3 ms·
There are multiple reasons for this architecture: 1. central index/indices 2. granular overlaying (you can overlay specific packages and don't have to maintai
by yrashk 16y ago
There are multiple reasons for this architecture:
1. central index/indices
2. granular overlaying (you can overlay specific packages and don't have to maintain your fork of a global repo)
3. granular permissions, allowing maintainers to work on their packages without going through the hassle of pull requests and associated bottleneck problems (the "homebrew problem")
4. usage of tags and branches for versioning primitives
5. independence from original repository
6. ability to use multiple repositories
7. (in the future) ability to use non-SCM fetching methods (aka "tarball")
That's for starters. Good enough? :)
- kunley 16y agoWell I can see your motivation and technical background. The question is: will coders bother ;) It's the people who finally decide so you'll probably have your bit of advertising/explaining to the community..
- yrashk 16y agoMost of people I talked to seem to be ok with this; also it only is important for those who decide to maintain or create their own packages — agner users don't have any hassles with it. And it's not that bad anyway :) — plus if the package author doesn't want to bother, he doesn't have to — anybody can submit a specification.
- kunley 16y agoThat's cool.