Peter, is it OK to consider your binaries under the Apache license for redistribution?
I have yet to check this but I'm happy about Eirik's report. So, just to confirm, does profiling still work with Java 8? --emi On Mon, Nov 26, 2018 at 5:42 PM Eirik Bakke <[email protected]> wrote: > > I tested the new DLL--it works!! > > Here's what I did: > 1) Start a Java profiling session via the "-agentpath" JVM argument and > invoking "Attach to External Process". Shortly thereafter the JVM crashes > with the "SUREFIRE-859: # EXCEPTION_ACCESS_VIOLATION (0xc0000005) at > pc=0x000000005f131f25, pid=10112, tid=23396" error (and a long JVM dump). > 2) Download Peter Hull's recompiled profiler binaries from > https://github.com/pedro-w/incubator-netbeans/releases/tag/v0.1-alpha , and > paste the two files (profilerinterface.dll and profilerinterface.map) into > NetBeans\profiler\lib\deployed\jdk16\windows-amd64 (overwriting the two > existing files). > 3) Repeat (1) again. No crash this time, everything works fine! > > This is great. Very happy that the profiler functionality will be working > again, for newer JDKs. Thanks, Peter! > > -- Eirik > > -----Original Message----- > From: Peter Hull <[email protected]> > Sent: Friday, November 16, 2018 5:58 AM > To: [email protected] > Subject: Re: Profiler crash for NB10 > > On Thu, Nov 15, 2018 at 6:02 AM Emilian Bold <[email protected]> wrote: > > PS: @Peter: if you could also make a recompiled .DLL available > > together with the branch it would be great. > Right, I have put changes into a fork here: > https://github.com/pedro-w/incubator-netbeans/tree/profiler-tmp > and the binaries (just the native code dll) > https://github.com/pedro-w/incubator-netbeans/releases/tag/v0.1-alpha > > I have tested it and it seems to be working on Windows 64. I've tested it > compiles on Windows 32. I have not tested or looked at any other platform > (though I don't think my changes should cause problems) > > I speculate the problem was some incorrect casting (can't cast a 64 bit > pointer to long on Windows?) which maybe didn't manifest on previous JDKs > because the values of jmethodid. > I don't know enough about the internals of the JVM but I'm still not sure if > the method used to convert a jmethodid to a jint is robust going forward. > > Please let me know what you think, > Pete > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > For further information about the NetBeans mailing lists, visit: > https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
