On Fri, Jul 10, 2020 at 02:55:34AM -0400, Kevin A. McGrail wrote: > > A branch is a bad suggestion due to the number of testers we need and it > doesn't test ruleqa/masscheck.
Sorry but I still don't agree at all. First and most important tester is _you_. Don't commit stuff that may break things in production. I consider trunk production, since I've used it always, many people use it, and masscheck uses it. I do my utmost not to commit stuff that breaks it or backwards compatibility. Renaming things is invasive and prone to break things. Especially when done in small batches over time, which makes no sense for review purposes. I would much rather see everything done (branch hello?) and _then_ you can call for a large mass to test it, and do a controlled test in masscheck, it's not rocket science. Of course it's all the SA committers duty to also test things out, not the general public at first. > The first commit was a proof of concept for just one function (whitelist_to) > to > see what would break and how it worked in the live system. It broke worse > than > expected because I forgot a part of the commit but led to the good idea of > using welcomelist. "part of the commit"? That's downplaying it, you didn't seem to even consider leaving old eval functions in.. > Sorry but no, a branch will just hide issues. We need to find bugs. So put this issue to an actual vote and wait for all the votes, instead of sounding like a dictator. But anyway, I'm so over this.. I will not be a part of this and not reply anymore. But if you break my trunk installation, I will be angry.
