srowen commented on a change in pull request #26636: [SPARK-27721][BUILD]
Switch to use right leveldbjni according to the platforms
URL: https://github.com/apache/spark/pull/26636#discussion_r349977256
##########
File path: pom.xml
##########
@@ -3081,5 +3094,17 @@
<profile>
<id>sparkr</id>
</profile>
+ <profile>
+ <id>aarch64</id>
+ <properties>
+ <leveldbjni.group>org.openlabtesting.leveldbjni</leveldbjni.group>
+ </properties>
+ <activation>
+ <os>
+ <family>linux</family>
+ <arch>aarch64</arch>
Review comment:
Actually, you've had prompt responses repeatedly:
https://mail-archives.apache.org/mod_mbox/spark-dev/201911.mbox/%3CCAEccTyxQ31BVP3vty0h_VxE_%3DfE7OWUg%3DqpumuZuK5kDys21JQ%40mail.gmail.com%3E
I won't reply a fourth time, but here, you're suggesting there is still
something that needs to be done to make aarch64 work. Why would anyone announce
it works then? No, the project is not going to take on this goal for you; you
are best placed to do this work. As before, you are getting prompt attention
for your efforts.
I do not expect that Spark will ever release a separate aarch64 build. So,
the ideal solution is to make the upstream version compatible with all
architectures. I do not know what the main leveldbjni project is or isn't going
to support; it's what I'm asking you. For now, relying on a fork under a
profile could be OK. Are there any other differences in the fork?
Yes, you do not want to activate based on the build machine's architecture.
It should be specified manually.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
