PR#15: Only run postun for headless package if package is being removed (BZ 1200302) - rpms/java-1.8.0-openjdk

rpms / java-1.8.0-openjdk

#15 Only run postun for headless package if package is being removed (BZ 1200302)

Closed 5 years ago by jerboaa. Opened 5 years ago by jgu.

Unknown source master into master

Only run postun for headless package if package is being removed (BZ 1200302)

Jonathan G. Underwood • 5 years ago

2766319

java-1.8.0-openjdk.spec

file modified

+10 -5

		`@@ -368,10 +368,12 @@`


		`%define postun_headless() %{expand:`
		`- alternatives --remove java %{jrebindir -- %{?1}}/java`
		`- alternatives --remove jre_%{origin} %{_jvmdir}/%{jredir -- %{?1}}`
		`- alternatives --remove jre_%{javaver} %{_jvmdir}/%{jredir -- %{?1}}`
		`- alternatives --remove jre_%{javaver}_%{origin} %{_jvmdir}/%{jrelnk -- %{?1}}`
		`+ if [ $1 -eq 0 ] ; then`
		`+ alternatives --remove java %{jrebindir -- %{?1}}/java`
		`+ alternatives --remove jre_%{origin} %{_jvmdir}/%{jredir -- %{?1}}`
		`+ alternatives --remove jre_%{javaver} %{_jvmdir}/%{jredir -- %{?1}}`
		`+ alternatives --remove jre_%{javaver}_%{origin} %{_jvmdir}/%{jrelnk -- %{?1}}`
		`+ fi`
		`}`

		`%define posttrans_script() %{expand:`
		`@@ -961,7 +963,7 @@`

		`Name: java-%{javaver}-%{origin}`
		`Version: %{javaver}.%{updatever}.%{buildver}`
		`- Release: 7%{?dist}`
		`+ Release: 8%{?dist}`
		`# java-1.5.0-ibm from jpackage.org set Epoch to 1 for unknown reasons`
		`# and this change was brought into RHEL-4. java-1.5.0-ibm packages`
		`# also included the epoch in their virtual provides. This created a`
		`@@ -2296,6 +2298,9 @@`
		`%endif`

		`%changelog`
		`+ * Sat Aug 25 2018 Jonathan Underwood <jonathan.underwood@gmail.com> - 1:1.8.0.181.b13-8`
		`+ - Only run postun for headless package if package is being removed (BZ 1200302)`
		`+`
		`* Wed Aug 01 2018 Jiri Vanek <jvanek@redhat.com> - 1:1.8.0.181.b13-7`
		`- build number moved from release to version`

jgu commented 5 years ago

I think this should fix BZ 1200302.

rebased onto 2766319

5 years ago

jerboaa commented 5 years ago

Thanks for the PR. We need to perform some extensive testing on this before we can merge. @jvanek is the domain expert on how this works exactly.

FWIW, the build failures of simple-koji-ci are due to this bug:
https://bugzilla.redhat.com/show_bug.cgi?id=1622186

jgu commented 5 years ago

Thanks. Note that the other postun scripts should also have a $1 == 0 guarding conditional too, but I wanted to keep the first PR simple.

jvanek commented 5 years ago

Yeh. This hunk was originally here for non-paralle-installable packages. It had to be removed. Tbh I do not remember exactly why. It is some 5 years ago. I guess that one paralle installation remoed alternatives of another one. BUt this is long time ago; Both packlages and alternatvies evolved.

A lot of testing will be done with this pull request. to se eif my concerns are still valid.

Tyvm for digging those stalled waters!

jerboaa commented 5 years ago

We need to keep the remove unconditionally, since this would otherwise break upgrades of packages. See https://bugzilla.redhat.com/show_bug.cgi?id=1200302#c57. It might be possible, though to update the path or the main alternative in case $1 != 0. I'll do some experiments.