-Xdebug a no-op?

[Jim Holmlund]

On 8/17/2012 2:52 AM, serguei.spitsyn at oracle.com wrote:
> Staffan,
>
> If I remember correctly, the -Xdebug options was to enable JVMDI which was deprecated in JDK 5.0.
> Please, refer to this page:
> http://docs.oracle.com/javase/1.5.0/docs/tooldocs/windows/java.html
>
> The JVMDI was still available in 5.0 but was implemented as wrapper api of the JVMTI.
> I've included Jim Holmlund, the former JPDA spec lead, to keep me honest.
>
You are an honest man!
- jjh

> This is the only code in VM that still depends on the -Xdebug flag value but not sure it is still 
> valid:
>
> src/share/vm/runtime/arguments.cpp:
> 2852 jint Arguments::finalize_vm_init_args(SysClassPath* scp_p, bool scp_assembly_required) {
> 2853   // This must be done after all -D arguments have been processed.
> 2854   scp_p->expand_endorsed();
> 2855
> 2856   if (scp_assembly_required || scp_p->get_endorsed() != NULL) {
> 2857     // Assemble the bootclasspath elements into the final path.
> 2858     Arguments::set_sysclasspath(scp_p->combined_path());
> 2859   }
> 2860
> 2861   // This must be done after all arguments have been processed.
> 2862   // java_compiler() true means set to "NONE" or empty.
> 2863   if (java_compiler()&&  !xdebug_mode()) {
> 2864     // For backwards compatibility, we switch to interpreted mode if
> 2865     // -Djava.compiler="NONE" or "" is specified AND "-Xdebug" was
> 2866     // not specified.
> 2867     set_mode_flags(_int);
> 2868   }
>
>
> Thanks,
> Serguei
>
> On 8/17/12 12:59 AM, Staffan Larsen wrote:
>> Can someone confirm that -Xdebug is mostly a no-op flag right now? I assume it is still there for backwards compatibility, but it does not seem to have any impact.
>>
>> Thanks,
>> /Staffan
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/serviceability-dev/attachments/20120820/2f3146c0/attachment.html