RE: gnu/testlet/java/lang/Thread/security10.java...

RE: gnu/testlet/java/lang/Thread/security10.java...

[Jeroen Frijters]

[-- Attachment #1: Type: text/plain, Size: 2822 bytes --]

Michael Koch wrote:
> Am Dienstag, 18. Januar 2005 20:04 schrieb David Daney:
> > David Daney wrote:
> > > Michael Koch wrote:
> > >> Am Dienstag, 18. Januar 2005 19:24 schrieb David Daney:
> > >>> Causes:
> > >>> Exception in thread "main" java.lang.SecurityException
> > >>>         at
> > >>> gnu.testlet.java.lang.Thread.security10$MySecurityManager.check
> > >>>Acce ss (security10.java:34)
> > >>>         at java.lang.Thread.checkAccess(Thread.java:1124)
> > >>>         at java.lang.Thread.setPriority(Thread.java:871)
> > >>>         at java.lang.Thread.init(Thread.java:293)
> > >>>         at java.lang.Thread.<init>(Thread.java:438)
> > >>>         at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
> > >>>         at java.security.AccessController.doPrivileged(Native
> > >>> Method) at
> > >>> java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:1
> > >>>21) at
> > >>> java.lang.ref.Finalizer.runFinalization(Finalizer.java:126) at
> > >>> java.lang.Runtime.runFinalization0(Native Method) at
> > >>> java.lang.Runtime.runFinalization(Runtime.java:656) at
> > >>> java.lang.System.runFinalization(System.java:759) at
> > >>> gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:23
> > >>>2) at
> > >>> gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)
> > >>>
> > >>> When run with Sun's jre.
> > >>
> > >> which version of the JRE ? I used 1.4.2 and it passes the tests.
> > >
> > > This is running on  FC3 on i686
> > >
> > > [daney@dl2 mauve]$ java -version
> > > java version "1.4.2_06"
> > > Java(TM) 2 Runtime Environment, Standard Edition (build
> > > 1.4.2_06-b03) Java HotSpot(TM) Client VM (build 1.4.2_06-b03,
> > > mixed mode)
> > >
> > > I have no default security policy installed that I am aware of. 
> > > It could be an issue with security policies.
> >
> > After further thought, it is not a security policy problem.  Your
> > security manager always throws SecurityException on any Thread
> > method. Sun's jre is trying to create a new Thread when
> > System.runFinalization() is called, and will always fail with your
> > SecurityManager installed.
> >
> > Perhaps you should save a reference to the original SecurityManager
> > and restore it when the test is finished.  That would leave the
> > test harness in the same state after the Thread.security10 test as
> > before.
> 
> Hmmm, setting "null" should be enough as no SecurityManager 
> may be set initiallly.
> 
> Daney: Can you try if this works for you ?

I ran into this problem as well and I'm committing the attached patch to
solve this.

Regards,
Jeroen

2005-01-31  Jeroen Frijters  <jeroen@frijters.net>

        * gnu/testlet/java/lang/Thread/security10.java:
        Added code to save and restore security manager.

[-- Attachment #2: security10.patch.txt --]
[-- Type: text/plain, Size: 813 bytes --]

Index: gnu/testlet/java/lang/Thread/security10.java
===================================================================
RCS file: /cvs/mauve/mauve/gnu/testlet/java/lang/Thread/security10.java,v
retrieving revision 1.1
diff -u -r1.1 security10.java
--- gnu/testlet/java/lang/Thread/security10.java	13 Jan 2005 09:49:35 -0000	1.1
+++ gnu/testlet/java/lang/Thread/security10.java	31 Jan 2005 11:24:08 -0000
@@ -37,6 +37,19 @@
   
   public void test (TestHarness h)
   {
+    SecurityManager secman = System.getSecurityManager();
+    try
+      {
+        testImpl(h);
+      }
+    finally
+      {
+        System.setSecurityManager(secman);
+      }
+  }
+
+  private void testImpl(TestHarness h)
+  {
     h.checkPoint("Thread creation");
     
     Thread testThread = new Thread();

Re: gnu/testlet/java/lang/Thread/security10.java...

[David Daney]

Jeroen Frijters wrote:
> 
> I ran into this problem as well and I'm committing the attached patch to
> solve this.
> 
> Regards,
> Jeroen
> 
> 2005-01-31  Jeroen Frijters  <jeroen@frijters.net>
> 
>         * gnu/testlet/java/lang/Thread/security10.java:
>         Added code to save and restore security manager.
> 
> 

Still not quite right.  I just committed this:

2005-02-02  David Daney <ddaney@avtrex.com>
	* gnu/testlet/java/lang/Thread/security10.java:
	Allow setSecurityManager.

