1. 9

  2. 2

    Sounds roughly like what seemed to be the consensus when there was a small Lua/luarocks meetup at FOSDEM a few years ago, (where at least @catwell and @etiene where also present iirc), and a few more people much more involved in Lua than I am - a good portion of users lament that there is no (big) standard library and the other part of the users (back then, including me) said they simply don’t need it[0]. Lua isn’t in widespread use in the way Ruby, Python, or PHP is, and the publicly available ecosystem is just much smaller.

    Also there’s the split of users between 5.1/5.2/5.3 and LuaJIT/RaptorJIT - when you’re embedding it in your app you probably want even more stability for your users, because they don’t care for new features, they want to script YOUR app. (No idea which version of Lua World of Warcraft includes, for example, and if they ever updated since 2004, or if they maintained their own fork, and how much it differs, it’s highly restricted for the users anyway).

    [0]: The company I was working for was building a product where latency for a HTTP request was the most important thing, let’s say ~5ms response time on a LAN roughly. The choices at the time (2013ish) were mostly C++ or LuaJIT if you wanted to at least stay in the realm of non-obscure languages. (No Java because of the GC, and it was maybe 1-2 years too early for Go, C might have worked but meh.)