 2003 (KainX)
      Gah!  New files. :P
      Ignore imlib2.pc since it's auto-generated.
      Sat Jul 12 21:33:20 EDT 2003 (KainX)
      Package names can vary.  Besides, if one doesn't know what's needed to build it, one shouldn't be building it. :-)
      Oops.  Forgot to nuke that.
      Fri Jul  2 14:41:17 2004                        Michael Jennings (mej)
      Tue Jul 20 17:23:57 2004                        Michael Jennings (mej)
      Sun Jul 25 17:45:53 2004                        Michael Jennings (mej)
      Wed Aug 25 15:53:59 2004                        Michael Jennings (mej)
      Thu Aug 26 13:25:22 2004                        Michael Jennings (mej)
      Ditto.
      Copyright -> License
      Ssssh!
      Thu Jan  6 10:27:24 2005                        Michael Jennings (mej)
      Sun Jan 16 14:27:18 2005                        Michael Jennings (mej)
      Wed Jan 19 17:10:29 2005                        Michael Jennings (mej)
      Fri Jan 21 00:57:08 2005                        Michael Jennings (mej)
      Sun Jan 23 22:30:28 2005                        Michael Jennings (mej)
      Fri Jan 28 20:26:06 2005                        Michael Jennings (mej)

Michael Thalmann (1):
      corrected order in autogen.sh, updated to automake 1.5

Mike Frysinger (3):
      allow for setting of env var to prevent running ./configure like the older autogen scripts allowed
      remove AC_CANONICAL_TARGET and use $host* variables instead of $target* variables since thats how it works
      we want to search for libX11 not libX ( http://bugs.gentoo.org/93300 )

Nathan Ingersoll (4):
      Explicitly link the modules to Imlib2 for portability to other platforms. (OS X in this case)
      Got a little over-zealous on the linking of Imlib2, removed the cases that aren't necessary.
      Remove the attempt to link freetype1, use the linking information generated by the configure.in.
      glibtoolize

Nigel Kostiuck (2):
      Added the change for BitBake the Standardized Openembedded Build System
      Python is too incompetent to parse hyphens

Peter Kjellerstedt (2):
      Removed this generated file from CVS (again).
      Corrected a typo.

Platon Fomichev (4):
      OS/2 fixes
      OS/2 binary open fix, I think in UNIX we can safely use "O_BINARY" too
      General cleanup of EMX things
      First preview of gzip & bzip2 loaders

Richard Lowe (1):
      Please Miss, I need -lm too.

Sytse Wielinga (14):
      Updated imlib2 debian packaging. It still doesn't configure correctly for me without some changes to autogen.sh and configure.ac; I'll have a look how to fix this correctly.
      - Fixed building with automake 1.6.3, by adding the 'CCASFLAGS' and 'CCAS'    substitutions to configure.ac  - Fixed building with autoconf 2.5, by making autogen.sh run aclocal and    autoconf also in libltdl. I'm not sure if it is all-right in all cases;    please have a look at it.  - Added some things to .cvsignore.
      1. autoheader and automake should also be run in libltdl 2. forgot autom4te.cache in .cvsignore
      Removed libltdl directory from imlib2; it's created by libtoolize.
      Readded libltdl/acconfig.h. That should be there.
      Raster isn't an 'author(s)'
      Big overhaul of a lot of the debian packages
      Updated most of the debian packaging. Everything I have missed is out of date, not important or not working.
      Fixed a couple of debs in e17/libs, added and updated a couple of .cvsignores.
      Updated the imlib2 debian packaging.
      Slight update for the debian packaging of imlib2, imlib2_loaders, ecore, edje; more to come later.
      Things may change sometimes. Let debian cope with it.
      A lot of moving around in imlib2{,loaders} caused lots of unignored files
      Made imlib2's tests and demos open the display correctly.

