Am 11/17/16 um 06:00 schrieb Igor Fedorenko: > FWIW, I ran maven integration tests (commit 73a2b7) against current > maven master and got this > >> Tests run: 771, Failures: 3, Errors: 21, Skipped: 0 > > Running the same test checkout with Maven 3.3.9 > >> Tests run: 771, Failures: 0, Errors: 0, Skipped: 0 >
I am currently looking into this. Whenever I commit to maven master, I start the 'maven-3.3-release-status' job on Jenkins and make sure, the ITs will not fail there. I did not receive an email from Jenkins and last time I started that job, the ITs ran successfully. I just started that job again, it has not caught up with recent commits so far. Let's see what will happen. There is no reason for anyone to apologize for anything. This is email. I am not sitting here with an angry face or something. It was more an "oh fuck it. a few weeks ago postponed for 3.5, branch created, work started to update JIRA issues, etc. and now it will not go into 3.x and the next major version will not need any of those issues solved because the new model solves all of them. what a waste of time." There is one thing I think needs to be outlined. Jason, I know nothing about the users you are talking about but I understand what is driving your decisions. I have not worked on issues of any user I know of personally. Every issue I have worked on has been reported by maven users in JIRA. Maybe some of them are your users. I don't know. There is no priority of my, your or anyone elses users here. If none of your users report issues in JIRA, then why ignore the users doing that and helping us by doing it. With "I did not change a thing" I meant "just wait until master is in a state it will be released in". There are ways to ship those changes without breaking backwards compatibility without the need to update model versions or something like this. I am currently working on making the one behavioural change not backwards compatible controllable by using a POM property 'maven.model.dependencyManagementImportInheritanceProcessing' (true/false). It's done that way in other areas. Similar to this [1] but without changing the maven model. POM property instead of model attribute. Would you be against doing that as well? If the anser is yes, then at least the Springframework guys and theire users will be disappointed when not shipping MNG-5971 for model version 4.0.0 in some backwards compatible way. [1] <https://git-wip-us.apache.org/repos/asf?p=maven.git;a=commit;h=4dde52072559d34feafb72f3784c265aacb14c8e> --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org