RFR 8168091: jlink should check security permission early when programmatic access is used
Sundararajan Athijegannathan
sundararajan.athijegannathan at oracle.com
Tue Oct 18 05:26:17 UTC 2016
Hi,
Thanks for your review. I've updated webrev with jdk.tools.jlink ->
jdk.tools.jlink.internal refactoring. But, when I attempted to get rid
of shell script in the test with your suggestion, I got:
Exception in thread "main" java.security.AccessControlException: access denied ("java.io.FilePermission" "/Users/SATHIJEG/src/jdk9-dev/jdk/test/tools/jlink/JTwork/tools/jlink/JLinkToolProviderTest.d/main.0.jta" "read")
at java.security.AccessControlContext.checkPermission(java.base at 9-internal/AccessControlContext.java:471)
at java.security.AccessController.checkPermission(java.base at 9-internal/AccessController.java:894)
at java.lang.SecurityManager.checkPermission(java.base at 9-internal/SecurityManager.java:548)
at java.lang.SecurityManager.checkRead(java.base at 9-internal/SecurityManager.java:887)
at java.io.FileInputStream.<init>(java.base at 9-internal/FileInputStream.java:127)
at java.io.FileInputStream.<init>(java.base at 9-internal/FileInputStream.java:93)
at java.io.FileReader.<init>(java.base at 9-internal/FileReader.java:58)
at com.sun.javatest.regtest.agent.MainWrapper.main(MainWrapper.java:46)
Looks like I've to give AllPermission to all code in jtreg itself and
leave the test only as sandbox! => I've to have use a complicated policy
file. Shell script avoids all that..
Updated webrev: http://cr.openjdk.java.net/~sundar/8168091/webrev.01/
Thanks,
-Sundar
On 18/10/16, 3:33 AM, Mandy Chung wrote:
>> On Oct 17, 2016, at 10:23 AM, Sundararajan Athijegannathan<sundararajan.athijegannathan at oracle.com> wrote:
>>
>> Please review http://cr.openjdk.java.net/~sundar/8168091/webrev.00/ for
>> https://bugs.openjdk.java.net/browse/JDK-8168091
> The shell test can be removed and use
> @run main/othervm/secure=java.lang.SecurityManager
>
> You may want to move the classes in jdk.tools.jlink package to jdk.tools.jlink.internal since they are now internal.
>
> Mandy
More information about the jigsaw-dev
mailing list