[OpenJDK 2D-Dev] [PATCH] Re: JDK-8012351

Sergey Bylokhov Sergey.Bylokhov at oracle.com
Tue Feb 11 00:02:09 UTC 2014


Hi, Phil.
The fix looks good.

On 11.02.2014 2:48, Phil Race wrote:
> On 2/10/2014 2:14 PM, Sergey Bylokhov wrote:
>> Hi, Phil.
>>  - Should we catch situations, when the getParentFile will return 
>> null, and change 'return newDir.equals(existDir)' to 'return 
>> Objects.equals(newDir, existDir)'?
>
> I can .. http://cr.openjdk.java.net/~prr/8012351.1/
> although for this to be necessary "newFont" would have to
> be stored in the "/" directory which seems very unlikely
>>  - I suppose it was a typo ">=" in SunFontManager?
>
> No, it was intended but overly cautious as it didn't give the FontFamily
> instance a chance to decide for itself.
>
> -phil.
>
>>
>> Otherwise the fix looks good, but note that I'm not  an expert in 
>> this area.
>>
>> On 05.02.2014 21:53, Phil Race wrote:
>>> So that patch wasn't the right thing but as a result of off-list
>>> discussion its become clearer how the problem might occur
>>> I've updated https://bugs.openjdk.java.net/browse/JDK-8012351
>>> with the evaluation.
>>>
>>> And I think this is the right fix for it: 
>>> http://cr.openjdk.java.net/~prr/8012351/
>>>
>>> Please review.
>>>
>>> Randy says it works for him. If anyone else is seeing this problem
>>> it would be nice to know if this fixes it for them too - or not !
>>> I never reproduced the problem myself except by 'forcing' Ubuntu
>>> Regular to be ignored in fontconfiguration.
>>>
>>> -phil.
>>>
>>>
>>>
>>> On 01/21/2014 09:34 AM, Ryan Tandy wrote:
>>>> Dear JDK team,
>>>>
>>>> Further to my earlier message, please find attached the patch that I
>>>> think solves JDK-8012351. I applied this change to the Ubuntu JDK7
>>>> package and made it available at the Launchpad bug for testing; so far
>>>> one other user commented that it works for them.
>>>>
>>>> Can JDK-8012351 be re-opened and this change considered as part of a
>>>> possible fix?
>>>>
>>>
>>
>>
>


-- 
Best regards, Sergey.




More information about the 2d-dev mailing list