RFR: 8362243: Devkit creation for Fedora base OS is broken

[Aleksey Shipilev]

On Tue, 19 Aug 2025 05:12:13 GMT, Volkan Yazici <vyazici at openjdk.org> wrote:

>> doc/building.md line 1404:
>> 
>>> 1402: | armel        | buster       | arm           | arm-linux-gnueabi        | zero                      |
>>> 1403: | ppc          | sid          | powerpc       | powerpc-linux-gnu        | zero                      |
>>> 1404: | ppc64be      | sid          | ppc64         | powerpc64-linux-gnu      | (all)                     |
>> 
>> This one is not about Fedora, but about Debian-based sysroot. I have not checked recently, but I suspect it would still work, as Debian PPC64 BE port is still alive: https://wiki.debian.org/PPC64
>
> @shipilev, thanks so much for the review. 🙇 I was thinking in the same with @magicus while removing this row. Do you still prefer to keep it?

This list is also _not_ about Fedora devkit, and the PR is about Fedora devkit, so we are doing unrelated work here, regardless if it makes sense or not :) So yes, I prefer to keep it.

> I think all our ppc64 code assumes le by now, so even if we could get a sysroot for ppc64be to work, I doubt it will do much good for the JDK.

I believe Zero is still fine with BE. 

Anyhow, we are doing these cross-compilation instructions largely to support the poor souls who are tasked with maintaining the unusual ports. So if Debian PPC64 BE port is alive, which I think it is, that would mean they build OpenJDK, that would mean they build _some_ PPC64BE config, whether Zero (more likely) or Server (unlikely, as it is probably broken).

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/26821#discussion_r2284832460