RFR (S): 8024087: Remove dead JVM_{Get, Set}PrimitiveFieldValues functions
Srinivas Ramakrishna
ysr1729 at gmail.com
Thu Oct 3 13:01:57 PDT 2013
The changes look good to me too.
-- ramki (ysr)
On Tue, Oct 1, 2013 at 5:22 PM, Mikael Vidstedt
<mikael.vidstedt at oracle.com> wrote:
>
> Coleen/Christian/Chris/Dmitry - Thanks for the reviews.
>
> Since this is only relevant on pre-1.4 JDKs I actually really think nobody
> depends on this.
>
> I noticed that I needed one small additional change in nativeLookup.cpp -
> the struct is iterated over in the function just below
> (lookup_special_native) and the start index was previously initialized to 2
> for jdk1.4+, skipping the first two entries in the array. With that change
> in place everything works as expected.
>
> I uploaded a webrev with that change included for the curious:
>
> http://cr.openjdk.java.net/~mikael/webrevs/8024087/webrev.01/webrev/
>
> Cheers,
> Mikael
>
>
> On 2013-09-30 18:27, Christian Thalinger wrote:
>>
>> SAP is still using this, I suppose, but they will maintain their own
>> version. I applaud old code removal. Looks good.
>>
>> On Sep 30, 2013, at 5:26 PM, Mikael Vidstedt <mikael.vidstedt at oracle.com>
>> wrote:
>>
>>> Folks,
>>>
>>> Please review:
>>>
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8024087
>>> Webrev:
>>> http://cr.openjdk.java.net/~mikael/webrevs/8024087/webrev.00/webrev/
>>>
>>>
>>> Summary:
>>>
>>> This change removes two JVM_* functions - JVM_GetPrimitiveFieldValues and
>>> JVM_SetPrimitiveFieldValues. These functions were used back in JDK 1.3.1 to
>>> intrinsify two java.io.Object{Input,Output}Stream methods, but the Java code
>>> was rewritten in the JDK 1.4 time frame and the corresponding Java methods
>>> are no longer there, meaning this code is well and truly dead and has been
>>> so for many years.
>>>
>>> On top of that the functions only have prototypes declared in
>>> jvm_misc.hpp (as opposed to jvm.h etc), meaning the risk of somebody
>>> depending on these functions is further limited. My non-extensive
>>> (re-)search of the web appears to confirm this.
>>>
>>> Thanks,
>>> Mikael
>>>
>
More information about the hotspot-dev
mailing list