Term (10):
      First commit. Woohoo!
      More typos (but in raster's defense, he's spelling checking every so often. ;)
      Cleaned up/partially rewrote README. Basically updated for the 1.0 release.
      Bored. Reading. Typo. Commit. Sleep.
      Add --with-edb=DIR argument (by request).
      Cleanup the spec file a bunch with a patch from Joakim Bodin <bodin@dreamhosted.com>. Changed a few things in to make it happy to build in rpm < 3.0.5, and a few other minor tweaks.
      Based on a suggestion from Richard Lowe <richlowe@btinternet.com>, and partially from a patch from him, updated README to include the dependancies. Also added a few more to imlib2.spec.in (libjpeg, libpng, edb).
      Blah. Isn't it great how things like "check the other packages in the spec file" occur to you RIGHT after you do a commit? I love being up late. :)
      Added imconvert, which stems from a conversation with raster about importing/exporting binary data (specifically imlib images from edb files, like the ebits files). This allows the user to export and import image data in and out of edb files, as well as arbitrary conversions to and from any format Imlib2 can handle.
      I should've done this a while back. Sorry about the big number of emails coming. ;)

Till Adam (3):
      small fix for the pnm loader. It didnt display the last two lines of a file. Ive only corrected the 24bit RGB one. Ill do the others if raster doesnt have a more elegant fix for this :).
      fix for the other binary versions in the pnm_loader. Could someone please check if the ascii ones work right? I didnt test them. and test the binary formats too while youre at it. Thanks :)
      ok, fix for the ascii ppm formats. They seemed to have the same problem. This should be it now, provided the save in ppm format works as expected which it seems to do from geist at least.

Tilman Sauerbeck (17):
      Fixed CFLAGS in pkg-config file
      im->real_file() is set on load, but it will be overwritten when the file is saved. so free the old contents before overwriting it.
      Imlib2 loaders don't need to ship with both load() _and_ save() anymore
      replaced loader_gzbz2 with loader_zlib and loader_bz2. removed empty save() in some loaders
      fixed some warnings
      Don't call DirtyPixmapsForImage() twice - it's already called by DirtyImage()
      DirtyPixmapsForImage() is only available when we build with X
      fixed a bunch of warnings. reverted last commit.
      don't attempt to save an image without image data
      updated Doxyfiles
      always print an error message if an image cannot be opened
      another bug in the bz2 loader: we need to duplicate the original filename so we don't access free'd memory
      fixed a fd leak and a bad memory access bug
      zlib/bz2 loaders, round 2: look for the real loader using a fake filename (original filename with the bz2/gz suffix cut)
      fix uncompression in bzip2 loader2
      fix uncompression for the zlib loader
      handle bzip2 errors

