Fix for 8030115 breaks windows build
David Holmes
david.holmes at oracle.com
Tue Mar 10 03:51:26 UTC 2015
Talk about history repeating itself, lessons not learned etc. :(
The original fix was broken when pushed to 9 and subsequently fixed
again under:
https://bugs.openjdk.java.net/browse/JDK-8053902
Hence we just need a backport of JDK-8053902. I'll send the approval
request immediately.
David
On 10/03/2015 1:40 PM, David Holmes wrote:
> The fix for 8030115 breaks the build on windows:
>
> c:/jprt/T/P1/021115.drlong/s/jdk/src/share/native/sun/tracing/dtrace/JVM.c(200)
> : error C2143: syntax error : missing ';' before 'type'
> c:/jprt/T/P1/021115.drlong/s/jdk/src/share/native/sun/tracing/dtrace/JVM.c(201)
> : error C2065: 'count' : undeclared identifier
> c:/jprt/T/P1/021115.drlong/s/jdk/src/share/native/sun/tracing/dtrace/JVM.c(201)
> : error C2065: 'count' : undeclared identifier
> c:/jprt/T/P1/021115.drlong/s/jdk/src/share/native/sun/tracing/dtrace/JVM.c(202)
> : error C2065: 'count' : undeclared identifier
> c:/jprt/T/P1/021115.drlong/s/jdk/src/share/native/sun/tracing/dtrace/JVM.c(204)
> : error C2065: 'count' : undeclared identifier
> c:/jprt/T/P1/021115.drlong/s/jdk/src/share/native/sun/tracing/dtrace/JVM.c(212)
> : error C2065: 'count' : undeclared identifier
> make[2]: ***
> [/cygdrive/c/jprt/T/P1/021115.drlong/s/build/windows-x86_64-normal-server-fastdebug/jdk/objs/libjsdt/JVM.obj]
> Error 2
>
> The older MSC compiler doesn't support declarations in the middle of
> blocks IIRC so barfs on this:
>
> @@ -195,16 +197,24 @@
> jvm_providers = (JVM_DTraceProvider*)calloc(
> num_providers, sizeof(*jvm_providers));
>
> - for (i = 0; i < num_providers; ++i) {
> - JVM_DTraceProvider* p = &(jvm_providers[i]);
> + int count = 0;
> + for (; count < num_providers; ++count) {
>
>
> Filed:
>
> https://bugs.openjdk.java.net/browse/JDK-8074795
>
> Not sure how this has escaped notice till now but the nightly build has
> been failing on Windows since this was pushed! :(
>
> I'm testing the simple fix and will send out the RFR/A as soon as it is
> confirmed.
>
> David
>
More information about the jdk8u-dev
mailing list