We have a number of add/remove rule tests, could you reproduce your problem, using the same classes and format as outlined in this test class. Ideally adding your reproducer as a single failing test method: https://github.com/droolsjbpm/drools/blob/master/drools-compiler/src/test... Creating JIRA’s and submitting pull requests. http://docs.jboss.org/drools/release/5.6.0.Final/droolsjbpm-introduction-... Mark On 6 Jun 2014, at 09:12, tia <rabarijaonadomoina(a)gmail.com&gt; wrote:

I am doing something similar, and other than a memory leak (which I have managed to track down and submit a patch for) it works pretty well... Try this... (Note that it would possibly be better to build a new ReleaseId every time, but more complicated) It has some code to clean up in case of errors, which seemed to be necessary. private static final String DRL_PATH_PREFIX = "src/main/resources"; private KieServices ks = null; private KieFileSystem kfs = null; private KieContainer kc = null; private KieBuilder kb = null; private ReleaseId snashotReleaseId = null; public boolean initRules(String commonDrl) { ks = KieServices.Factory.get(); kfs = ks.newKieFileSystem(); kb = ks.newKieBuilder(kfs).buildAll(); snapshotReleaseId = ks.getRepository().getDefaultReleasId(); // This is a snapshot release id, so it is ok if it stays the same kfs.generateAndWritePomXML(snapshotReleaseId); kfs.write(String.format(%s/common.drl", DRL_PATH_PREFIX), commonDrl); kb.buildAll(); kc = ks.newKieContainer(snapshotReleaseId); if ( kb.getResults().hasMessages( org.kie.api.builder.Message.Level.ERROR )) { for( org.kie.api.builder.Message result : kb.getResults().getMessages() ) { System.err.println("DRL error " + message.getText() + " in path: " + message.getPath() + " on line: " + message.getLine() + " column: " + message.getColumn()); } return false; } } private String getRulePath(String uniqueRuleId) { return String.format("%s/rule_%s.drl", DRL_PATH_PREFIX, uniqueRuleId); } private java.util.List<String> buildIncremental(String path) { java.util.Set<String> paths = new java.util.HashSet<String>(); paths.add(path); return buildIncremental(paths, true); } private java.util.Set<String> buildIncremental(java.util.Collection<String> paths, boolean finish) { String[] pathsArray = paths.toArray(new String[0]); java.util.Set<String> failedPaths = new java.util.HashSet<String>(); IncrementalResults results = ((InternalKieBuilder)kBuilder).createFileSet(pathsArray).build(); List<Message> addedMessages = results.getAddedMessages(); for (Message message : addedMessages) { if (Message.Level.ERROR == message.getLevel()) { System.err.println("DRL error " + message.getText() + " adding path: " + message.getPath() + " on line: " + message.getLine() + " column: " + message.getColumn()); String failedPath = message.getPath(); if (!failedPath.startsWith(DRL_PATH_PREFIX)) { failedPath = String.format("%s/%s", DRL_PATH_PREFIX, failedPath); } failedPaths.add(failedPath); if (finish) { kfs.delete(failedPath); } } } if (finish && !failedPaths.isEmpty()) { buildIncremental(paths, false); } List<Message> removedMessages = results.getRemovedMessages(); for (Message message : removedMessages) { if (Message.Level.ERROR == message.getLevel()) { System.err.println("DRL error " + message.getText() + " removing path: " + message.getPath() + " on line: " + message.getLine() + " column: " + message.getColumn()); String failedPath = message.getPath(); if (!failedPath.startsWith(DRL_PATH_PREFIX)) { failedPath = String.format("%s/%s", DRL_PATH_PREFIX, failedPath); } failedPaths.add(failedPath); } } if (finish) { kc.updateToVersion(snapshotReleaseId); } return failedPaths; } public boolean removeRule(String uniqueRuleId) { String rulePath = getRulePath(uniqueRuleId); kfs.delete(rulePath); return buildIncremental(rulePath).isEmpty(); } public boolean addRule(String uniqueRuleId, String rule){ if (ks == null) { initRules(); } String rulePath = getRulePath(uniqueRuleId); kfs.write(rulePath, rule); return buildIncremental(rulePath).isEmpty(); } On Fri, Jun 6, 2014 at 4:12 AM, tia <rabarijaonadomoina(a)gmail.com&gt; wrote:

Ok, I still don't understand why my new rule is not taken in account... So, I came back to my previous code and I realize that I have the expected results anyway. What I'm doing is: 1- FireAllRules (mandatory call to not have the NullPointerException) 2- Add a new rule 3- FireAllRules again As the session used is the same all along the process, all the expected rules to be triggered are triggered once at step 1 and step 3. So, it's ok -- View this message in context: http://drools.46999.n3.nabble.com/NullPointerException-at-org-drools-comp... Sent from the Drools: User forum mailing list archive at Nabble.com.