On 14/03/2016 19:37, Phil Race wrote:
:
http://cr.openjdk.java.net/~alanb/8142968/2/jdk/src/java.desktop/macosx/classes/com/apple/laf/AquaUtils.java.sdiff.html
368 // can we use ClassLoader.getSystemClassLoader here?
369 final ClassLoader launcherClassLoader =
ClassLoaders.appClassLoader();
I suspect yes, although it doesn't really matter and maybe this
workaround method
is not needed any more but that is for another day.
Right, although we haven't really changed anything here. I assume the
check for com.installshield.wizard.platform.macosx.MacOSXUtils is
something that came from Apple originally.
http://cr.openjdk.java.net/~alanb/8142968/2/jdk/src/java.desktop/share/classes/java/awt/color/ICC_Profile.java.sdiff.html
:
This bug would affected mainline so why was this change not fixed there ?
I'm curious how it was found since there is no bug report mentioned.
Also I wonder if we really have to have to open a stream to test
existence ?
I think we can revert this. There was a time where it wasn't possible to
get a URL to a resource.
:
Not sure why deleting a few unused imports needed to be part of this
change.
We had to remove several unused imports in places where it attempted to
import types from non-exported packages or from modules that are
declared as a dependency.
:
http://cr.openjdk.java.net/~alanb/8142968/2/jdk/src/java.desktop/share/classes/sun/print/ServiceDialog.java.sdiff.html
Why were these changes made in jake ? I thought the policy was to make
all changes in mainline unless they could only be made in jake ?
This is also related to URL to resources, I think we can revert this one
too.
Also I
1) don't see where/how the stream is closed
2) The doPrivileged at line 2823 no longer seems likely to be necessary.
BTW: try (in) { .. } will ensure that the input stream is closed. Also
the doPrivileged is needed as there may be user code on the stack.
I'll leave it to Yuri for the AWT tests as I think he contributed all
the changes to the tests.
-Alan.