Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
branch: master
Commits on Jan 5, 2015
  1. @directhex

    Bump to 3.12 and tag

    directhex authored
Commits on Nov 25, 2014
  1. Newer giflib defines the EGifOpen correctly. Prevent build failures

    Marek Habersack authored
    by making its redefinition conditional on presence of the EgifOpen
    function declaration.
Commits on Sep 29, 2014
  1. @nealef
Commits on Sep 4, 2014
  1. @directhex

    Bump to 3.8 and tag

    directhex authored
Commits on Sep 1, 2014
  1. @akoeplinger

    Fixed wrong order of arguments to gdip_unit_conversion() calls

    akoeplinger authored
    The argument order in the gdip_unit_conversion calls were wrong, i.e. dpi and type were switched.
    Reported by Michael Tautschnig in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=688361
  2. @akoeplinger

    Properly fix the issue in GdipAddPathStringI from 429cf5d

    akoeplinger authored
    It wasn't a GCC bug at all, the code in the function is actually invalid C as the pointer to rect stops being valid after the if.
    The proper fix is to move the declaration of rect outside the if, so it stays valid in the entire function.
Commits on Aug 27, 2014
  1. @directhex
  2. @alexrp

    Merge pull request #25 from mono/gcc-bug

    alexrp authored
    Workaround GCC 4.7+ optimizer bug by disabling optimization for the affected function
  3. @alexrp

    Merge pull request #23 from ralovich/image-fixes2

    alexrp authored
    Image: fix function declarations
  4. @alexrp

    Merge pull request #22 from akoeplinger/gitignore

    alexrp authored
    Ignored a bunch of generated files
  5. @akoeplinger

    Workaround GCC 4.7+ optimizer bug by disabling optimization for the a…

    akoeplinger authored
    …ffected function
    
    GCC 4.7+ incorrectly optimizes the GdipAddPathStringI() function, resulting in the value of r->Y being incorrect in the called GdipAddPathString() function.
    By compiling with -O0 or adding a printf("Test") after r = &rect this doesn't happen.
Commits on Jul 17, 2014
  1. Image: fix function declarations

    RALOVICH, Kristof authored
    First, there was a typo in both GdipRemovePropertyItem and GdipSetPropertyItem. Second, add missing declaration of already defined GdipGetAllPropertyItems function
    
    Commit licensed under MIT/X11.
Commits on Jul 5, 2014
  1. @akoeplinger

    Ignored a bunch of generated files

    akoeplinger authored
    Those are created by autotools (test-driver, compile) and test runs
Commits on Jun 18, 2014
  1. @alexrp

    Merge pull request #21 from ranma42/tests

    alexrp authored
    Enable testing
  2. @ranma42

    Fix testbits

    ranma42 authored
    The testbits test failure indicates an issue in the definition and use
    of the BitmapData structure. Document the issue and work around it in
    the test, to allow the test to actually check the behaviour of the
    implementation.
    
    Also see https://bugzilla.xamarin.com/show_bug.cgi?id=20682
    
    Commit licensed under MIT/X11.
  3. @ranma42

    Enable autotools-based testing

    ranma42 authored
    Automakes provides facilities to run tests and collect their
    results. This can be enabled simply by adding a TESTS setting which
    contains the names of the test programs to be run.
    
    All of the tests in the "tests/" folder are now enabled except for
    testgdi, which requires interaction with the user.
    
    Commit licensed under MIT/X11.
Commits on Jun 11, 2014
  1. @alexrp

    Merge pull request #20 from akoeplinger/fix-dependency-check

    alexrp authored
    Check that Cairo is installed in configure.ac again
  2. @akoeplinger

    Check that Cairo is installed in configure.ac again

    akoeplinger authored
    This was removed accidently in PR#18.
    
