anything can be shadowed by closer nested variables
fixing bugs in let-syntax and expanding optional arg default values
improving expansion algorithm on internal define
some small optimizations to the compiler
maintaining interpreter for bootstrapping
making branch destinations relative so we can access the code
using *ip instead of code[ip]
doing the same to bp (argument access)
small improvement to map, builtin->instruction
better profiling
always fast and gives conservative correct answers, at the expense of
fidelity on medium and large size structures. for example (hash (iota 15))
gives the same result as (hash (iota 14)).
accepting r6rs IEEE literals +-nan.0 and +-inf.0
printing distinguished -0.0, indicating float with .0f instead
of #float, double with .0 instead of #double
more renaming (? on predicates, ! on mutating operators)
changing T to #t :( all those #s are so ugly
porting code to sort out which NILs are false and which are
empty lists
switching to scheme-style special forms. however you feel about
scheme names vs. CL names, using both is silly.
mostly switching to scheme predicate names, with compatibility
aliases for now. adding set-constant! to make this efficient.
adding null?, eqv?, assq, assv, assoc, memq, memv, member
adding 2-argument form of if
allowing else as final cond condition
looking for init file in same directory as executable, so flisp
can be started from anywhere
renaming T to FL_T, since exporting a 1-character symbol is
not very nice
adding opaque type boilerplate example file
adding correctness checking for the pattern-lambda benchmark
bugfix in int2str
that could access the current environment dynamically. It also
differed unnecessarily from common lisp set in this respect.
So now setq is a builtin special form that sets lexical or
global variables, and set is a function that sets global variables.
Rather than eliminate the power of the dynamic set, I extended it
by adding eval*, which evaluates its argument in the current
environment. The justification for this is that the interpreter
is already dynamic enough to allow it with no overhead, so the
ability might as well be exposed.
cleanup; removing some magic numbers
beginning hash tables
JeffBezanson