sand
is for creating interpreters, like the Python interpreter and Haskell interpreter.
It can also be used for creating text based games and CLI test environments.
For examples, check out the examples folder.
sand
implements a concurrent model. It views an interpreter as two seperate components:
the User Interface, sand.UI
, and the Command Processor,sand.Engine
. The following
diagram shows how under the hood sand
operates. Every square is a goroutine.
+--------+ +--------------------------+
| | +-------------> Engines Manager +--------------+
| Read <----------+ | +--------------------------+ |
| | | | |
+----+---+ | | |
| +-+---+------+ +-------v------+
| | | | | +----------+
+------------> UI | | Engine | | Engine |
| (usually +----------------------------------------> Runner +---->+ Exec |
+------------> main) | | | | |
| | | XXXXXXXXXXXXXXXXXXXXXXXXXXXX | | +----------+
| +-+----------+ X Manager connects UI X +--------------+
+----+---+ | X to Engine Runner X
| | | XXXXXXXXXXXXXXXXXXXXXXXXXXXX
| Write <----------+
| |
+--------+
sand.UI
is a struct
that is provided for you and is implemented as broad as possible;
however there are few features missing, which are commonly found in popular interpreters,
namely: Line history and Auto-completion. These features may be added later, but as for
now they are not planned for.
sand.Engine
is an interface
, which must be implemented by the user. Implementations
of sand.Engine
must have a comparable underlying type, see Go Spec
for comparable types in Go.