[OpenJDK 2D-Dev] RFR: 8231556: Wrong font ligatures used when 2 versions of same font used

[Philip Race]

 > probably this check can be added to the current method
 > in PhysicalFont and removed from the CFont?

Addressing this point, it *could* but if that was the best solution
I would have done it back then. It is a macos fix in macos code which
seems best to me.

-phil.

On 12/7/19, 4:39 PM, Philip Race wrote:
>
> I am not sure. I don't remember doing the follow on fix mentioned there
> to make the full name of the components of logical fonts on macos have
> different names. And since at least two of the logical fonts on macos are
> in .ttc files then they'd have the same file name .. so I think that 
> over-ride
> is still needed.
>
> -phil.
>
> On 12/7/19, 4:00 PM, Sergey Bylokhov wrote:
>> Hi, Phil.
>>
>> A few years ago in JDK-8139176 we added a check for the "style" to the
>> CFont.equals(), probably this check can be added to the current method
>> in PhysicalFont and removed from the CFont?
>>
>> On 12/7/19 11:59 am, Philip Race wrote:
>>> The equals() method was refactored a bit to make it easier to add
>>> the extra tests, and some of the complexity is due to paranoia that
>>> platName *might* be null .. although I don't think that should ever 
>>> be the case.
>>
>> I think code can be simplified if Objects.equals() will be used
>> Don't we need to update the hashcode as well?:
>>     return Objects.hash(fullName, platName);
>>
>>