Sure. I'm sorry for being abrupt. I get frustrated when people show up to threads and want to talk and have their opinions heard, but they're ignoring the rest of the discussion in the thread. I'm not trying to attack you with that statement, please hear me out.swick wrote:Can you please stop telling me I'm wrong and instead tell me *why* I'm wrong? I actually really don't understand your problem. I don't understand what can't be solved by better communication and conventions.
I'm also kinda pissed that you're condescending here and don't have trust in the people lurking on this forum. I'm with this project for years now, I learned how to do open source here and while I'm not contributing regularly anymore I'm still testing the linux generic builds (probably the only one doing that).
We're having several issues with our release process, which I have outlined in this thread several times. One solution to this, which is not the solution I prefer, is to make some parts of the release process, that are absolutely irrelevant to our end users, private for a temporary period. All you are saying is "no don't do this" but not offering any counter-solution that actually works. You say we should name the RC packages RC packages: we actually already do that. So that's not a solution, because if it would work, it would already be working.
If you guys hadn't noticed, my preferred proposed solution to our issues involves absolutely no "secrecy" whatsoever. Everyone seems to be screaming and crying about this "secrecy" but no one wants to propose or even support an alternate, workable solution.
Again I am sorry for being abrupt, I am sorry for being a jerk. Please accept my apology. But please understand where I am coming from.
This is actually something I had not thought of, and might work really well. Instead of divorcing the project from the release packages, or keeping the packages under wraps, we could as a project publicly advertise the RCs and ask for testing. Then everyone knows they are RCs because we are making blog and PR posts about them being RCs, and anyone who gets confused has the entire rest of the community there to correct them.lysol wrote:I get the point with having RC:s public. OpenTTD announce their RC:s on their news page for example. It makes sense and the argument that no one will confuse an RC for a final release makes sense, and having more testers for the RC:s are great of course.
The only issue I see with this is 95% of the time there's no change between our RCs and our release packages. We don't need widespread testing on our RCs, we just need someone (who didn't package them) to give them a whirl and make sure there's not something gravely wrong with them. The extra work of publicizing the RC and handling the community engagement isn't really justified by the need for testing.
But that's why there's really no issue with having the RCs be "private" for a few days. You're not missing anything... the only thing you would miss is the 5% of the time where something silly didn't make it into that package, and it has to get rebuilt. Remember that the code for the project is still completely public: if you really want to test the release before the release packages come out, just build it yourself. And if you're too lazy to do that, then be really lazy and let someone else test the RC for you before it ends up on your computer.
We're getting completely off track here. If people really don't understand or support the whole private testing phase being added to the release process, then support Option 1, or propose an alternate solution that actually addresses our issues. Turning this into some political drama-fest isn't constructive at all...