Below is a copy my exchange with Jeffrey, now I think we should go with option 2.
Xavier ---------- Forwarded message ---------- From: Xavier Hanin <[EMAIL PROTECTED]> Date: Apr 26, 2007 4:01 PM Subject: Re: IVY-475 To: jeff <[EMAIL PROTECTED]> On 4/26/07, jeff <[EMAIL PROTECTED]> wrote:
hi xavier, the reason why i made this haltonunresolved is because how i wrote the code, it only halts on unresolved, and does not halt on a general failure. i did this because i though the user should have the choice of NOT failing on non-resolve (seems like a valid use case), but still failing on a general error. however, i did not add the haltonfailure feature in the patch.
The point is that unresolved dependencies are considered as a general error, so if you halt on failure you will halt on unresolved too. There is no other example of NOT halting on unresolved whlie halting on failure. So I think we will implement haltonfailure only for the moment, and if you want an halt on unresolved feature, create another issue, and we'll see how we can address it consistently in both resolve and install tasks. Xavier
i would suggest leaving haltonunresolved, and also adding haltonfailure, which would default to true. Xavier Hanin <[EMAIL PROTECTED]> wrote: Hi Jeffrey, I'm not sure if you follow the ivy-dev mailing list or not, but we'd need your opinion about IVY-475. Could you check this thread: http://www.nabble.com/Re%3A-svn-commit%3A-r532027---in--incubator-ivy-core-trunk%3A-CHANGES.txt-doc-doc-use-install.html-src-java-org-apache-ivy-ant-IvyInstall.java-test-java-org-apache-ivy-ant-IvyInstallTest.java-tf3641359.html Thanks, Xavier
