| Summary: | dev-java/swt-* fails with dev-java/openjdk-11 due to libjawt.so search paths | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Mike Auty (RETIRED) <ikelos> |
| Component: | Current packages | Assignee: | Java team <java> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | dschridde+gentoobugs, gyakovlev |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | add a new path | ||
yeah, -11 changed installation paths a bit.
openjdk-11/lib is always lib, no matter of what arch/subarch etc.
also there is not x86 openjdk:11, but there is one for arm.
I guess we need another condition in swt, which will cover openjdk-bin:11 and openjdk:11.
basically [[ -f "${JAVA_HOME}/lib//${JAWTSO}" ]]
should do for all :11 slots.
(In reply to Georgy Yakovlev from comment #1) > I guess we need another condition in swt, which will cover openjdk-bin:11 > and openjdk:11. > > > basically [[ -f "${JAVA_HOME}/lib//${JAWTSO}" ]] > should do for all :11 slots. Are there plans to implement this? Is something holding it back? dev-java/swt-4.10 still exhibits this behavior. Any plan to implement this? Created attachment 633438 [details, diff]
add a new path
as suggested. The patch is basic, can you check and push it please?
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=f239e95ecddb44dc379f9c12be10b0a36e2dd0c1 commit f239e95ecddb44dc379f9c12be10b0a36e2dd0c1 Author: Georgy Yakovlev <gyakovlev@gentoo.org> AuthorDate: 2020-04-25 09:34:55 +0000 Commit: Georgy Yakovlev <gyakovlev@gentoo.org> CommitDate: 2020-04-25 09:35:17 +0000 dev-java/swt: add openjdk11 libjawt.so path Closes: https://bugs.gentoo.org/689976 Package-Manager: Portage-2.3.99, Repoman-2.3.22 Signed-off-by: Georgy Yakovlev <gyakovlev@gentoo.org> dev-java/swt/swt-4.10.ebuild | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) |
I know that openjdk-11 is still masked for use as a gentoo-vm, but I've been testing it and haven't run into many problems at all, in fact, I think this is the first one. In swt (all versions) there is the following check: if [[ -f "${JAVA_HOME}/jre/lib/${AWT_ARCH}/${JAWTSO}" ]]; then export AWT_LIB_PATH="${JAVA_HOME}/jre/lib/${AWT_ARCH}" elif [[ -f "${JAVA_HOME}/jre/bin/${JAWTSO}" ]]; then export AWT_LIB_PATH="${JAVA_HOME}/jre/bin" elif [[ -f "${JAVA_HOME}/$(get_libdir)/${JAWTSO}" ]] ; then export AWT_LIB_PATH="${JAVA_HOME}/$(get_libdir)" else eerror "${JAWTSO} not found in the JDK being used for compilation!" Openjdk-11 doesn't seem to install a jre directory (I don't know if that's different to openjdk-bin-11, or something else, either way), so the first two are out. The third option: [[ -f "${JAVA_HOME}/$(get_libdir)/${JAWTSO}" ]] uses get_libdir, which on an amd64 system returns lib64, however, openjdk-11 installs directly under /usr/lib64/openjdk-11/lib/libjawt.so. I have symlinked lib64 to lib under /usr/lib64/openjdk-11/ and that builds fine, what I'm not sure of it which package needs to change to accommodate the build (hence CCing both maintainers)? Should openjdk actually store its files under lib64, or should the swt's get_libdir be changed? Given that openjdk-bin-8 installs the following: /usr/lib64/openjdk-8/jre/lib/amd64/libjawt.so /usr/lib64/openjdk-8/lib/amd64/libjawt.so it isn't clear that get_libdir would work any better there? Either way, I thought I'd start the discussion/process of getting this resolved... 5:)