mesonbuild
Wrapdb relaunching
Now that wrapdb is back up (thanks for all who fixed it), maybe we should do some work on relaunching it and build back some trust. Maybe something like:
- write a blog post/mailing list post about how the crash went down and how it was eventually fixed
- explain about the new site reliability features such as mirroring, maybe ask for companies to donate us some virtual servers (or whatever it is the system requires) to get extra mirrors
- organize a "wrap day" to properly launch the new site
The concept of "wrap day" is just to have one day when people come together, create new wrap files, submit them to the db, review them etc. We would be there to help newbies with their questions and so on. Hopefully this would provide us with more people who want to help with both creating and rewiewing wraps.
Why is a crash or hosting a thing?
Isn't the vcpkg model 'quite good' where the index is a repo that can be git pulled, and the search is then a local operation etc? Then all this site reliability stuff is github's problem.
@jpakkane still working on actually making the proxy robust, will take a bit more time. It's basically an untested code which I started on Tim's machine. I don't think it's ready to be deployed by people outside of the core meson team. Things to do here: remove standalone mode, add test coverage for the proxy mode, add configuration for uwsgi server and create package for at least one distro, possibly docker as well.
@jmansion the idea is to avoid lock-in if we have to migrate away from github.
