<AWT Dev>  Gcc 8.1 HarfBuzz library compilation issues
philip.race at oracle.com
Tue Nov 20 15:04:10 UTC 2018
I have removed awt-dev and replaced it with the CORRECT 2d-dev list.
Please reply to this one .. to get awt-dev out of the loop.
I already have an open bug to upgrade harfbuzz I will be handling soon.
If the necessary changes are there we are all good.
We do not much like at all locally modifying sources we get from upstream.
It is a maintenance nightmare.
On 11/20/18, 6:48 AM, Simon Tooke wrote:
> On 2018-11-20 5:31 a.m., Magnus Ihse Bursie wrote:
>> On 2018-11-19 22:22, Simon Tooke wrote:
>>> I've been looking at compiling the JDK with GCC 8.1, and trying to fix
>>> issues upstream as I find them.
>> Nice! If you find issues in the JDK source code per se, please let us
>> know. :) Sooner or later, we'll have to support gcc 8 properly anyway,
>> might just as well try to clear the path already.
> I noticed yesterday that
> https://bugs.openjdk.java.net/browse/JDK-8213153 from Dmitry Chuyko is
> also working on GCC 8. I am not sure how far down the rabbit hole he
> is, but I (with the kind assistance of Severin Gehwolf since I'm not a
> committer) am submitting a series of patches to move towards that. I
> have a JDK tree compiling with 8.1, but it still has to disable warnings
> in more places than I feel comfortable with.
>>> For example the version of HarfBuzz shipped in the JDK has some issues
>>> with template specialization and with clearing a struct via memset()
>>> (thus bypassing C++ member access restrictions).
>>> I've had changes to address these issues accepted upstream , and I
>>> was wondering if it is appropriate for me to file an issue to backport
>>> those fixes (I have a patch ready) or do you prefer to wait until
>>> HarfBuzz is upgraded to the latest version during the natural course of
>>> JDK development.
>>> (Be aware that the JDK is on a 1.7 version, while these fixes are in a
>>> 2.X version).
>>>  https://github.com/harfbuzz/harfbuzz/pull/1334
More information about the build-dev