This library is a high level interface between Rust and Lua. Its major goal is to expose as easy to use, practical, and flexible of an API between Rust and Lua as possible, while also being completely safe.
There are other high level Lua bindings systems for rust, and this crate is an exploration of a different part of the design space. The other high level interface to Lua that I am aware of right now is hlua which you should definitely check out and use if it suits your needs. This crate has the following differences with hlua:
The key difference here is that rlua handles rust-side references to Lua values
in a fundamentally different way than hlua, more similar to other Lua bindings
systems like Selene for C++. Values like
rlua::Table
and rlua::Function
that hold onto Lua values in the Rust stack,
instead of pointing at values in the Lua stack, are placed into the registry
with luaL_ref. In this way, it is possible to have an arbitrary number of
handles to internal Lua values at any time, created and destroyed in arbitrary
order. This approach IS slightly slower than the approach that hlua takes of
only manipulating the Lua stack, but this, combined with internal mutability,
allows for a much more flexible API.
There are currently a few notable missing pieces of this API:
_ENV
upvalue of a loaded chunk to a table other than _G
, so that you can
have different environments for different loaded chunks.Additionally, there are ways I would like to change this API, once support lands in rustc. For example:
It is also worth it to list some non-goals for the project:
This library is very much Work In Progress, so there is a some API churn. Currently, it follows a pre-1.0 semver, so all API changes should be accompanied by 0.x version bumps.
The goal of this library is complete safety, it should not be possible to cause undefined behavior whatsoever with the API, even in edge cases. There is, however, QUITE a lot of unsafe code in this crate, and I would call the current safety level of the crate "Work In Progress". Still, UB is considered the most serious kind of bug, so if you find the ability to cause UB with this API at all, please file a bug report.
There are, however, a few ways to cause panics and even aborts with this API that are not currently considered bugs. Usually these panics or aborts are alternatives to what would otherwise be unsafety. A near term goal of this project is to remove the ability for lua to cause a panic or abort, and then panic / abort behavior will be considered a bug just like UB is, but that is currently not the case.
Panic / abort considerations when using this API:
Lua
instance should continue to be usable.gcc
crate (the
default), LUA_USE_APICHECK
is enabled. Any abort caused by this internal
Lua API checking should be considered a bug, particularly because without
LUA_USE_APICHECK
it would generally be unsafe.__gc
metamethod specified in lua scripts will abort if the metamethod causes an
error rather than longjmp like a normal error would. Lua objects can also
be resurrected with user provided __gc
metamethods
(See here for details),
and this includes userdata, so it is possible to trigger a panic from lua by
resurrecting a userdata and re-using it after it has been garbage collected.
It is an eventual goal of the library to ensure that lua scripts cannot
cause panics or aborts, but currently this is not true and this is a known
limitation. Lua scripts should NOT be able to cause unsafety, though, this
is always considered a bug.LUA_USE_APICHECK
abort, from exceeding LUAI_MAXCCALLS. This may be a source of unsafety if
LUA_USE_APICHECK
is disabled, and is considered a bug.rlua
are FnMut, so if you trigger your own callback
recursively, currently this will panic.rlua::Variadic
. I believe
this would be unsafe without LUA_USE_APICHECK
and should be considered a
bug.