[OpenJDK 2D-Dev] PING: JDK-8017773: OpenJDK7 returns incorrect TrueType font metrics

[Phil Race]

Hi,

It does seem like it may have been wrong to adjust the height of the
overall font metrics because of
a widened outline .. although there is likely to be less visible line
gap and more chance of clipping,
nonetheless I think this looks OK.. If we see such problems we can
re-consider.

- Please get a "+1" from some one else too ..
- Please push it to 9-client, not 9-dev. ie
http://hg.openjdk.java.net/jdk9/client/jdk

.. and thanks for the off-list ping on this.

-phil.

On 4/28/2014 7:00 AM, Yasumasa Suenaga wrote:
> Hi Phili,
>
> I've created patch for this issue and uploaded webrev:
>
> http://cr.openjdk.java.net/~ysuenaga/JDK-8017773/webrev.00/
>
>
> Could you help me?
>
> Yasumasa
>
>
> On 04/10/2014 09:29 PM, Yasumasa Suenaga wrote:
>> Hi Phili,
>> Thank you for replying.
>>
>>> I don't know the details of what is on the end of the various links you
>>> provide and
>>> I can't go grabbing code from somewhere else.
>> You can get patches as following:
>>    http://icedtea.classpath.org/bugzilla/attachment.cgi?id=873&action=diff
>>    http://icedtea.classpath.org/bugzilla/attachment.cgi?id=1014&action=diff
>>
>>
>>> So step #1 would be for you to provide a patch I can consider under
>>> terms of the
>>> OCA against hg.openjdk.java/net/jdk9/client/jdk ..
>> I've agreed the OCA, and I'm an author of jdk9 project.
>> http://openjdk.java.net/census#ysuenaga
>>
>>
>> Thanks,
>>
>> Yasumasa
>>
>>
>> On 04/10/2014 05:12 AM, Phil Race wrote:
>>> All bugs should be fixed first in the main development release and back
>>> ported if appropriate.
>>> And in any case any targeting you see is purely what the manager-types
>>> are doing
>>> because of current rules that say a bug needs to be targeted to a release.
>>> It doesn't actually mean a thing here as I currently have no time to
>>> even consider this bug.
>>> I agree with the general premise that bold fonts should only be wider,
>>> not taller but
>>> I don't know the details of what is on the end of the various links you
>>> provide and
>>> I can't go grabbing code from somewhere else.
>>>
>>> So step #1 would be for you to provide a patch I can consider under
>>> terms of the
>>> OCA against hg.openjdk.java/net/jdk9/client/jdk ..
>>>
>>> -phil.
>>>
>>>
>>> On 4/3/2014 7:13 AM, Yasumasa Suenaga wrote:
>>>> Hi all,
>>>>
>>>> Have you checked this issue?
>>>>
>>>> This issue will be fixed in JDK9.
>>>> However, I want fix it in JDK7 and JDK8.
>>>>
>>>>
>>>> Yasumasa
>>>>
>>>>
>>>> On 2014/03/17 9:02, Yasumasa Suenaga wrote:
>>>>> Hi all,
>>>>>
>>>>> The "Fix version" of this issue has been changed into 9 from 8-pool.
>>>>> Will this issue not be fixed in JDK7/8 ?
>>>>>
>>>>> I think this issue is bug of OpenJDK implementation.
>>>>> So this issue should be fixed ASAP.
>>>>>
>>>>>
>>>>> Yasumasa
>>>>>
>>>>> On 2014/02/13 16:38, Yasumasa Suenaga wrote:
>>>>>> Hi all,
>>>>>>
>>>>>> I've reported about JDK-8017773 . Have you checked it ?
>>>>>>
>>>>>> I guess that many Linux distributor will ship OpenJDK8 after JDK8 GA is released.
>>>>>> If this issue will not fix at JDK8 GA, many users who use 2D on JDK8 may encounter it.
>>>>>>
>>>>>> This issue occurs in OpenJDK only.
>>>>>> However, I think there are many users who want to use 2D on OpenJDK.
>>>>>> So I hope to fix this issue ASAP and hope to backport to JDK7.
>>>>>>
>>>>>>
>>>>>> Please cooperate.
>>>>>>
>>>>>> Yasumasa
>>>>>>
>>>>>>
>>>>>> On 2014/02/03 15:12, Yasumasa Suenaga wrote:
>>>>>>> Hi all,
>>>>>>>
>>>>>>> My customer encountered this issue.
>>>>>>> I think this issue is caused by Bug 1435 [1] and 1659 [2] in IcedTea.
>>>>>>>
>>>>>>> I read source code of "freetype-2.5.0-4.fc20" .
>>>>>>> Bold style should affect Font Glyph only. However, current implementation
>>>>>>> of OpenJDK affects Font Face. Thus value of Ascent/Descent is incorrect.
>>>>>>>
>>>>>>> OracleJDK seems to use sun.font.T2KFontScaler to calculate these values.
>>>>>>> So this issue occurs in OpenJDK only.
>>>>>>>
>>>>>>>
>>>>>>> I think we need to merge patches of Bug 1435 and 1659 in IcedTea
>>>>>>> to fix JDK-8017773 .
>>>>>>>
>>>>>>> Could you merge these patches ?
>>>>>>> Please cooperate.
>>>>>>>
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> Yasumasa
>>>>>>>
>>>>>>>
>>>>>>> [1] http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=1435
>>>>>>> [2] http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=1659
>>>>>>>
>>>>>>>
>>>>>>> ------ D E T A I L S ------
>>>>>>>
>>>>>>> I've created testcase as "JavaApplication1.java" .
>>>>>>> Result as:
>>>>>>>
>>>>>>> OracleJDK 6 Update 31
>>>>>>> ----------------
>>>>>>> [root at RHEL6-5 font]# /usr/local/jdk1.6.0_31/bin/java JavaApplication1
>>>>>>> JDK: 1.6.0_31
>>>>>>>          VM: Java HotSpot(TM) 64-Bit Server VM
>>>>>>>
>>>>>>> java.awt.Font[family=IPAMincho,name=IPAMincho,style=plain,size=50]
>>>>>>>           Ascent:  43.99414
>>>>>>>           Descent: 6.0058594
>>>>>>>
>>>>>>> java.awt.Font[family=IPAMincho,name=IPAMincho,style=bold,size=50]
>>>>>>>           Ascent:  43.99414
>>>>>>>           Descent: 6.0058594
>>>>>>> ----------------
>>>>>>>
>>>>>>> OpenJDK7 (in RHEL6.5)
>>>>>>> ----------------
>>>>>>> [root at RHEL6-5 font]# java JavaApplication1
>>>>>>> JDK: 1.7.0_45
>>>>>>>          VM: OpenJDK 64-Bit Server VM
>>>>>>>
>>>>>>> java.awt.Font[family=IPAMincho,name=IPAMincho,style=plain,size=50]
>>>>>>>           Ascent:  43.99414
>>>>>>>           Descent: 6.0058594
>>>>>>>
>>>>>>> java.awt.Font[family=IPAMincho,name=IPAMincho,style=bold,size=50]
>>>>>>>           Ascent:  45.60547
>>>>>>>           Descent: 4.3945312
>>>>>>> ----------------