On Tue, Apr 23, 2013 at 9:30 PM, Jorge <xxopxe@gmail.com> wrote:
Yep, there is a flurry of traffic on diegonehabs github, but it would be great to have some insight on what is going to happen in the future ref. luasocket.
It's such a crucial piece of the Lua infrastructure, true. Part of the trouble can be expressed by rephrasing the late great Grace Hopper's ironical quote about standards: "modern open source software is so great because there are so many forks to choose from".
For instance, I have a fork of LuaSocket embedded in the LuaBuild project that used Paul K's fork plus a fix for IPV6-crippled XP, works fine on everything I've tried on. By 'fine' I mean 'passes tests' - except for the one about setting the REUSEADDR flag. Tests are only as good as their coverage, however, and the LuaSocket tests basically just exercise basic TCP/IP functionality. No UDP or the derived modules like socket.http.
It feels like a little coordination would be useful here...
