[8u] RFR bad source/target in Makefile of J2DBench

Jiri Vanek jvanek at redhat.com
Tue Nov 9 09:41:48 UTC 2021


hmm. Webrev messed up.
Now am using the one from https://hg.openjdk.java.net/code-tools/webrev/raw-file/tip/webrev.ksh

And output is better - the  patch itself is same, but the whole history of files is correctly omitted.
updated:  http://cr.openjdk.java.net/~jvanek/bad_j2d_bytecodeversion/webrev/

The  J2DBench  is not included in the main build cycle. You have to:

$ pushd  jdk/src/share/demo/java2d/J2DBench/
$ make
$ java -jar dist/J2DBench.jar

It starts the starter gui.
 From that you can load options/default.opt and press start
Without gui,  you can do:
$ java -jar dist/J2DBench.jar -loadopts options/default.opt  -batch -title review -description review  -saveres review.raw

which runs 1 hour.

https://bugs.openjdk.java.net/browse/JDK-8042199 amended as you sugested


Thanx for patience!
   J.


On 11/8/21 18:07, Jiri Vanek wrote:
> Hello!
> 
> Metadata updated
> http://cr.openjdk.java.net/~jvanek/bad_j2d_bytecodeversion/webrev/
> 
> 
> I'mnot a commiter, so you aor Andrew will need commit on my behalf. TY!
> 
> J.
> 
> On 11/5/21 21:38, Hohensee, Paul wrote:
>> Still has the CMMTests.java issue. I can see the change in the new version of the file but not the webrev, so try "webrev -b". The actual patch will need to include the original commit metadata. You can get that using "hg export --git -r 
>> <commit-id>".
> 
> ^ cleared.
>>
>> The hg-updater daemon will create the backport issue when the backport is pushed, so you don't have to do it yourself.
>>
>> -----Original Message-----
>> From: Jiri Vanek <jvanek at redhat.com>
>> Date: Friday, November 5, 2021 at 8:15 AM
>> To: "Hohensee, Paul" <hohensee at amazon.com>, Andrew Hughes <gnu.andrew at redhat.com>
>> Cc: "jdk8u-dev at openjdk.java.net" <jdk8u-dev at openjdk.java.net>
>> Subject: RE: [8u] RFR bad source/target in Makefile of J2DBench
>>
>> http://cr.openjdk.java.net/~jvanek/bad_j2d_bytecodeversion/webrev/
>>
>> updated. Now is 1:1 with original.
>>
>> TY!
>>
>> On 11/5/21 15:46, Jiri Vanek wrote:
>>>
>>> On 11/4/21 21:47, Hohensee, Paul wrote:
>>>> The use of -source 1.5 -target 1.5 in J2DBench/Makefile doesn't correspond to the comment in the README, which says that 1.6 is the minimum due to jdk9's javac. The original jdk9 patch used -source 1.6 and -target 1.6, so I'd do the same
>>>> and also leave the README comment alone.  A reference to a later JDK is fine in a backport.
>>>
>>> Ok. Will do.
>>>
>>>>
>>>> The patch to CMMTests.java is missing the spelling update to the definition and use of usePlatfromProfiles and usePlatfrom. To be sure, it's a nop, but should be done to keep any future backport(s) consistent.
>>>
>>> How that happened. Will double check.
>>>
>>>
>>>>
>>>> Otherwise lgtm.
>>>>
>>>> Thanks,
>>>> Paul
>>>>
>>>> -----Original Message-----
>>>> From: jdk8u-dev <jdk8u-dev-retn at openjdk.java.net> on behalf of Jiri Vanek <jvanek at redhat.com>
>>>> Date: Thursday, November 4, 2021 at 12:29 PM
>>>> To: Andrew Hughes <gnu.andrew at redhat.com>
>>>> Cc: "jdk8u-dev at openjdk.java.net" <jdk8u-dev at openjdk.java.net>
>>>> Subject: Re: [8u] RFR bad source/target in Makefile of J2DBench
>>>>
>>>> ping please!
>>>>
>>>> On 10/25/21 20:14, Jiri Vanek wrote:
>>>>> Thanx!
>>>>>
>>>>> webrev adapted: http://cr.openjdk.java.net/~jvanek/bad_j2d_bytecodeversion/webrev/
>>>>>
>>>>> Compared to original, I had dropepd the build.xml change, as it was not relevant.
>>>>>
>>>>> The https://bugs.openjdk.java.net/browse/JDK-8042199 is already closed.  Should it be reopened, or new bug will be created?
>>>>>
>>>>> Thanx!
>>>>>     J.
>>>>>
>>>>>
>>>>>
>>>>> On 10/25/21 15:37, Andrew Hughes wrote:
>>>>>> On 18:56 Thu 21 Oct     , Jiri Vanek wrote:
>>>>>>> http://cr.openjdk.java.net/~jvanek/bad_j2d_bytecodeversion/
>>>>>>>
>>>>>>> Hello! the build.xml (https://hg.openjdk.java.net/jdk8u/jdk8u/jdk/file/d2bbff5b7006/src/share/demo/java2d/J2DBench/build.xml#l52) is corerect,
>>>>>>> and have -source/target correctly on 1.5
>>>>>>>
>>>>>>> However Mkaefile was left behind, and is on 1.2, which jdk8 itself can not play.
>>>>>>>
>>>>>>>
>>>>>>> Am looking forward for kind soul to create a bug a commit on my behalf.
>>>>>>>
>>>>>>> Happy hacking,
>>>>>>>        J.
>>>>>>> -- 
>>>>>>> Jiri Vanek Mgr.
>>>>>>> Principal QA Software Engineer
>>>>>>> Red Hat Inc.
>>>>>>> +420 775 39 01 09
>>>>>>>
>>>>>>
>>>>>> This looks like a subset of:
>>>>>>
>>>>>> https://bugs.openjdk.java.net/browse/JDK-8042199
>>>>>>
>>>>>> I think it would be preferable to backport this to keep this code aligned
>>>>>> across JDKs. It fixes some other differences between the Makefile &
>>>>>> build.xml, as well as actually making it buildable with 1.4, as the
>>>>>> README claims.
>>>>>>
>>>>>> The -g:none addition seemed strange to me, but it seems debug=false
>>>>>> has been in build.xml since the start of the OpenJDK project [0].
>>>>>> If we want to remove that, we should probably do so in HEAD and
>>>>>> backport for consistency across releases.
>>>>>>
>>>>>> [0] https://mail.openjdk.java.net/pipermail/2d-dev/2014-August/004736.html
>>>>>>
>>>>>
>>>>
>>>> -- 
>>>> Jiri Vanek Mgr.
>>>> Principal QA Software Engineer
>>>> Red Hat Inc.
>>>> +420 775 39 01 09
>>>>
>>>>
>>>
>>
>> -- 
>> Jiri Vanek Mgr.
>> Principal QA Software Engineer
>> Red Hat Inc.
>> +420 775 39 01 09
>>
>>
> 

-- 
Jiri Vanek Mgr.
Principal QA Software Engineer
Red Hat Inc.
+420 775 39 01 09



More information about the jdk8u-dev mailing list