On 19 October 2014 17:51, Tanu Kaskinen <tanu.kaskinen at linux.intel.com> wrote: > Hi, > > In D?sseldorf we discussed a bit about the plan regarding the routing > patches. The situation remains that nobody else than David has looked into > the patches, and David would rather not merge the patches, because the > infrastructure alone doesn't solve any problems. The point was repeated that > it's hard to review the infrastructure if it's not visible how it will be > used. > > The idea of rewriting the infrastructure as a module was brought up. Having > the code in a module was considered to be easier to accept in upstream, and > my impression was that everyone thought that would be a good idea. It > occurred to me later that we didn't actually discuss it in very concrete > terms how a module-based approach would work better, so let's do it now. > > The routing patches that have been written so far are divided into several > sets. If we imagine that I would have originally put the infrastructure in a > module, would the patch sets have been accepted to the master branch without > waiting for more code? Reviewing would have been equally hard, because the > final picture wouldn't have been available in this case either. Would I have > got the permission to push the code without any review, or with > lower-quality review? I guess so, but I'd like to get a confirmation for > that before I start from scratch again. Yes, I think the idea would be to get broad review and not nitpick so much on the concepts which were a concern for David/me to merge into the core since at least I found it hard to evaluate it without more context, usage examples, etc. Regards, Arun
