RFR(XS) 8020791: [TESTBUG] runtime/jsig/Test8017498.sh failed to compile (original) (raw)
David Holmes david.holmes at oracle.com
Thu Jul 18 18:53:58 PDT 2013
- Previous message: RFR(XS) 8020791: [TESTBUG] runtime/jsig/Test8017498.sh failed to compile
- Next message: RFR(XS) 8020791: [TESTBUG] runtime/jsig/Test8017498.sh failed to compile
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi Calvin,
On 19/07/2013 10:12 AM, Calvin Cheung wrote:
Please review this small fix for a testcase.
webrev: http://cr.openjdk.java.net/~ccheung/8020791/webrev/ Bug: https://jbs.oracle.com/bugs/browse/JDK-8020791 http://bugs.sun.com/viewbug.do?bugid=8020791 When the testcase was developed, it was using an incomplete jnimd.h without the following: #if defined(SOLARIS) || defined(LINUX) || defined(ALLBSDSOURCE) The default typedef long jlong will be used during compilation of the native code of the testcase on linuxx64. With the above #if defined in the jnimd.h, if "LINUX" isn't defined, the following typedef will be used and causing the compile error. _typedef int64 jlong; The fix is to define LINUX (-DLINUX) in the gcc command line.
It seems to me that assuming there is a gcc in the path on the test machine is not very reliable. We have 4 tests that make that assumption.
That aside the changes seem valid.
David
The change in TestJNI.c is just a minor cleanup.
Test: Ran jtreg on the testcase. thanks, Calvin
- Previous message: RFR(XS) 8020791: [TESTBUG] runtime/jsig/Test8017498.sh failed to compile
- Next message: RFR(XS) 8020791: [TESTBUG] runtime/jsig/Test8017498.sh failed to compile
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]