kwo pushed a commit to branch master. http://git.enlightenment.org/legacy/imlib2.git/commit/?id=67011d6252c2cd2c3a4df9a3a0c5759c65d9b634
commit 67011d6252c2cd2c3a4df9a3a0c5759c65d9b634 Author: Kim Woelders <[email protected]> Date: Sat Apr 15 15:59:22 2017 +0200 1.4.10. --- ChangeLog | 358 +++++++++++++++++++++++++++++++---------------------------- configure.ac | 4 +- 2 files changed, 188 insertions(+), 174 deletions(-) diff --git a/ChangeLog b/ChangeLog index ccc2107..01fdd32 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,5 +1,19 @@ *** +*** Version 1.4.10 *** +*** + +Kim Woelders (2): + PNM loader: Fix reading PNM bitmaps. + Fix missing break. + +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 + + +*** *** Version 1.4.9 *** *** @@ -235,7 +249,7 @@ Kim Woelders (29): 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). + 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). @@ -275,7 +289,7 @@ Carsten Haitzler (1): 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). + 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+-. @@ -304,21 +318,21 @@ 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 + 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. + 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. + remove duplicate line. clean install files for test package. fix test package and clean up some install files. generalize imlib2 library install files @@ -328,12 +342,12 @@ Kim Woelders (19): 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. + 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?). + 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). @@ -357,7 +371,7 @@ Sebastian Dransfeld (2): 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 + * improve autotools stuff * move libtool versioning from src/bin/Makefile.am to configure.in * formatting *** @@ -374,7 +388,7 @@ Carsten Haitzler (6): 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. + 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): @@ -387,7 +401,7 @@ Mike Frysinger (4): 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. + 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. @@ -409,7 +423,7 @@ Carsten Haitzler (9): 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. :) + 1. autofoot patches. 2. fix maximize to work again. :) fix blah-config includes remove openembedde pkg info - old and dead fix x detect @@ -422,21 +436,21 @@ 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: + 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. + 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. + 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. + 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. @@ -455,12 +469,12 @@ Ben Rockwood (1): 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 [...] + 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) ... + 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: ---------------------------------------------------------------------- ��������� :) + 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 @@ -470,7 +484,7 @@ Carsten Haitzler (31): 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. :) + 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 @@ -493,7 +507,7 @@ 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). + 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. @@ -528,8 +542,8 @@ Mike Frysinger (23): 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. + 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 @@ -540,12 +554,12 @@ Sebastian Dransfeld (13): Silence Silence. If the version from config has something after x.y.z, drop it. - * Add X headers if needed * Formatting + * 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 unused variables. Remove signedness warning. Print pointers with %p Remove excessive strlen usage. Build fix. @@ -556,7 +570,7 @@ Sebastian Dransfeld (13): 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. + 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 @@ -564,27 +578,27 @@ Azundris (7): * wonky versioning. Carsten Haitzler (486): - adding imlib2 code in.. NOT a lib yet... :) but playable code and loader system + 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. :) + 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... :) + 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 :) + 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.. :) + 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 + 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 :) @@ -597,7 +611,7 @@ Carsten Haitzler (486): 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 :) + 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... :) @@ -608,12 +622,12 @@ Carsten Haitzler (486): 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 :) + 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... :) + 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... @@ -629,9 +643,9 @@ Carsten Haitzler (486): 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 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) + 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... @@ -642,7 +656,7 @@ Carsten Haitzler (486): 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... + 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! :) @@ -650,13 +664,13 @@ Carsten Haitzler (486): 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... + 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) + 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). + 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... @@ -677,13 +691,13 @@ Carsten Haitzler (486): 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. + 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 :) + 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 :) + 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 ! :) @@ -697,7 +711,7 @@ Carsten Haitzler (486): 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.. :) + 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... @@ -706,7 +720,7 @@ Carsten Haitzler (486): 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? + 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......................... :) @@ -718,13 +732,13 @@ Carsten Haitzler (486): 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 :) + 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 :) + 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... @@ -743,7 +757,7 @@ Carsten Haitzler (486): 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. + 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.......... @@ -772,37 +786,37 @@ Carsten Haitzler (486): 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) + 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 + 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. + 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 + 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! + 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 :) + 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.. :) + 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... :( + 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... :) + 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... + 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 + 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) @@ -812,7 +826,7 @@ Carsten Haitzler (486): 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. + 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.. @@ -836,9 +850,9 @@ Carsten Haitzler (486): 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... + 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 + 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.. @@ -846,7 +860,7 @@ Carsten Haitzler (486): 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) + 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 @@ -859,8 +873,8 @@ Carsten Haitzler (486): 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 + 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......... :) @@ -870,7 +884,7 @@ Carsten Haitzler (486): 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... :) + 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 @@ -878,7 +892,7 @@ Carsten Haitzler (486): 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.... :) ) + 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 @@ -897,17 +911,17 @@ Carsten Haitzler (486): 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. + 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 :) + 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 + * 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 :) @@ -919,16 +933,16 @@ Carsten Haitzler (486): 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. :) + 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 + 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. + 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 :) + oooooooooooooooops fix fix fix bug bug :) up to 1.0.3 imconvert.c added to tarball franz's patches to support other color spaces :) @@ -936,17 +950,17 @@ Carsten Haitzler (486): add :) in cvs :) clear function :) add a color clearer too - bmtext dithers.... filters work.. :) + 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) + 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. + 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... :) + 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 @@ -971,7 +985,7 @@ Carsten Haitzler (486): bugsie! fixed! no savies! :) <-tilman tga loader fixed. now it works! :) - Michel Briand <[email protected]> mmap tga loader + Michel Briand <[email protected]> mmap tga loader kwo imlib2 patch actually use cache in 32bpp/24bpp kwo's patch @@ -995,7 +1009,7 @@ Carsten Haitzler (486): 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... :( + 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 @@ -1010,7 +1024,7 @@ Carsten Haitzler (486): fix install no asflags recurse oosp accidentally put this in. damn! - openembedded build files... this makes life so easy to build efl for embedded... :) + openembedded build files... this makes life so easy to build efl for embedded... :) cleaner configure.in oopsies. fixies bart patch for debian package stuff @@ -1024,12 +1038,12 @@ Carsten Haitzler (486): 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". :) + 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... + 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... @@ -1044,29 +1058,29 @@ Carsten Haitzler (486): 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 + make e17 pass make distcheck.... and... ASPARAGUS time... http://enlightenment.freedesktop.org for tarballs asparagus! full asparagus xpm segv fix - asparagus... make e17 distcheck. + 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 + * 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 + * 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. + 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. + 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. @@ -1076,10 +1090,10 @@ Chris Ross (20): 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. + 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. + 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? @@ -1098,11 +1112,11 @@ Dan Sinclair (1): 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 :) + 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 ... just some cleanup ... + 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 ... just some cleanup ... Horms (4): rpm -ta now works on a tarball produced by "make distcheck" and friends @@ -1115,7 +1129,7 @@ Ibukun Olumuyiwa (2): 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. + 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): @@ -1133,89 +1147,89 @@ Laurence J. Lane (15): stuff removed dh_testversion stuff - added patch by David N. Welton to cleanup configure handling Thanks, David. + 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 + 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. + 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) + 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) + 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. + 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. + 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. + 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 <[email protected]>. - BMP loader from Isaac Richards <[email protected]>. It currently has issues with progressive loading, so don't use it with feh. :-) + BMP loader from Isaac Richards <[email protected]>. 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. + 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 + "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) + 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. + 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) + 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) + 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) + 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. :-) + 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) @@ -1241,9 +1255,9 @@ Mike Frysinger (3): 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. + 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): @@ -1264,37 +1278,37 @@ 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 + 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. + 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. + 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. + 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 <[email protected]>. 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 <[email protected]>, 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. ;) + Cleanup the spec file a bunch with a patch from Joakim Bodin <[email protected]>. 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 <[email protected]>, 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. + 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 @@ -1318,39 +1332,39 @@ Tilman Sauerbeck (17): Tom Gilbert (85): Shaddup ;) shaddup ;) - AARGH. Godamn file decriptor leak which has been driving me CRAZY for a WEEK! Got the BASTARD. DIE! + 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 + 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) + 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... + 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. + 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. + 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. + 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). + 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 @@ -1372,7 +1386,7 @@ Tom Gilbert (85): 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? + 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 @@ -1390,17 +1404,17 @@ Tom Gilbert (85): 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. + 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 <[email protected]> foo Fix broken ordering. - Fix from [email protected] for segvs when tiling images seamlessly (if they are an odd number of pixels in height). + Fix from [email protected] 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. + ->data was getting leaked, as _tidyup() only free()s data for nodes of type CHAR. Vincent Torri (5): Doxygen documentation diff --git a/configure.ac b/configure.ac index 6fa7fac..bc5a8d0 100644 --- a/configure.ac +++ b/configure.ac @@ -1,7 +1,7 @@ # get rid of that stupid cache mechanism rm -f config.cache -AC_INIT([imlib2], [1.4.9], [[email protected]]) +AC_INIT([imlib2], [1.4.10], [[email protected]]) AC_CONFIG_SRCDIR(configure.ac) AC_CONFIG_MACRO_DIR([m4]) AC_CANONICAL_BUILD @@ -23,7 +23,7 @@ define([AC_LIBTOOL_LANG_F77_CONFIG], [:])dnl LT_INIT VERSION_CUR=5 -VERSION_REV=9 +VERSION_REV=10 VERSION_AGE=4 lt_version=${VERSION_CUR}:${VERSION_REV}:${VERSION_AGE} AC_SUBST(lt_version) --
