RFR: Checking for JNI Exceptions in jni_util.c

roger riggs roger.riggs at oracle.com
Tue Feb 4 21:05:19 UTC 2014


Hi,

Updated the webrev with the recommendations.

On 2/4/2014 3:36 PM, Alan Bateman wrote:
> On 04/02/2014 18:11, roger riggs wrote:
>> Hi Alan,
>>
>> Thanks for the review and suggestions; the webrev has been updated
>> with the recommendations.
>>
>> Webrev:
>> http://cr.openjdk.java.net/~rriggs/webrev-check_exception-8030993/
> The updated webrev looks okay. I note in the JNU_ClassXXX functions 
> that the return from NewGlobalRef is not checked but I believe it's 
> safe to call DeleteLocalRef even if there is a pending exception.
>
> One minor comment is that in initializeEncoding it now assumes that 
> the reference returned by JNU_ClassString is a global ref (because it 
> obtains the reference before expanding the space for local refs). An 
> alternative (which might be clearer for future readers) is to do the 
> EnsureLocalCapacity first.
Corrected;  but I note in reading the jni specification for 
EnsureLocalCapacity
that the VM ensures that there is capacity for 16 local references before
calling the native method so checking for smaller numbers is a noop.

Roger

>
> -Alan




More information about the core-libs-dev mailing list