I moved the FontPrivilege test into open. I did one small change to it, instead of calling setSecurityManager() in the beginning of the test, I added the tag /security=java.lang.SecurityManager which should have about the same effect.
http://cr.openjdk.java.net/~rkennke/6904882/webrev.02/ Is this ok to push? Also needs one more review, Igor, Dmitri, anyone else left in 2d? /Roman Am Montag, den 30.11.2009, 11:49 -0800 schrieb Phil Race: > Like the last one, this could use a regression test - even though > there's a JCK failure. FWIW a trivial one like this :- > > import java.awt.*; > import java.io.File; > > public class CreateFontSMTest { > public static void main(String[] args) throws Exception { > System.setSecurityManager(new SecurityManager()); > Font font = new Font("Dialog", Font.PLAIN, 12); > System.out.println(font.getFontName()); > } > } > > will work, but I also found we already have a test .. which I > guess never got run :-(, called FontPrivilege.java that does > something very similar. Its also in closed and I leave it > up to you whether to open that or add a new one but opening > it would be best. > > -phil. > > > > Roman Kennke wrote: > > This patches fixes bug #6904882. As suggested by Phil it puts the whole > > static initializer in a privileged block, grouping the existing 4 > > smaller blocks into one, and importantly including the offending > > File.exists() call. This requires to make the static fields non-final. > > > > http://cr.openjdk.java.net/~rkennke/6904882/webrev.00/ > > > > Ok to commit? > > > > /Roman > > > >
