summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* Add get tone curve params functionMarti Maria2020-06-152-1/+12
|
* Update cmsalpha.cMarti Maria2020-06-151-1/+1
| | | | Fixed bound check
* Fix datarace conditionMarti Maria2020-06-081-21/+8
| | | | | - Fix datarace condition on cmsGetContext() Thanks to Jan Wassenberg for finding this elussive bug - Ged rid of commented out code
* Added a code guardMarti Maria2020-06-051-0/+2
| | | | To prevent segfaults
* Fix slight change of behaviour on create transformslcms2.10AMarti Maria2020-06-051-39/+42
| | | | | | in 2.10 cmsFLAGS_NOOPTIMIZE changed slightly the previous behaviour as prevented any kind of optimization. That was enough to crash some complex workflows of cerain apps,, so here is a rollback.
* Fix Lut16 matrixMarti Maria2020-06-021-10/+7
| | | | Matrix should always be 3x3
* More fixes from Artifexlcms2.10Marti Maria2020-05-263-4/+13
| | | | | More fixes from Artifex Thank you!
* Merged SSE matrix-shaper optimizerMarti Maria2020-05-251-4/+6
| | | | | | It makes small difference, but here you go! In my computer, matrix shaper runs times x 2.4 faster the core lcms2 without SSE is about times x 2
* Changes from ArtifexMarti Maria2020-05-254-19/+20
| | | | Many thanks for sharing!
* Update cmscgats.cMarti Maria2020-05-231-41/+23
| | | | | - Get rid of a bunch of code from uncertain origin (it was doing nonsense anyway) - Format the function layout
* Regeneration of configuremm22020-05-231-37/+110
|
* fix overoptimization issueMarti Maria2020-05-211-1/+1
| | | | The cmsFLAGS_NOOPTIMIZE was ignored when a plug-in was used
* cmsalpha.c : fix swapped-endian caseTanner2020-05-201-4/+5
| | | | | | | | Without this fix, FormatterPos() is incapable of handling swapped-endian requests (note that no case returns `2`, the index into the swapped-endian row!) Fix is simple: when 2-bytes are requested, but half-float is not, check endianness. If T_ENDIAN16, return 16SE; otherwise, return 16. (Also a minor typo fix on a neighboring comment.)
* Merge pull request #213 from kmilos/exportsMarti Maria2020-05-171-3/+1
|\ | | | | | | Add missing and remove duplicate exports. Looks good, thank you.
| * Add missing and remove duplicate exportsMiloš Komarčević2020-05-031-3/+1
| |
* | Added contibutorsMarti Maria2020-05-171-70/+38
| | | | | | | | Many thanks to Philipp Knechtges for the improved PS code
* | Cleanup PostScript CSA GenerationPhilipp Knechtges2020-05-161-28/+80
|/ | | | | | | | As reported some time ago [1], some PS interpreters have issues with the CSA code lcms2 emits when gamma tables are encountered in the icc profile. This patch picks up the ideas from the mentioned mailing list thread and implements them in lcms2. [1] https://sourceforge.net/p/lcms/mailman/message/33182987/
* Fix end of lines messMarti Maria2020-04-291-0/+1
| | | | Hopefully it is now fixed
* Added an extra checkMarti Maria2020-04-271-0/+1
| | | | Just to avoid link warnings
* add missing functions to the def fileMarti Maria2020-04-271-0/+24
| | | | The DEF file allows fine tune name decoration on 32-bit DLL builds.
* Attempt to upgrade configure script to 2.10Marti Maria2020-04-251-3/+2
| | | | Fix plugins issue as well.
* adjust copyright yearMarti Maria2020-04-102-2/+2
| | | | and remove duplicated symbol
* Revert "Update autotools"Marti Maria2020-04-051-2/+3
| | | | This reverts commit 08fe5c96ed4089b21fa6e6b6248a54a5a4396cb4.
* Update MINGW supportMarti Maria2020-04-021-0/+2
| | | | Fixed a situation where MINGW fails
* Update autotoolsMarti Maria2020-04-011-3/+2
| | | | Update configure script to 2.10
* Prepare 2.10 release (I)Marti Maria2020-04-0126-42/+33
| | | | Some minor changes to prepeare 2.10 release
* Update cmsvirt.cMarti Maria2019-10-031-1/+2
| | | | Fix wrong white fixup on V2 PCS when building input profiles. Thanks to Maurice Luttmer for finding the bug.
