openjfx-11.0.7_p1.ebuild has a typo in the java VM detection code, appears twice: if has_version --host-root dev-java/openjdk:${SLOT}; then export JAVA_HOME=${EPREFIX}/usr/$(get_libdir)/openjdk-${SLOT} export JDK_HOME="${JAVA_HOME}" export ANT_RESPECT_JAVA_HOME=ture ^^^^ else if [[ ${MERGE_TYPE} != "binary" ]]; then JDK_HOME=$(best_version --host-root dev-java/openjdk-bin:${SLOT}) [[ -n ${JDK_HOME} ]] || die "Build VM not found!" JDK_HOME=${JDK_HOME#*/} JDK_HOME=${EPREFIX}/opt/${JDK_HOME%-r*} export JDK_HOME export JAVA_HOME="${JDK_HOME}" export ANT_RESPECT_JAVA_HOME=ture ^^^^ fi fi It looks like this was first introduced in openjfx-11.0.3_p1.ebuild, and then carried forward in subsequent openjfx-11.0.6_p2.ebuild, openjfx-11.0.7_p0.ebuild, and openjfx-11.0.7_p1.ebuild I can't find the code that actually reads the ANT_RESPECT_JAVA_HOME variable; maybe it is simply testing if the variable is set and so interprets any value as truth. This has then been copy-pasted into other Java ebuilds in bugzilla (but none that have hit the tree yet as far as I can tell), such as work on an updated icedtea-web package (#715316).
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=b5f10f9e818ff21a58476933fd4c60906a7a1d78 commit b5f10f9e818ff21a58476933fd4c60906a7a1d78 Author: Georgy Yakovlev <gyakovlev@gentoo.org> AuthorDate: 2020-08-30 22:58:32 +0000 Commit: Georgy Yakovlev <gyakovlev@gentoo.org> CommitDate: 2020-08-30 22:58:52 +0000 dev-java/openjfx: fix typo Closes: https://bugs.gentoo.org/728790 Package-Manager: Portage-3.0.4, Repoman-3.0.1 Signed-off-by: Georgy Yakovlev <gyakovlev@gentoo.org> dev-java/openjfx/openjfx-11.0.9_p0.ebuild | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
yeah according to my testing it checks the presence. for some reason I've fixed it already, just noticed it was not the case. Maybe I git stashed it during updates and missed. thanks for catching.