Round two Re: Please review fix for 6819847 [Re: [Fwd: Problems with building jmx for OpenJDK.]]

Andrew John Hughes gnu_andrew at
Tue Mar 31 21:27:51 UTC 2009

2009/3/31 Joe Darcy <Joe.Darcy at>:
> Andrew John Hughes wrote:
>> 2009/3/31 Tim Bell <Tim.Bell at>:
>>> I wrote:
>>>> One question - if the user goes to the trouble to set any of
>>>> (See the comments in jdk/make/common/Defs.gmk starting at line 127),
>>>> should we assume they must really, really want the binary plugs and
>>>> force
>>>> IMPORT_BINARY_PLUGS=true?  Or have them set it as well?
>>>> I am thinking we shouldn't do this automatically.  I notice in the logs
>>>> that JPRT is setting these even when IMPORT_BINARY_PLUGS=false.
>>> Review feedback convinced me to reverse that last statement.
>>> Here is round two - please take a look:
>>> With these changes the default behavior is as if
>>> Unless the user explicitly sets any of the PLUGS variables:
>>> Setting these will force IMPORT_BINARY_PLUGS=true
>>> The other case is if a user outside the Sun network sets
>>> IMPORT_BINARY_PLUGS=true.  Here they will need to set one of the
>>> ALT_..._BINARY_PLUGS variables to point to the plugs anyway.
>> That suggests to me that outside Sun, there is no need for
>> IMPORT_BINARY_PLUGS at all (i.e. its value is implied by the presence
>> of one of the path variables).  However, I assume it is needed inside
>> Sun, for whatever reason.
> The only remaining need for the plugs in either OpenJDK 6 or 7 is for SNMP
> functionality.  If there was an open source replacement of that code, the
> plugs could go away altogether.
> -Joe

Are there some applications that use this?  Are there test cases?  At
present, all we have to go on for writing a replacement are the API
calls made by the code in OpenJDK.
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (

Support Free Java!
Contribute to GNU Classpath and the OpenJDK

PGP Key: 94EFD9D8 (
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8

More information about the build-dev mailing list