RFR: 8351493: Revert build changes that are no longer needed [v2]
Magnus Ihse Bursie
ihse at openjdk.org
Thu Mar 20 22:20:28 UTC 2025
On Mon, 10 Mar 2025 13:38:30 GMT, Magnus Ihse Bursie <ihse at openjdk.org> wrote:
>> I compared the hermetic-java-runtime branch with mainline, and noticed some old build changes lingering, that are no longer needed. Furthermore, I checked a handful of other files and noticed a few spurious changes that just make the branch differ from mainline for no good reason. I also reverted those.
>>
>> With this patch, the difference between hermetic-java-runtime and mainline gets even more minimal.
>
> Magnus Ihse Bursie has updated the pull request incrementally with one additional commit since the last revision:
>
> Also remove added extra line
I don't really see how this differs from the dynamic case. The distributor building the JDK must decide whether to use the system libraries (typically done by Linux distributions, since they have control over such things), or bundle it (most other). The same goes for when building a static library. Even if we allow some kind of "mix and match" with jlink, there needs to be some kind of limit on how much a single version of the JDK can provide for. In one extreme, mixing different address sizes or CPU architectures is not possible, and will never be. On the other hand, libjsig has been provided as an optional library. So somewhere between there will libharfbuzz need to fall. My instinct is that it is not like libjsig but rather like the OS; that is, for a specific instance of the JDK a distributor builds, they decide not only OS and CPU, but also if libraries like harfbuzz should be included or not, both for static and dynamic.
-------------
PR Comment: https://git.openjdk.org/leyden/pull/47#issuecomment-2741789229
More information about the leyden-dev
mailing list