MozillaFirefox/mozilla-bmo1504834-part3.patch
Wolfgang Rosenauer 351b951583 - Mozilla Firefox 129.0
https://www.mozilla.org/en-US/firefox/129.0/releasenotes
  MFSA 2024-33 (bsc#1228648))
  * CVE-2024-7518 (bmo#1875354)
    Fullscreen notification dialog can be obscured by document content
  * CVE-2024-7519 (bmo#1902307)
    Out of bounds memory access in graphics shared memory handling
  * CVE-2024-7520 (bmo#1903041)
    Type confusion in WebAssembly
  * CVE-2024-7521 (bmo#1904644)
    Incomplete WebAssembly exception handing
  * CVE-2024-7522 (bmo#1906727)
    Out of bounds read in editor component
  * CVE-2024-7523 (bmo#1908344)
    Document content could partially obscure security prompts
  * CVE-2024-7524 (bmo#1909241)
    CSP strict-dynamic bypass using web-compatibility shims
  * CVE-2024-7525 (bmo#1909298)
    Missing permission check when creating a StreamFilter
  * CVE-2024-7526 (bmo#1910306)
    Uninitialized memory used by WebGL
  * CVE-2024-7527 (bmo#1871303)
    Use-after-free in JavaScript garbage collection
  * CVE-2024-7528 (bmo#1895951)
    Use-after-free in IndexedDB
  * CVE-2024-7529 (bmo#1903187)
    Document content could partially obscure security prompts
  * CVE-2024-7530 (bmo#1904011)
    Use-after-free in JavaScript code coverage collection
  * CVE-2024-7531 (bmo#1905691)

OBS-URL: https://build.opensuse.org/package/show/mozilla:Factory/MozillaFirefox?expand=0&rev=1168
2024-08-10 12:42:53 +00:00

18 lines
964 B
Diff

# HG changeset patch
# Parent 09cd4ac2cc607e85aa572425b824fbab386af607
For FF68, AntiAliasing of XULTexts seem to be broken on big endian (s390x). Text and icons of the sandwich-menu to the
right of the address bar, as well as plugin-windows appears transparant, which usually means unreadable (white on white).
diff --git a/gfx/skia/skia/src/opts/SkBlitMask_opts.h b/gfx/skia/skia/src/opts/SkBlitMask_opts.h
--- a/gfx/skia/skia/src/opts/SkBlitMask_opts.h
+++ b/gfx/skia/skia/src/opts/SkBlitMask_opts.h
@@ -210,6 +210,8 @@ namespace SK_OPTS_NS {
// ~~~>
// a = 1*aa + d(1-1*aa) = aa + d(1-aa)
// c = 0*aa + d(1-1*aa) = d(1-aa)
+ // TODO: Check this for endian-issues!
+ // Do we need to switch 255 to the front for all of those tuples?
return (aa & Sk4px(skvx::byte16{0,0,0,255, 0,0,0,255, 0,0,0,255, 0,0,0,255}))
+ d.approxMulDiv255(aa.inv());
};