Backporting changeset from 7 to 6 for bugfix

[Andrew John Hughes]

On 10 June 2010 22:17, Joe Darcy <joe.darcy at oracle.com> wrote:
> Andrew John Hughes wrote:
>>
>> On 10 June 2010 02:02, Joe Darcy <joe.darcy at oracle.com> wrote:
>>
>>>
>>> Andrew Su wrote:
>>>
>>>>
>>>> Hello,
>>>>
>>>> This is the link[1] to the changeset which fixes the bug which renders a
>>>> circle that has a side clipped.
>>>> Requesting permission to backport this fix.
>>>>
>>>> [1] http://hg.openjdk.java.net/jdk7/jdk7/jdk/rev/fb03586d68b6
>>>>
>>>> -Andrew
>>>>
>>>>
>>>
>>> Hello Andrew.
>>>
>>> I approve this fix being applied to OpenJDK 6 b20 *as long as* the file
>>> test/sun/pisces/ScaleTest.java is added initially with an Oracle
>>> copyright
>>> rather than a Sun one; in other words, use this version of the test file:
>>>
>>> http://hg.openjdk.java.net/jdk7/jdk7/jdk/file/00cd9dc3c2b5/test/sun/pisces/ScaleTest.java
>>>
>>>
>>
>> Done.
>>
>>
>>>
>>> FYI, here are some other fixes from JDK 7 that might be applicable and
>>> worthwhile to include in OpenJDK 6:
>>>
>>> 6697721 OpenJDK: rotated text baseline different between TextLayout and
>>> drawString
>>>
>>
>> This patch updates a testcase introduced in:
>>
>> 6683472: Incorrect handling of translation component of font transform.
>>
>> and further updated by:
>>
>> 6687298: Reg testcase java/awt/Graphics2D/DrawString/RotTransText.java
>> fails on windows
>>
>> Ok to backport these as well?
>>
>
> Yes.
>

Done.

>>
>>>
>>> 6587560 OpenJDK problem handling bitmaps returned when LCD text is
>>> requested.
>>>  (may want to pull over any other subsequent fixes involving
>>> src/share/native/sun/font/freetypeScaler.c too,
>>>
>>
>> I've brought over:
>>
>> 6703377: freetype: glyph vector outline is not translated correctly
>> 6587560: OpenJDK problem handling bitmaps returned when LCD text is
>> requested
>> 6887292: memory leak in freetypeScaler.c
>>
>> The others have already been backported.
>>
>
> Good.
>>
>>
>>>
>>>   but the changeset for 6307603: "[X11] Use RENDER extension for complex
>>> operations done in software"
>>>  should be considered separately.)
>>>
>>
>> I was going to ask about this.  I think it should wait until b21
>>
>
>> It's not even in the jdk7 forest yet, and needs time to soak there
>> first.
>>
>
> Agreed on both counts.
>
>> An older version has been in IcedTea for years, so we'll update that
>> and get some testing there before including it in OpenJDK6 upstream.
>>
>>
>>>
>>> 6829678 StrokeShapeTest: createStrokedShape() behaves differently
>>>
>>
>> Backported.
>>
>>
>>>
>>> 6829673 ThinLineTest: A line < 1 pixel disappears.
>>>
>>
>> Backported.
>>
>>
>>>
>>> 6793344 BasicStroke's first element dash pattern is not a dash
>>>
>>
>> Backported.
>>
>> There are some Nimbus bugfixes we should bring over:
>>
>> changeset:   2337:042eb92f89ad
>> user:        peterz
>> date:        Thu May 06 12:57:30 2010 +0400
>> summary:     6919629: Nimbus L&F Nimbus.Overrides option leaks
>> significant amounts of memory
>>

Done.

>> changeset:   2214:325823a26aac
>> user:        peterz
>> date:        Thu Mar 18 12:02:12 2010 +0300
>> summary:     6932524: NIMBUS: 3 constructors of JSplitPane creates new
>> jsp with continuous layout - they should not.
>>

Done.

>> changeset:   2212:1224c1388e86
>> user:        rupashka
>> date:        Wed Mar 17 12:48:21 2010 +0300
>> summary:     6933784: NIMBUS: ImageView getNoImageIcon and
>> getLoadingImageIcon returns nulls instead of an icon
>>

Done.

>> changeset:   2140:fffd21bc5657
>> user:        peterz
>> date:        Fri Dec 25 17:47:08 2009 +0300
>> summary:     6860438: [Nimbus] Code to globally set slider's thumb
>> background doesn't work as specified
>>

Done.

>> changeset:   2139:e2f7e92c30f1
>> user:        peterz
>> date:        Mon Dec 21 19:26:58 2009 +0300
>> summary:     6860433: [Nimbus] Code to set a single slider's thumb
>> background doesn't work as specified
>>
>

Done.
This one relies on a new method being added to SynthLookAndFeel in:

changeset:   1999:f012e2c094ff
user:        peterz
date:        Thu Dec 10 12:08:58 2009 +0300
summary:     6827653: Make Synth UI classes public

which we can't do in OpenJDK6.
I've worked round this so that SynthLookAndFeel instead calls down to
NimbusLookAndFeel directly (rather than via the new protected method)
when it detects
an instance of it.  This has the same effect as the OpenJDK7 fix
without introducing a protected method in SynthLookAndFeel and passes
the test (pre-patched, OpenJDK6 doesn't).

> Approved.
>
>
>> and some timezone data updates:
>>
>> changeset:   2296:2571720800a6
>> user:        peytoia
>> date:        Tue Apr 20 15:01:31 2010 +0900
>> summary:     6945238: (tz) Support tzdata2010i
>>

Done.

>> changeset:   2216:f799c62ad4f8
>> user:        peytoia
>> date:        Tue Mar 30 18:35:47 2010 +0900
>> summary:     6939021: (tz) Support tzdata2010g
>>

Done.

>> changeset:   2209:0e0ce1aa1bba
>> user:        peytoia
>> date:        Thu Mar 11 11:54:17 2010 +0900
>> summary:     6933032: (tz) Support tzdata2010e
>>

Done.

>> changeset:   2208:0137b5857c63
>> parent:      2183:90248595ec35
>> user:        okutsu
>> date:        Wed Mar 10 14:32:47 2010 +0900
>> summary:     6932473: (tz) javazic produces incorrect SimpleTimeZone
>> parameters with Sun<=7
>>

Done.

>> Do these look ok?
>>
>>
>
> Yes; also approved to go back.
>
> Since some Sun copyrighted files have leaked back in, we'll done another
> pass to clean up any such files before b20 goes final.
>

Ok.  I used the Oracle copyright on those files introduced in the
patches above, rather than the original.

> Thanks,
>
> -Joe
>
>
>



-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8