Index: security10.java
===================================================================
RCS file: /cvs/mauve/mauve/gnu/testlet/java/lang/Thread/security10.java,v
retrieving revision 1.2
diff -c -p -r1.2 security10.java
*** security10.java     31 Jan 2005 11:26:16 -0000      1.2
--- security10.java     2 Feb 2005 18:17:43 -0000
***************
*** 1,6 ****
   // Tags: JDK1.0

! // Copyright (C) 2004 Free Software Foundation, Inc.
   // Written by Michael Koch (konqueror@gmx.de)

   // This file is part of Mauve.
--- 1,6 ----
   // Tags: JDK1.0

! // Copyright (C) 2004, 2005 Free Software Foundation, Inc.
   // Written by Michael Koch (konqueror@gmx.de)

   // This file is part of Mauve.
*************** package gnu.testlet.java.lang.Thread;
*** 24,29 ****
--- 24,30 ----

   import gnu.testlet.Testlet;
   import gnu.testlet.TestHarness;
+ import java.security.Permission;

   public class security10 implements Testlet
   {
*************** public class security10 implements Testl
*** 33,38 ****
--- 34,47 ----
       {
         throw new SecurityException();
       }
+
+     public void checkPermission(Permission perm)
+     {
+       RuntimePermission p = new RuntimePermission("setSecurityManager");
+       if (p.implies(perm))
+         return;
+       super.checkPermission(perm);
+     }
     }

     public void test (TestHarness h)

RE: gnu/testlet/java/lang/Thread/security10.java...

[Jeroen Frijters]

David Daney wrote:
> Still not quite right.  I just committed this:

Indeed. Thanks. The default Classpath security policy is too lenient, so
I didn't notice this.

Regards,
Jeroen

Re: gnu/testlet/java/lang/Thread/security10.java...

[David Daney]

Michael Koch wrote:
> Am Dienstag, 18. Januar 2005 20:04 schrieb David Daney:
> 
>>David Daney wrote:
>>
>>
>>After further thought, it is not a security policy problem.  Your
>>security manager always throws SecurityException on any Thread
>>method. Sun's jre is trying to create a new Thread when
>>System.runFinalization() is called, and will always fail with your
>>SecurityManager installed.
>>
>>Perhaps you should save a reference to the original SecurityManager
>>and restore it when the test is finished.  That would leave the
>>test harness in the same state after the Thread.security10 test as
>>before.
> 
> 
> Hmmm, setting "null" should be enough as no SecurityManager may be set 
> initiallly.
> 
> Daney: Can you try if this works for you ?
> 

I could be several days before I can try it.

David Daney.

Re: gnu/testlet/java/lang/Thread/security10.java...

[Michael Koch]

Am Dienstag, 18. Januar 2005 20:04 schrieb David Daney:
> David Daney wrote:
> > Michael Koch wrote:
> >> Am Dienstag, 18. Januar 2005 19:24 schrieb David Daney:
> >>> Causes:
> >>> Exception in thread "main" java.lang.SecurityException
> >>>         at
> >>> gnu.testlet.java.lang.Thread.security10$MySecurityManager.check
> >>>Acce ss (security10.java:34)
> >>>         at java.lang.Thread.checkAccess(Thread.java:1124)
> >>>         at java.lang.Thread.setPriority(Thread.java:871)
> >>>         at java.lang.Thread.init(Thread.java:293)
> >>>         at java.lang.Thread.<init>(Thread.java:438)
> >>>         at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
> >>>         at java.security.AccessController.doPrivileged(Native
> >>> Method) at
> >>> java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:1
> >>>21) at
> >>> java.lang.ref.Finalizer.runFinalization(Finalizer.java:126) at
> >>> java.lang.Runtime.runFinalization0(Native Method) at
> >>> java.lang.Runtime.runFinalization(Runtime.java:656) at
> >>> java.lang.System.runFinalization(System.java:759) at
> >>> gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:23
> >>>2) at
> >>> gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)
> >>>
> >>> When run with Sun's jre.
> >>
> >> which version of the JRE ? I used 1.4.2 and it passes the tests.
> >
> > This is running on  FC3 on i686
> >
> > [daney@dl2 mauve]$ java -version
> > java version "1.4.2_06"
> > Java(TM) 2 Runtime Environment, Standard Edition (build
> > 1.4.2_06-b03) Java HotSpot(TM) Client VM (build 1.4.2_06-b03,
> > mixed mode)
> >
> > I have no default security policy installed that I am aware of. 
> > It could be an issue with security policies.
>
> After further thought, it is not a security policy problem.  Your
> security manager always throws SecurityException on any Thread
> method. Sun's jre is trying to create a new Thread when
> System.runFinalization() is called, and will always fail with your
> SecurityManager installed.
>
> Perhaps you should save a reference to the original SecurityManager
> and restore it when the test is finished.  That would leave the
> test harness in the same state after the Thread.security10 test as
> before.

