Created attachment 469132 [details] stacktrace I'll present you with the solution / request first -- the explanation for how I arrived there is given below. dev-libs/nss should probably be sub-slotted, so that dev-java/icedtea can be rebuild if dev-libs/nss is being updated: "The issue is that NSS has been updated since you last built IcedTea. If you re-emerge IcedTea, the problem should be fixed." --- Andrew John Hughes <gnu.andrew@redhat.com> The full report, as reported upstream, is replicated here: When I run http://repo1.maven.org/maven2/io/swagger/swagger-codegen-cli/2.2.0/swagger-codegen-cli-2.2.0.jar to fetch from https://raw.githubusercontent.com/... I get attached stacktrace. The full command-line is: # java -jar swagger-codegen-cli-2.2.0.jar generate -i https://raw.githubusercontent.com/kubernetes/kubernetes/v1.5.5/api/swagger-spec/v1.json -l clojure --additional-properties 'projectName=kubernetes,projectVersion=1.5.5-0.0.1,projectDescription="Clojure API bindings for Kubernetes",projectUrl=https://github.com/devurandom/clj-kubernetes,projectLicenseName="Apache License, Version 2.0",projectLicenseUrl=http://www.apache.org/licenses/LICENSE-2.0.html' The full script I am executing is available at: https://github.com/devurandom/clj-kubernetes/blob/master/update.sh I am using IcedTea 3.3.0: # java -version openjdk version "1.8.0_121" OpenJDK Runtime Environment (IcedTea 3.3.0) (Gentoo icedtea-3.3.0) OpenJDK 64-Bit Server VM (build 25.121-b13, mixed mode) See-Also: http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=3344 See-Also: https://bugs.gentoo.org/show_bug.cgi?id=605430
Created attachment 469134 [details] emerge --info icedtea nss
I've already said in bug #605430 that subslotting NSS isn't an option. Think of the bigger picture. This will inconvenience almost everybody just for the few that build IcedTea from source. It only affects IcedTea because of the weird way it uses the library and NSS upstream have already said they will try and get IcedTea to do it the right way. They also said that the first breakage that happened in 3.28 was unintended and it only broke again in 3.29 because they put it back the way it was before. They will try not to break it again before a formal ABI change when a subslot would be necessary.
(In reply to James Le Cuirot from comment #2) Thanks for the explanation. From #605430 it was not entirely clear what exactly is wrong and how it shall be fixed.
