| commit | 842cf8bf97c8c12b64c5c189e2f0781a3330aba8 | |
| author | erikchen <erikchen@chromium.org> | |
| Wed, 15 Jul 2015 20:45:01 +0000 (15 13:45 -0700) | ||
| committer | Commit bot <commit-bot@chromium.org> | |
| Wed, 15 Jul 2015 20:45:51 +0000 (15 20:45 +0000) | ||
| tree | a79c145596c3f3f243f72d0678c3a69774cee9a4 | treesnapshot (tar.gz zip) |
| parent | b53d5d2078793f80b78ee4b6229c22fae731add1 | commitdiff |
mac: Fix SkiaUtilsMacTest.NSImageRepToSkBitmap.
This test used the function CreateNSImage(), which first created an NSImage, and
then attached a NSBitmapImageRep to the NSImage. The problem is that the
NSBitmapImageRep was implicitly created, and its colorspace depended on the
linked version of OSX and hardware on the machine. I changed the implementation
of CreateNSImage() to use an explicitly created NSBitmapImageRep. I also changed
the colorspace of the function CreateNSImage() and the unit test to always be
+[NSColorSpace genericRGBColorSpace], to avoid potential wacky scenarios that
could arise with +[NSColorSpace deviceRGBColorSpace].
BUG=463170
Review URL: https://codereview.chromium.org/1225233004
Cr-Commit-Position: refs/heads/master@{#338909}
This test used the function CreateNSImage(), which first created an NSImage, and
then attached a NSBitmapImageRep to the NSImage. The problem is that the
NSBitmapImageRep was implicitly created, and its colorspace depended on the
linked version of OSX and hardware on the machine. I changed the implementation
of CreateNSImage() to use an explicitly created NSBitmapImageRep. I also changed
the colorspace of the function CreateNSImage() and the unit test to always be
+[NSColorSpace genericRGBColorSpace], to avoid potential wacky scenarios that
could arise with +[NSColorSpace deviceRGBColorSpace].
BUG=463170
Review URL: https://codereview.chromium.org/1225233004
Cr-Commit-Position: refs/heads/master@{#338909}
| skia/ext/skia_utils_mac_unittest.mm | diffblobblamehistory |