Fix all the issue reports first
#1
您们好,

I have a suggestion to all developers of MCS - I think that we should for the time being, instead of adding new features, make a concentrated effort to either fix or close all of the bug reports/feature requests, both on the old and new bug trackers.

In my opinion, some 150+ bug reports is a little steep, and doesn't look good for a newcomer.

Newcomer: zomg! So many bugs! *heads to Mineserver*

I'll try to follow this doctrine, art thouest behind this idea?
Reply
Thanks given by:
#2
If you're referring to those ~120 open issues on GitHub, keep in mind that those are:
- feature requests
- pull request
- true bugs

If you want a real bug count, filter by tag "bug" or "crash"; there are only about 20 such issues. Most of those are somewhat esoteric bugs, too, such as "MCA files should be padded to 4K boundaries".

I think the current stream of development is okay - some are fixing bugs, some are implementing new stuff, and it generally works.
Reply
Thanks given by:
#3
Gosh, so contraryTongue

By bugs, I mean all that issues, not just bugs, should be looked at, fixed, or closed as invalid.

For example, you could work on implementing all the 1.7 biomes (of which there is a bug report for every single one) - after all, you did say that you would be looking forwards to implementing them.

(Updated thread title)
Reply
Thanks given by:
#4
Personally I don't think the issue number is a problem. Other metrics are healthy, number of bugs is low and code churn are high. The only reason that we have so many issues is that we're about half way through a major project and accelerating. So we're getting more core users and developers but we have a deficit in issues fixed as it takes less time to learn to use the server than to get up to speed developing it.

Have a look at our commit activity graph
Reply
Thanks given by:
#5
If anyone wants to switch to mineserver, they're gonna find out soon enough that the decision was bad Smile They support 1.5.2 as the latest protocol. I think the only viable C++ alternative to MCS is hCraft, and even that has been left untouched for some 2 months.
Reply
Thanks given by:
#6
Someone can also see those issues and say: "Wow this project has many people looking at it, because there are many issues created by different people. I'll try to help and fix the issues." Smile
Reply
Thanks given by:
#7
You have a point, but there will be issues which people don't want to/can't be bothered to fix, which will end up being untouched for two years. I suppose I'll focus on what I can do now.
Reply
Thanks given by:




Users browsing this thread: 1 Guest(s)