Page MenuHomeFreeBSD

java/openjdk{25,26}: Don't req xorg for headless
ClosedPublic

Authored by haraldei on Tue, Aug 25, 9:39 AM.
Tags
None
Referenced Files
F169739523: D59167.id185017.diff
Wed, Sep 2, 10:33 AM
F169676855: D59167.id185152.diff
Wed, Sep 2, 6:38 AM
F169600036: D59167.id185152.diff
Wed, Sep 2, 1:02 AM
F169587068: D59167.id185192.diff
Wed, Sep 2, 12:06 AM
Unknown Object (File)
Tue, Sep 1, 4:39 PM
Unknown Object (File)
Tue, Sep 1, 10:52 AM
Unknown Object (File)
Tue, Sep 1, 7:14 AM
Unknown Object (File)
Mon, Aug 31, 6:56 AM

Details

Summary

Turn xorg and related libs into BUILD_DEPENDS to avoid requiring them
when installing from packages.

PR: 294358
Reported by: raivo@lehma.com

Diff Detail

Repository
R11 FreeBSD ports repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

haraldei created this revision.

Mainly just a check that I got the field in the commit message right, but feel free to have a look to see if the patches looks ok too.

java/openjdk25/Makefile
4

typo

71–75

Spaces here should be tabs.
Also, I'm unclear on why these BUILD_DEPENDS are necessary. I did a poudriere testport with them removed, and the build succeeded.
https://pkg.ftfl.ca/data/16amd64-default/2026-08-25_11h23m47s/logs/openjdk25-headless-25.0.4+7.f.1.log

java/openjdk26/Makefile
71–75

Same BUILD_DEPENDS comments as above.

Address review comments.

Kept the BUILD_DEPENDS lines while waiting for feedback from original author.

Removed BUILD_DEPENDS lines no longer needed after 25.0.4 and 26.0.2, and
updates commit message to reflect this.

haraldei marked an inline comment as done.
This revision is now accepted and ready to land.Fri, Aug 28, 7:16 PM

Is there any specific reason for checking flavor twice? Putting USES + USE_XORG into same conditional would be nicer. Or just leaving USE_XORG as is because it's no-op without USES=... xorg ....

Is there any specific reason for checking flavor twice? Putting USES + USE_XORG into same conditional would be nicer. Or just leaving USE_XORG as is because it's no-op without USES=... xorg ....

I think the checks were just added around where the USES and USE_XORG lines already were placed.

Portclippy complains if I put the USES and USE_XORG lines together, but moving the additions to RUN_DEPENDS into the same conditional as USES causes no complains. I'll clean up some more portclippy issues, and test a bit more before submitting a new patch for review.