    Commit licensed under MIT/X11.
  3. @alexrp

    Merge pull request #18 from akoeplinger/remove-cairo-and-pixman

    alexrp authored
    Removed bundled cairo and pixman
Commits on Jun 10, 2014
  1. @akoeplinger

    Removed bundled cairo and pixman.

    akoeplinger authored
    Commit licensed under MIT/X11.
  2. @alexrp

    Merge pull request #16 from ranma42/fix-testgdi

    alexrp authored
    Fix testgdi
  3. @alexrp

    Merge pull request #15 from ranma42/improve-autotools

    alexrp authored
    Improve autotools
  4. @alexrp

    Merge pull request #19 from akoeplinger/fix-cairo-pen-adjustment-on-l…

    alexrp authored
    …inux
    
    Removed pen adjustment for Cairo >= 1.12
Commits on Jun 9, 2014
  1. @kumpera

    Merge pull request #17 from akoeplinger/fix-libgdiplus-crash

    kumpera authored
    Fix libgdiplus crash
Commits on Jun 8, 2014
  1. @akoeplinger

    Removed pen adjustment for Cairo >= 1.12

    akoeplinger authored
    On modern versions of the Cairo library pen adjustment is no longer necessary.
    This fixes errors in the Mono System.Drawing testsuite caused by wrong adjustment when using such a version.
  2. @akoeplinger

    Move the call to png_get_PLTE() up so it initializes the png_palette …

    akoeplinger authored
    …and num_palette variables in all cases.
    
    This fixes accessing an uninitialized variable e.g. in the 2bpp case.
    
    Commit licensed under MIT/X11.
  3. @akoeplinger

    Use correct type for variable

    akoeplinger authored
    This was changed in PR#8, but the correct variable type is png_charp according to a clang warning and the documentation (http://refspecs.linuxfoundation.org/LSB_3.1.0/LSB-Desktop-generic/LSB-Desktop-generic/libpng12.png.get.iccp.1.html).
    
    Commit licensed under MIT/X11.
  4. @akoeplinger

    Fix crash due to uninitialized variable

    akoeplinger authored
    png_get_tRNS doesn't initialize supplied variables when no transparency chunk is found (see http://sourceforge.net/p/png-mng/mailman/message/27133707/).
    In that case the trans_alpha variable still has its initial (indeterminate) value which leads to the crash when using it in set_pixel_bgra() later on.
    We now check if the transparency chunk is valid before accessing it.
    
    Commit licensed under MIT/X11.
  5. @ranma42

    Cleanup autogen.sh

    ranma42 authored
    Remove dead code and do not warn about configure being run without
    arguments if configure is not going to be run at all.
    
    This also makes it more consistent with the mono autogen.sh.
  6. @ranma42

    Silence autotools warning

    ranma42 authored
    Modern autotools complain that
    
    warning: autoconf input should be named 'configure.ac', not
    'configure.in'
    
    The file is explicitly used in autogen.sh, which needs to be updated
    consistently.
  7. @ranma42

    Fix testgdi

    ranma42 authored
    The testgdi program makes GDI+ calls without invoking
    GdiplusStartup().
    
    This fixes the following errors:
    
    got st: 3 expected Okgot st: 2 expected Okjpg drawn
    got st: 3 expected Okgot st: 2 expected Oktif drawn
    got st: 3 expected Okgot st: 2 expected Okgif drawn
    got st: 3 expected Okgot st: 2 expected Okpng drawn
    got st: 3 expected Okgot st: 2 expected Okbmp drawn
Commits on Apr 11, 2014
  1. @kumpera

    Fix float-to-int conversion to handle recent versions of clang.

    kumpera authored
    Recent versions of clang are much more aggressive in exploiting undefined
    behavior. Given MaxInt cannot be encoded as float, the compiler would miscompile
    the macro and produce a negative number.
Commits on Apr 10, 2014
  1. @kumpera

    Make sure we include cairo's feature test header so CAIRO_HAS_QUARTZ_…

    kumpera authored
    …FONT has a change of been properly checked against.
Commits on Dec 16, 2013
  1. @alexrp

    Merge pull request #14 from directhex/master

    alexrp authored
    Use FreeType macros for tttables.h inclusion
  2. @directhex

    Use FreeType macros for tttables.h inclusion

    directhex authored
    As of FreeType 2.1.6 (November 2003), using #include to include Freetype libraries directly is not supported.
    
    This has come to a head, as in FreeType 2.5.0, the location of headers has been moved around, breaking building of libgdiplus.
    
    This slight change uses the "official" way to include the required header file, without breaking building on older versions of the library.
Something went wrong with that request. Please try again.