* module/language/cps2/optimize.scm: New file.
* module/language/cps2/simplify.scm: New file, factored out of
simplify2.scm.
* module/language/cps/simplify2.scm: Remove, as it's obsolete.
* module/language/cps2/compile-cps.scm: Optimize the CPS.
* module/Makefile.am: Adapt for added and deleted files.
This is an intermediate step. We'll replace CPS bit by bit. If it
turns out to be a terrible idea we can just revert.
* module/Makefile.am (TREE_IL_LANG_SOURCES): Remove compile-cps.scm.
(CPS_LANG_SOURCES): Remove arities.scm.
* module/language/cps/arities.scm: Remove.
* module/language/tree-il/compile-cps.scm: Remove.
* module/language/tree-il/spec.scm: Remove use of compile-cps.scm.
* module/language/cps/compile-bytecode.scm: Remove use of arities.scm.
Instead, incoming terms are expected to call their continuations
with the correct number of arguments.
* module/language/tree-il/compile-cps2.scm (adapt-arity): New
procedure. This is equivalent to (language cps arities), but as it
is a necessary pass and not an optimization it's more proper to put
it in the converter itself. Unlike with the nested CPS
representation, it's possible to look up continuations without
making a DFG.
(convert): Adapt arities as necessary.
* lib/jit_ia64-fpu.c, lib/jit_ia64.c: Correct movi_d_w
and movi_f_w implementation to work when not using a
data buffer. This causes the check varargs.tst to
work when passing "-d" to the lightning test tool.
* module/language/cps/intset.scm (intset-add!): If the root is a
branch but isn't editable, be sure to actually update the transient
intset's root to store the writable root.
The tentative plan is to replace CPS with CPS2, and to rename CPS2 to
CPS. We will add a pass to compile tree-il to CPS2, then work from
the top down to replace the CPS compiler passes.
* module/language/cps2.scm:
* module/language/cps2/compile-cps.scm:
* module/language/cps2/renumber.scm:
* module/language/cps2/utils.scm: New files.
* module/Makefile.am: Add new files to build.
* module/language/cps/intmap.scm (intmap->alist, print-intmap)
(print-transient-intmap): New functions.
Install the new functions as printers for <intmap> and <transient-intmap>.
* lib/jit_mips-cpu.c, lib/jit_mips-fpu.c, lib/jit_mips.c:
Add base support to jit vararg functions to the mips backend.
Currently only supported on the o32 abi, until access to a
n32 system is arranged.
* include/lightning.h, include/lightning/jit_private.h,
lib/jit_names.c, lib/lightning.c: Add initial support
for the new jit_va_start, jit_va_arg, jit_va_arg_d, and
jit_va_end interfaces. The jit_va_start call is supposed
to return a va_list compatible pointer, but not yet
decided if it will be "declared" stdarg compatible,
as for now only x86 support has been added (and should
be compatible), but issues may arise on other backends.
* check/lightning.c: Add wrappers to call the new jit_va_*
interfaces.
* lib/jit_x86-cpu.c, lib/jit_x86.c: Implement the new
jit_va_* for x86.
* lib/jit_x86-sz.c: Add fields, but not yet fully updated,
as this is an intermediate commit.
* lib/jit_aarch64-sz.c, lib/jit_aarch64.c,
lib/jit_alpha-sz.c, lib/jit_alpha.c,
lib/jit_arm-sz.c, lib/jit_arm.c,
lib/jit_hppa-sz.c, lib/jit_hppa.c,
lib/jit_ia64-sz.c, lib/jit_ia64.c,
lib/jit_mips-sz.c, lib/jit_mips.c,
lib/jit_ppc-sz.c, lib/jit_ppc.c,
lib/jit_s390-sz.c, lib/jit_s390.c,
lib/jit_sparc-sz.c, lib/jit_sparc.c: Prepare for the
new jit_va_* interfaces. Not yet implemented, and will
cause an assertion if used.
* check/va_list.tst: Simple early test case, that works
on x86_64, x32, ix86, cygwin, and cygwin64.
* module/language/cps/intset.scm (make-atomic-reference)
(get-atomic-reference, set-atomic-reference!): New functions.
(*branch-size-with-edit*, *edit-index*): New constants.
(<transient-intset>): New data type.
(new-branch, clone-branch-and-set): Adapt to set edit field.
(transient-intset, persistent-intset): New exports.
(intset-add!): New interface, supporting "transient" intsets.
(intset-ref, intset-next, intset-prev, intset-fold, intset-fold2):
Work with transients.
* module/language/cps/intmap.scm (make-atomic-reference):
(get-atomic-reference, set-atomic-reference!): New helpers.
(*branch-size-with-edit*, *edit-index*): Branches now have a trailing
field, an atomic reference to their owner.
(<transient-intmap>): New record type. A mutable intmap.
(new-branch): Set the "edit" field on the branch.
(clone-branch-and-set): No editor for this field.
(assert-readable!, writable-branch): New helpers.
(transient-intmap, persistent-intmap): New exported functions.
(intmap-add!): New function.
(intmap-next, intmap-prev, intmap-ref): Work on transient intmaps.
(intmap-fold): Persist the intmap before folding over it.
* module/language/cps/intmap.scm (*branch-bits*): Switch to 32-way
branching. Marginally faster on lookup, and creation costs can be
amortized via using the transient interface. Marginal speedup for
assembler.scm compilation.
* module/language/cps/intmap.scm (meet-error): New helper.
(intmap-add, intmap-union, intmap-intersect): The "meet" argument is
optional and defaults to meet-error.
* module/language/cps.scm ($rec): Replace $letrec with $rec, which is an
expression, not a term. This means that the names bound by the letrec
appear twice: once in the $rec term, and once in the continuation.
This is not very elegant, but the situation is better than it was
before. Adapt all callers.
* doc/ref/compiler.texi (CPS in Guile): Incomplete documentation
updates. I'll update these later when the IL settles down.
* module/ice-9/eval.scm (primitive-eval): Factor out primitive=?
helper. Simplify compile-top-call. Add compile-top-branch for
primcall branches, so the compiler can see the specialized branch
operator.
* module/ice-9/eval.scm (primitive-eval): Lazily compile lambda bodies.
Special-case calls to top-level or module variables, and recognize
some of those calls as primcalls. In that case, emit closures with
the primcalls.
* module/language/cps/compile-bytecode.scm (optimize, compile-bytecode):
Marginally improve bootstrap memory usage by not retaining stale
copies of already-rewritten terms.
