Julian Foad <julianf...@btopenworld.com> wrote:
>What holds me back from committing it is lack of a test that at least >exercises the 'keep-going' functionality (that is, reports on another >revision after reporting a failed revision) and lack of evidence that >it has been tested and found to work as expected in a few different >failure scenarios. I tried some quick tests by hand with a previous >version of the patch and found some wrong behaviour, and haven't tested >the latest version at all. Mainly I am unwilling to commit myself to >ensuring that necessary testing and follow-ups do in fact happen. > Thanks Julian. I have now handled the errors that can occur while verifying-one-revision in a single place, which would evade falling into wrong behaviour. I am still writing test cases to test a few possible scenarios. >The other issues I raised are not blockers for me and could be >addressed in follow-up commits. I wouldn't veto you or anyone else >committing it in its current state with follow-ups afterwards. > >Prabhu is welcome to work on this on a branch from my point of view, if >that would help. > Sure, that would be more comfortable to work on the feature. >Prabhu, you might also find it helpful to ask for help if you're having >difficulty with some part. For example, if you find it hard to write >tests for it, it's quite possible that someone here would be willing to >help or even do that for you. (In fact it's good engineering practice >for someone else to write the tests.) Don't be shy of asking. > Sure Julian.. Regards Prabhu -- Sent from my Android phone with K-9 Mail. Please excuse my brevity.
