RFR: 8027314: Java should recognize Diagnostic options if -XX:+UnlockDiagnosticVMOptions is not specified and print an informative message

[Christian Thalinger]

On Jan 15, 2014, at 11:34 AM, Calvin Cheung <calvin.cheung at oracle.com> wrote:

> On 1/15/2014 11:22 AM, Christian Thalinger wrote:
>> On Jan 15, 2014, at 10:56 AM, Calvin Cheung <calvin.cheung at oracle.com> wrote:
>> 
>>> Christian,
>>> 
>>> Thanks for your review.
>>> I've made some changes based on your suggestion.
>>> You can view the webrev at the same location: http://cr.openjdk.java.net/~ccheung/8027314/webrev/
>>> Only globals.[c|h]pp have been modified.
>>> Previous version of webrev is at: http://cr.openjdk.java.net/~ccheung/8027314/webrev.00/
>> These two webrevs seem to be the same.
> They're different. If you look at the diff of globals.hpp:
> http://cr.openjdk.java.net/~ccheung/8027314/webrev/src/share/vm/runtime/globals.hpp.cdiff.html
> 
> +   static bool is_product_build();
> 
> 
> 
> http://cr.openjdk.java.net/~ccheung/8027314/webrev.00/src/share/vm/runtime/globals.hpp.cdiff.html
> 
> +   bool is_product_build() const;

Oh.  What I meant was making is_product_build() static local method, not a class method:

+static bool is_product_build() {
+#ifdef PRODUCT
+  return true;
+#else
+  return false;
+#endif
+}

> 
> Calvin
> 
>> 
>>> Calvin
>>> 
>>> On 1/14/2014 7:01 PM, Christian Thalinger wrote:
>>>> +bool Flag::is_product_build() const {
>>>> 
>>>> This is an odd method but I can see why you have it.  Making it a static local method would be better though.
>>>> 
>>>> Otherwise this looks good.  Thanks for fixing the "Did you mean” message; one of my changes broke that one.
>>>> 
>>>> On Jan 14, 2014, at 5:22 PM, Calvin Cheung <calvin.cheung at oracle.com> wrote:
>>>> 
>>>>> This fix is for clarifying the error messages associated with vm options of type
>>>>> diagnostic, experimental, develop, and notproduct.
>>>>> 
>>>>> Summary of fix:
>>>>> - added an optional boolean (return_flag) parameter in Flag::find_flag() so that
>>>>> Arguments::process_argument() can inspect the returned flag and call
>>>>> into get_locked_message();
>>>>> - added new function is_product_build() in globals.[c|h]pp;
>>>>> - in get_locked_message(), issues the appropriate error message;
>>>>> - added a new test VMOptionWarning.java;
>>>>> - minor cleanup in the existing tests - ConfigFileWarning, CompilerConfigFileWarning.
>>>>> 
>>>>> Sample output:
>>>>> ---------------
>>>>> 
>>>>> Before fix:
>>>>> 
>>>>> C:\jdk8_i586\jdk1.8.0\bin\java -XX:+PrintInlining -version
>>>>> Improperly specified VM option 'PrintInlining'
>>>>> Error: Could not create the Java Virtual Machine.
>>>>> Error: A fatal exception has occurred. Program will exit.
>>>>> 
>>>>> After fix:
>>>>> 
>>>>> C:\jdk8_i586\jdk1.8.0\bin\java -XX:+PrintInlining -version
>>>>> Error: VM option 'PrintInlining' is diagnostic and must be enabled via -XX:+UnlockDiagnosticVMOptions.
>>>>> Error: Could not create the Java Virtual Machine.
>>>>> Error: A fatal exception has occurred. Program will exit.
>>>>> 
>>>>> Before fix:
>>>>> 
>>>>> c:\jdk9_i586\jdk1.9.0\bin\java  -XX:+TraceJNICalls -version
>>>>> Unrecognized VM option 'TraceJNICalls'
>>>>> Did you mean '(+/-)TraceJNICalls'?
>>>>> Error: Could not create the Java Virtual Machine.
>>>>> Error: A fatal exception has occurred. Program will exit.
>>>>> 
>>>>> After fix:
>>>>> 
>>>>> C:\jdk8_i586\jdk1.8.0\bin\java -XX:+TraceJNICalls -version
>>>>> Error: VM option 'TraceJNICalls' is develop and is available only in debug version of VM.
>>>>> Error: Could not create the Java Virtual Machine.
>>>>> Error: A fatal exception has occurred. Program will exit.
>>>>> 
>>>>> Testing:
>>>>>    manual test via command line and the new test VMOptionWarning.
>>>>> 
>>>>> Bug:
>>>>>    https://bugs.openjdk.java.net/browse/JDK-8027314
>>>>> 
>>>>> Webrev:
>>>>>    http://cr.openjdk.java.net/~ccheung/8027314/webrev/
>>>>> 
>>>>> thanks,
>>>>> Calvin
>