See the PR, a subset is enough.
The problem with all classes comes from the fact that HPPC uses a
template mechanism (like NIO Buffer classes in JDK) to generate the
source code for all those combinations. We need to only fork all classes
we use. Actually we do not fork classes part of the HPP
> On 27 May 2024, at 09:08, Chris Hegarty
> wrote:
>
>> ...
>
> That sounds like quite a lot of classes. How much is actually necessary to
> allow to remove the dependency? And/Or is there a place of natural place
> where it makes logical sense to subset?
Please ignore this comment. I see
Hi,
> +1 to moving the hppc fork to oal.internal.
+1
> On 26 May 2024, at 13:33, Bruno Roustant wrote:
>
> Currently the hppc fork in Lucene is composed of 15 classes and 8 test
> classes.
> Forking everything in hppc would mean 525 classes and 193 test classes. I'm
> not sure we want to for
Thanks Bruno - didn't want to step in the way but I can work on the PR now
(is this fine)?
D.
On Mon, May 27, 2024 at 9:03 AM Bruno Roustant
wrote:
> Ah, I started to work on this. So I just sent an incomplete PR[1] to share
> with you Dawid, so you don't do the work twice. Feel free to modify
Ah, I started to work on this. So I just sent an incomplete PR[1] to share
with you Dawid, so you don't do the work twice. Feel free to modify the PR
if needed.
[1] https://github.com/apache/lucene/pull/13422
Le lun. 27 mai 2024 à 08:40, Dawid Weiss a écrit :
>
> Hi Mike,
>
> These changes are
