https://bugzilla.redhat.com/show_bug.cgi?id=1557371 --- Comment #26 from Jie Kang <jkang@xxxxxxxxxx> --- (In reply to jiri vanek from comment #24) > > Fully versioned dependency in subpackages if applicable. > Note: No Requires: %{name}%{?_isa} = %{version}-%{release} in > > Unluckily, we have it mostly opposite. java-openjdk requires > %{name}-headless%{?1}%{?_isa} = %{epoch}:%{version}-%{release} (?1 is > nothing xor debug) and rest is usually transitive. Main package have: > Requires: %{name}-headless%{?1}%{?_isa} = %{epoch}:%{version}-%{release} > correctly > > java-openjdk-headless - requires nothing, as rest requires it > java-openjdk-devel - Requires: %{name}%{?1}%{?_isa} = > %{epoch}:%{version}-%{release} ?1 confuses rpmlint? > java-openjdk-jmods - Requires: %{name}-devel%{?1} = > %{epoch}:%{version}-%{release > java-openjdk-demo , Requires: %{name}%{?1}%{?_isa} = > %{epoch}:%{version}-%{release} > java-openjdk-src ,Requires: %{name}-headless%{?1}%{?_isa} = > %{epoch}:%{version}-%{release} > java-openjdk-javadoc-zip , hmm.. no requires, no provides... hmmm Issue? > java-openjdk-accessibility , Requires: %{name}%{?1}%{?_isa} = > %{epoch}:%{version}-%{release (agaiin, ?1 confused it?) > java-openjdk-javadoc , no requirement. imho it deserves to be included > without jre itself. > > If non-debug are good, then those are also good, as the macro is generating > them. In addition, thoise will never be discovered by rpmlint, as those > depends on java-openjdk-whatever-debug version (in same rationale as normla > ones). And (imho) have no reason te depend on nonrmal ones. > java-openjdk-debug > java-openjdk-headless-debug , > java-openjdk-devel-debug > java-openjdk-jmods-debug , > java-openjdk-demo-debug > java-openjdk-src-debug , > java-openjdk-javadoc-debug , > java-openjdk-javadoc-zip-debug , > java-openjdk-accessibility-debug , > > java-openjdk-debugsource - What is this? > java-openjdk-debuginfo , And what is this? > > > So imho the only real issue (unlessomebody tells me whta are those last two) > is java-openjdk-javadoc-zip. > Imho should be fixed. No requires - same reasoning as javadoc, but some > virtual provides shouldbe there. > So taken from javadoc: > > %define java_javadoc_rpo() %{expand: > OrderWithRequires: %{name}-headless%{?1}%{?_isa} = > %{epoch}:%{version}-%{release} > # Post requires alternatives to install javadoc alternative > Requires(post): %{_sbindir}/alternatives > # in version 1.7 and higher for --family switch > Requires(post): chkconfig >= 1.7 > # Postun requires alternatives to uninstall javadoc alternative > Requires(postun): %{_sbindir}/alternatives > # in version 1.7 and higher for --family switch > Requires(postun): chkconfig >= 1.7 > > # Standard JPackage javadoc provides > Provides: java-javadoc-zip%{?1} = %{epoch}:%{version}-%{release} > Provides: java-%{javaver}-javadoc-zip%{?1} = %{epoch}:%{version}-%{release} > Provides: java-%{javaver}-%{origin}-javadoc-zip = > %{epoch}:%{version}-%{release} > } > > Including also alternatives, as spec contains > alternatives --install %{_javadocdir}/java-zip javadoczip .... lines > > > What do you think? Fix for javadoc-zip sounds okay to me. Is zip format for javadoc used by many people? Just curious. debuginfo and debugsource package info: https://fedoraproject.org/wiki/Changes/SubpackageAndSourceDebuginfo I'm not sure if these are supposed to apply for openjdk package. Do you know? -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component _______________________________________________ package-review mailing list -- package-review@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to package-review-leave@xxxxxxxxxxxxxxxxxxxxxxx