Tom Gilbert (85):
      Shaddup ;)
      shaddup ;)
      AARGH. Godamn file decriptor leak which has been driving me CRAZY for a WEEK! Got the BASTARD. DIE!
      Removed a crufty bit.
      nothing major
      Leak plugged. Thanks Eric :)
      Okay. The loader list is now trimmed. Where it would previously contain: argb.a   bmp.a   gif.a   jpeg.a   png.a   pnm.a   tiff.a argb.la  bmp.la  gif.la  jpeg.la  png.la  pnm.la  tiff.la argb.so  bmp.so  gif.so  jpeg.so  png.so  pnm.so  tiff.so
      Bite me =P
      Partial loader_tiff rewrite from Eric Dorland. Much nicer :)
      Thought you could use a ChangeLog. I filtered it from muy cvs-commits-list mbox, so bin it if you don't like =)
      Thu Apr 27 02:59:57 GMT 2000 (gilbertt)
      Consolidate one ChangeLog in the root dir.
      Thu Apr 27 03:16:59 GMT 2000 (gilbertt)
      Thu Apr 27 04:00:28 GMT 2000 (gilbertt)
      Thu Apr 27 04:22:06 GMT 2000 (gilbertt)
      Thu Apr 27 13:41:11 GMT 2000 (gilbertt)
      Thu Apr 27 13:43:49 GMT 2000 (gilbertt)
      This is weird. I'm sure I added the prototypes for these context_get_* functions to Imlib2.h before... Yet they aren't there.... Hrm...
      SHUT YOUR HOLE CVS BITCH!
      Default x,y to 0 and w, h to image dimensions, so if you don't specify, the filter applies to the whole image.
      Willem's rotation patch.
      fix core on imlib_list_fonts()
      Fixed imlib_list_fonts()
      Don't show duplicates in imlib_list_fonts().
      Okay. imlib_free_image_and_decache() was leaking images. I hope I made the right fix here. Basically, the imlib_free_image_and_decache() call in api.c sets the flag F_INVALID then calls the internal __ImlibFreeImage(), this checks if the flag F_UNCACHEABLE is set, and only frees it if so.
      Urm. oops. hehe :)
      Added line drawing with clipping.
      Added rectangle clipping, in the form:
      The line clipping function is quite useful, so I made it public. Sometimes it's handy to work out where your line was/would be drawn.
      Scratch that. Start again.
      Polygons. Not filled ones yet =P Empty ones are easier ;-)
      Added function to calculate bounds of a polygon.
      imlib_image_draw_ellipse()
      Polygon filling. Right now only works for convex polygons. Works with a clipping rect, but highly suboptimally (I'm not doing proper polygon clipping here yet, just clipping slowly on each point drawn - really nasty).
      Better API, simpler polygon struct. imlib_polygon_new() now, no type member. Then _draw_polygon(polygon, unsigned int closed), and _fill_polygon(poly).
      imlib_image_fill_ellipse()
      killed a rounding error in line clipping
      slight speedup
      tidy up
      Much faster polygon clipping, made span() more sensible.
      Much faster clipped ellipse filling.
      final fix for _list_fonts()
      more docs
      update my email addy
      Lots of changes. Macro-ised the point_on_segment code.
      shuddup
      Use spans and the span list clipper to do filled ellipses too. Much better.
      more inlining
      Commiting what I have so far. More to do tomorrow, but it's 3am.
      nm I fixed it anyway
      quick warning hunt
      not quite there yet
      Fix for _polygon_get_bounds.
      Better fix.
      EEeeeeek. Segv.
      SHUT UP!
      png loader now pays head to images "quality" tag, just like the jpeg loader. Now, the png lib takes values 1-9 for compression. I decided to standardise loaders on a 1-100 quality value, and do some sums in the loader to convert to 1-9 compression. That was you can set quality and not care what file format is used. Sound reasonable?
      jpeg and png should do the right thing with quality _or_ compression now
      let's not have a coredumping example app ;-)
      dunno why there were two of those
      Adam's font fixer-upper patch :)
      fix memory leak
      warnings suck
      This is what I meant.
      safer
      er no
      sorry, debugging stuff
      ewps!
      that doesn't actually do anything different
      another one
      Expunged all the raw_file stuff and fixed a warning.
      fixed filled ellipses - note, this isn't the cleanest fix in the world.
      bad logic there tom
      Once more into the breech.
      You are required to set `AS' and `ASFLAGS' via `configure.in'.  The autoconf macro `AM_PROG_AS' will do this for you.  Unless they are already set, it simply sets `AS' to the C compiler and `ASFLAGS' to the C compiler flags.
      No, I didn't mean to get rid of that bit.
      *snicker
      Patch from Brian Lindholm <lindholm@aol.com>
      foo
      Fix broken ordering.
      Fix from Lindholm@aol.com for segvs when tiling images seamlessly (if they are an odd number of pixels in height).
      A bugfix!
      Fix bug in ellipse drawing introduced who knows when by who knows who.
      memory leak busted - valgrind is great.
      ->data was getting leaked, as _tidyup() only free()s data for nodes of type CHAR.

Vincent Torri (5):
      Doxygen documentation
      Doxygen scripts, description of imlib2
      Doxygen doc: css and html files
      Doxygen doc: images
      Configuration summary