* Merge branch 'master' of https://github.com/mm2/Little-CMSMarti Maria2019-09-271-1/+1
|\
| * Buffer overflow fix.Gaurav2019-09-241-1/+1
| | | | | | | | BuildColorantList() concatenates 32 bit data in each iteration. The max value of iteration can be 16. Thus buffer overflow can occur with storage of 128 bytes. It would need 512 bytes in worst case.
* | port of several fixupsMarti Maria2019-09-274-9/+12
|/ | | | From assorted sources, small fixes
* merge branchMarti Maria2019-07-261-2/+19
| | | | plane bytes, not pixels
* fix array bounds on writing matricesMarti Maria2019-05-121-44/+40
| | | | fixed a out-of-bounds error on certain conditions
* Merge branch 'master' of https://github.com/mm2/Little-CMSMarti Maria2019-04-1812-466/+466
|\
| * provide a way to avoid "register" storage classMarti Maria2019-01-2112-466/+466
| | | | | | | | C++17 seems to require to travel back in time and modify C99 spec
* | expose MD5Marti Maria2019-04-181-16/+12
|/ | | | let people to use MD5 internals. This is added in the plug-ins API
* Fix double free of curve data structure.Sebastian Rasmussen2018-12-151-6/+5
| | | | | | | | | | | | When the call to CurvesAlloc() fails in OptimizeByJoiningCurves() the cleanup code will try to free the ObtainedCurves twice. This is because ObtainedCurves has also been inserted into the pipeline by the call to cmsPipelineInsertStage(). This commit NULLs the ObtainedCurves variable immediately after the insertion into the pipeline was succesful and also moves later use of ObtainedCurves earlier to compensate for setting ObtainedCurves to NULL. This avoids the double free.
* #174: Cast error in cmsxform.c reported by GCC 8tp2018-12-071-1/+1
|
* Reported bug from quangnh89Marti Maria2018-09-101-0/+1
| | | | | | | | | | | | | - Turns a warning into a fatal error. I have tested the profile against a normal use and found despite this is a real bug, it is harmless in the security context. The out of bounds condition can only be reached by building a special program. Yet-existing, normal code are unaffected as it never needs to do this particular sequence. I agree it is better to turn this warning into a fatal error in order to keep safe. NOT RELATED WITH SECURITY. IT CANNOT BE USED TO DO ANY EXPLOIT.
* Upgrade Visual studio 2017 15.8Marti Maria2018-08-151-4/+10
| | | | | | - Upgrade to 15.8 - Add check on CGATS memory allocation (thanks to Quang Nguyen for pointing out this)
* add extra checkMarti Maria2018-07-141-0/+5
| | | | For out of mem confition
* fixed code for very unusual platformMarti Maria2018-07-131-2/+2
| | | | big endian machines without int64 support
* Extra typosluz.paz2018-05-212-6/+6
|
* Misc. typosluz.paz2018-05-217-10/+10
| | | Found via `codespell -q 3`
* Merge branch 'master' of https://github.com/mm2/Little-CMSMarti Maria2018-05-028-42/+42
|\
| * Merge pull request #147 from sebras/masterMarti Maria2018-03-075-22/+22
| |\ | | | | | | Reindent some lines, fixing gcc indentation warnings. Looks great, thank you
| | * Reindent some lines, fixing gcc indentation warnings.Sebastian Rasmussen2017-12-135-22/+22
| | | | | | | | | | | | | | | LittleCMS now compiles without warnings using: CFLAGS=-Wall ./configure && make
| * | Fix some typos in comments and messages (found by codespell)Stefan Weil2018-03-071-1/+1
| | | | | | | | | | | | Signed-off-by: Stefan Weil <sw@weilnetz.de>
| * | Merge pull request #152 from stweil/masterMarti Maria2018-03-073-19/+19
| |\ \ | | | | | | | | Replace non-ASCII characters. Looks great, thank you!
| | * | Replace non-ASCII charactersStefan Weil2018-02-243-19/+19
| | | | | | | | | | | | | | | | Signed-off-by: Stefan Weil <sw@weilnetz.de>
* | | | add swapped-endian formats in alpha handlingMarti Maria2018-05-021-18/+94
|/ / / | | | | | | | | | fixes #159
View Lorry Controller Status