Eating our own dog food

Mads Kiilerich mads at kiilerich.com
Mon Jul 17 23:10:20 UTC 2017


On 07/14/2017 08:22 PM, Thomas De Schampheleire wrote:
> However, we need to be careful about disk space too. I guess the
> multitude of Kallithea repositories is not such a big problem, but if
> users can create other repositories we need to be more careful. I
> don't think it is the intention that OOK becomes a public repository
> host. Perhaps we can restrict each user to have one repository
> 'kallithea' and nothing else.
>
> With one kallithea-incoming, this problem does not exist.
> Mads, Andrew, what is your view?

Hosting of code from open source contributors requires some continous 
amount of work and resources. Both for establishing trust and for 
cleaning up when rogue players abuse the system anyway. But it can be 
done, assuming we want to do it and establish a realistic process and 
tooling.

Also, Kallithea is mainly used internally inside organisations. It has 
not been optimized for hosting open source projects with "random" 
contributors. I think some of the short-comings are quite obvious - we 
don't need dogfooding to find these. A commitment to making dogfooding 
work smoothly would force us to change priorities and would postpone 
other work. We or someone could do that.

Increased use of https://kallithea-scm.org/repos/ would mainly influence 
the areas Andrew currently is taking care of, so I think it mainly is up 
to him how we should do it.

/Mads


More information about the kallithea-general mailing list