Hmmm, setting "null" should be enough as no SecurityManager may be set 
initiallly.

Daney: Can you try if this works for you ?


Michael
-- 
Homepage: http://www.worldforge.org/

Re: gnu/testlet/java/lang/Thread/security10.java...

[David Daney]

David Daney wrote:
> Michael Koch wrote:
> 
>> Am Dienstag, 18. Januar 2005 19:24 schrieb David Daney:
>>
>>
>>> Causes:
>>> Exception in thread "main" java.lang.SecurityException
>>>         at
>>> gnu.testlet.java.lang.Thread.security10$MySecurityManager.checkAcce
>>> ss (security10.java:34)
>>>         at java.lang.Thread.checkAccess(Thread.java:1124)
>>>         at java.lang.Thread.setPriority(Thread.java:871)
>>>         at java.lang.Thread.init(Thread.java:293)
>>>         at java.lang.Thread.<init>(Thread.java:438)
>>>         at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
>>>         at java.security.AccessController.doPrivileged(Native
>>> Method) at
>>> java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:121)
>>>         at
>>> java.lang.ref.Finalizer.runFinalization(Finalizer.java:126) at
>>> java.lang.Runtime.runFinalization0(Native Method) at
>>> java.lang.Runtime.runFinalization(Runtime.java:656) at
>>> java.lang.System.runFinalization(System.java:759) at
>>> gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:232)
>>>         at
>>> gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)
>>>
>>> When run with Sun's jre.
>>
>>
>>
>> which version of the JRE ? I used 1.4.2 and it passes the tests.
>>
> 
> This is running on  FC3 on i686
> 
> [daney@dl2 mauve]$ java -version
> java version "1.4.2_06"
> Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_06-b03)
> Java HotSpot(TM) Client VM (build 1.4.2_06-b03, mixed mode)
> 
> I have no default security policy installed that I am aware of.  It 
> could be an issue with security policies.
> 

After further thought, it is not a security policy problem.  Your 
security manager always throws SecurityException on any Thread method. 
Sun's jre is trying to create a new Thread when System.runFinalization() 
is called, and will always fail with your SecurityManager installed.

Perhaps you should save a reference to the original SecurityManager and 
restore it when the test is finished.  That would leave the test harness 
in the same state after the Thread.security10 test as before.

David Daney.

Re: gnu/testlet/java/lang/Thread/security10.java...

[Michael Koch]

Am Dienstag, 18. Januar 2005 19:40 schrieb David Daney:
> Michael Koch wrote:
> > Am Dienstag, 18. Januar 2005 19:24 schrieb David Daney:
> >>Causes:
> >>Exception in thread "main" java.lang.SecurityException
> >>         at
> >>gnu.testlet.java.lang.Thread.security10$MySecurityManager.checkAc
> >>ce ss (security10.java:34)
> >>         at java.lang.Thread.checkAccess(Thread.java:1124)
> >>         at java.lang.Thread.setPriority(Thread.java:871)
> >>         at java.lang.Thread.init(Thread.java:293)
> >>         at java.lang.Thread.<init>(Thread.java:438)
> >>         at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
> >>         at java.security.AccessController.doPrivileged(Native
> >>Method) at
> >>java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:121
> >>) at
> >>java.lang.ref.Finalizer.runFinalization(Finalizer.java:126) at
> >>java.lang.Runtime.runFinalization0(Native Method) at
> >>java.lang.Runtime.runFinalization(Runtime.java:656) at
> >>java.lang.System.runFinalization(System.java:759) at
> >>gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:232)
> >>         at
> >>gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)
> >>
> >>When run with Sun's jre.
> >
> > which version of the JRE ? I used 1.4.2 and it passes the tests.
>
> This is running on  FC3 on i686
>
> [daney@dl2 mauve]$ java -version
> java version "1.4.2_06"
> Java(TM) 2 Runtime Environment, Standard Edition (build
> 1.4.2_06-b03) Java HotSpot(TM) Client VM (build 1.4.2_06-b03, mixed
> mode)
>
> I have no default security policy installed that I am aware of.  It
> could be an issue with security policies.

