Try adding retries to our full CI steps

These run a bunch at a time in parallel, and a failure in any one means having to
re-run all again which has a pretty high chance of hitting some other flaky
failure.

Author: nickva

build-aux/Jenkinsfile.full

@@ -233,12 +233,12 @@ def generateContainerStage(platform) {
                   dir( "${platform}/build" ) {
                     sh "${configure(meta[platform])}"
                     sh 'make'
-                    sh 'make eunit'
-                    sh 'make elixir'
-                    sh 'make elixir-search'
-                    sh 'make mango-test'
-                    sh 'make weatherreport-test'
-                    sh 'make nouveau-test'
+                    retry(3) {sh 'make eunit'}
+                    retry(3) {sh 'make elixir'}
+                    retry(3) {sh 'make elixir-search'}
+                    retry(3) {sh 'make mango-test'}
+                    retry(3) {sh 'make weatherreport-test'}
+                    retry(3) {sh 'make nouveau-test'}
                   }
                 }
                 catch (err) {