Apricot should do this, just like Rubinius saves the bytecode it generates from a file so it doesn't have to recompile again until the file is changed. Most of the start up time in Apricot is in compiling the core library, so it could be a lot faster if this was cached.
Note that Apricot's bytecode caching will be slightly more complicated than Rubinius' because in Apricot if file A uses a macro from file B then A must be recompiled also if B changes.
Apricot should do this, just like Rubinius saves the bytecode it generates from a file so it doesn't have to recompile again until the file is changed. Most of the start up time in Apricot is in compiling the core library, so it could be a lot faster if this was cached.
Note that Apricot's bytecode caching will be slightly more complicated than Rubinius' because in Apricot if file
A
uses a macro from fileB
thenA
must be recompiled also ifB
changes.