Can someone with acces to JDK 1.5 please try it ?


Michael

-- 
Homepage: http://www.worldforge.org/

Re: gnu/testlet/java/lang/Thread/security10.java...

[David Daney]

Michael Koch wrote:
> Am Dienstag, 18. Januar 2005 19:24 schrieb David Daney:
> 
> 
>>Causes:
>>Exception in thread "main" java.lang.SecurityException
>>         at
>>gnu.testlet.java.lang.Thread.security10$MySecurityManager.checkAcce
>>ss (security10.java:34)
>>         at java.lang.Thread.checkAccess(Thread.java:1124)
>>         at java.lang.Thread.setPriority(Thread.java:871)
>>         at java.lang.Thread.init(Thread.java:293)
>>         at java.lang.Thread.<init>(Thread.java:438)
>>         at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
>>         at java.security.AccessController.doPrivileged(Native
>>Method) at
>>java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:121)
>>         at
>>java.lang.ref.Finalizer.runFinalization(Finalizer.java:126) at
>>java.lang.Runtime.runFinalization0(Native Method) at
>>java.lang.Runtime.runFinalization(Runtime.java:656) at
>>java.lang.System.runFinalization(System.java:759) at
>>gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:232)
>>         at
>>gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)
>>
>>When run with Sun's jre.
> 
> 
> which version of the JRE ? I used 1.4.2 and it passes the tests.
> 

This is running on  FC3 on i686

[daney@dl2 mauve]$ java -version
java version "1.4.2_06"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_06-b03)
Java HotSpot(TM) Client VM (build 1.4.2_06-b03, mixed mode)

I have no default security policy installed that I am aware of.  It 
could be an issue with security policies.

David Daney.

Re: gnu/testlet/java/lang/Thread/security10.java...

[Michael Koch]

Am Dienstag, 18. Januar 2005 19:24 schrieb David Daney:

> Causes:
> Exception in thread "main" java.lang.SecurityException
>          at
> gnu.testlet.java.lang.Thread.security10$MySecurityManager.checkAcce
>ss (security10.java:34)
>          at java.lang.Thread.checkAccess(Thread.java:1124)
>          at java.lang.Thread.setPriority(Thread.java:871)
>          at java.lang.Thread.init(Thread.java:293)
>          at java.lang.Thread.<init>(Thread.java:438)
>          at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
>          at java.security.AccessController.doPrivileged(Native
> Method) at
> java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:121)
>          at
> java.lang.ref.Finalizer.runFinalization(Finalizer.java:126) at
> java.lang.Runtime.runFinalization0(Native Method) at
> java.lang.Runtime.runFinalization(Runtime.java:656) at
> java.lang.System.runFinalization(System.java:759) at
> gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:232)
>          at
> gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)
>
> When run with Sun's jre.

which version of the JRE ? I used 1.4.2 and it passes the tests.


Michael
-- 
Homepage: http://www.worldforge.org/

gnu/testlet/java/lang/Thread/security10.java...

[David Daney]

2005-01-13  Michael Koch  <konqueror@gmx.de>

	* gnu/testlet/java/lang/Thread/security10.java:
	New testcases.



Causes:
Exception in thread "main" java.lang.SecurityException
         at 
gnu.testlet.java.lang.Thread.security10$MySecurityManager.checkAccess 
(security10.java:34)
         at java.lang.Thread.checkAccess(Thread.java:1124)
         at java.lang.Thread.setPriority(Thread.java:871)
         at java.lang.Thread.init(Thread.java:293)
         at java.lang.Thread.<init>(Thread.java:438)
         at java.lang.ref.Finalizer$1.run(Finalizer.java:112)
         at java.security.AccessController.doPrivileged(Native Method)
         at 
java.lang.ref.Finalizer.forkSecondaryFinalizer(Finalizer.java:121)
         at java.lang.ref.Finalizer.runFinalization(Finalizer.java:126)
         at java.lang.Runtime.runFinalization0(Native Method)
         at java.lang.Runtime.runFinalization(Runtime.java:656)
         at java.lang.System.runFinalization(System.java:759)
         at 
gnu.testlet.SimpleTestHarness.runtest(SimpleTestHarness.java:232)
         at gnu.testlet.SimpleTestHarness.main(SimpleTestHarness.java:421)

When run with Sun's jre.

You install a security manager and then leave it there to interfere with 
  any tests that follow.  That seems wrong.

David Daney.