* libguile/Makefile.am (INSTANTIATE): New variable.
(install-data-hook): Use it.
* libguile/libguile-2.2-gdb.scm: Autoload (system vm debug).
Augment %load-path and %load-compiled-path, and reload (system base
types).
* module/system/base/types.scm: Remove #:hide to be 2.0-compatible.
Use (system syntax internal) conditionally when on 2.2.
* libguile/throw.h (scm_ithrow, scm_throw): Mark as SCM_NORETURN.
* libguile/throw.c (scm_throw, scm_ithrow): Adapt to not return.
* libguile/vm-engine.c (throw, throw/value, throw/value+data): New
instructions.
* libguile/vm.c (vm_throw, vm_throw_with_value)
(vm_throw_with_value_and_data): New helpers.
* module/language/cps/compile-bytecode.scm (compile-function): Add cases
for new instructions.
* module/language/cps/prune-bailouts.scm (prune-bailouts): More simple,
now that there are no $kreceives in play.
* module/language/cps/reify-primitives.scm (reify-clause): Update
reification of no-clause functions to use new throw op.
* module/language/tree-il/compile-cps.scm (convert): Convert invocations
of the variable-arity 'throw primitive from Tree-IL to the new
fixed-arity CPS instructions.
* module/system/vm/assembler.scm (emit-throw/value*)
(emit-throw/value+data*, emit-throw): Export new instructions.
* module/system/vm/disassembler.scm (code-annotation): Add annotation.
* libguile/vm-engine.c (lsh, rsh, lsh/immediate, rsh/immediate): New
instructions taking unboxed bit counts.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/effects-analysis.scm:
* module/language/cps/specialize-numbers.scm (specialize-f64-unop):
(specialize-u64-unop): Add ability to specialize add/immediate, etc,
and add lsh/immediate as well.
(specialize-u64-binop, specialize-u64-shift): Move rsh/lsh
specialization to its own procedure, given that the bit count is
already unboxed.
(specialize-operations): Adapt to support more /immediate
instructions.
* module/language/cps/type-fold.scm (mul): Reify an lsh/immediate
instead of an ash.
* module/language/cps/types.scm (compute-ash-range): Add type inferrers
for lsh, rsh, and their immediate variants.
* module/system/vm/assembler.scm: Export emit-lsh and so on.
* module/language/tree-il/compile-cps.scm (convert): Convert "ash" on
immediates to rsh/immediate or lsh/immediate.
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Don't restrict /imm params to encodeable immediates; specialize any
imm. Rely on reify-primitives to undo the transformation if needed.
* module/system/base/target.scm (target-max-size-t):
(target-max-size-t/scm, target-max-vector-length): New public
functions.
* module/language/cps/types.scm (type-entry-saturating-union): Remove
restriction of polymorphic types to be within max-size-t; this could
incorrectly apply constraints on numeric values.
(&max/size, &max/scm-size): Use target-max-size-t.
(*max-size-t*): Remove; replace uses with (target-max-size-t).
* module/language/cps/optimize.scm (optimize-first-order-cps): Move up
the simplify-primcalls pass, to allow DCE to take away unneeded
constants. An incremental step
* module/language/cps/reify-primitives.scm (reify-primitives): Add pass
to re-reify constant arguments for primcalls with immediate parameters
that can't be encoded as bytecode.
* module/language/cps/compile-bytecode.scm (compile-function): Update
add/immediate, etc.
* module/language/cps/slot-allocation.scm (compute-needs-slot):
Simplify.
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Rework for add/immediate, etc.
* module/language/cps/types.scm (define-unary-result!)
(define-binary-result!): Take types as params instead of variables, so
we can share this code with /imm variants.
(add/immediate, sub/immediate, uadd/immediate, usub/immediate)
(umul/immediate, ulsh/immediate, ursh/immediate): Update type
inferrers.
* module/language/cps/closure-conversion.scm (convert-one):
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/effects-analysis.scm (define-primitive-effects*)
(expression-effects, primitive-effects): Only fall back to passing
constant table if the immediate parameter is false. Adapt closure
effects analysis.
* module/language/cps/slot-allocation.scm (compute-needs-slot): Remove
special cases for free-ref/free-set!.
* module/language/cps/compile-bytecode.scm (compile-function): Make
load-f64, load-s64, and load-u64 take an immediate parameter instead
of a CPS value.
* module/language/cps/effects-analysis.scm: Remove CPS argument from
immediate load instructions.
* module/language/cps/slot-allocation.scm (compute-needs-slot): Remove
special case for load-64 etc.
* module/language/cps/specialize-numbers.scm
(specialize-u64-scm-comparison): Adapt.
* module/language/cps/specialize-primcalls.scm (specialize-primcalls):
Adapt.
* module/language/cps/types.scm (define-type-inferrer*): Also take param
argument.
(define-type-inferrer, define-predicate-inferrer): Adapt.
(define-type-inferrer/param): New helper.
(load-f64, load-s64, load-u64): Adapt inferrers to pass on value from
param.
* module/language/cps/utils.scm (compute-constant-values): Adapt.
* module/language/cps.scm ($primcall): Add "param" member, which will be
a constant parameter to the primcall. The idea is that constants used
by primcalls as immediates don't need to participate in optimizations
in any way -- they should not participate in CSE, have the same
lifetime as the primcall so not part of DCE either, and don't need
slot allocation. Indirecting them through a named $const binding is
complication for no benefit. This change should eventually improve
compilation time and memory usage, once we fully take advantage of it,
as the number of labels and variables will go down.
* module/language/cps/closure-conversion.scm:
* module/language/cps/compile-bytecode.scm:
* module/language/cps/constructors.scm:
* module/language/cps/contification.scm:
* module/language/cps/cse.scm:
* module/language/cps/dce.scm:
* module/language/cps/effects-analysis.scm:
* module/language/cps/elide-values.scm:
* module/language/cps/handle-interrupts.scm:
* module/language/cps/licm.scm:
* module/language/cps/peel-loops.scm:
* module/language/cps/prune-bailouts.scm:
* module/language/cps/prune-top-level-scopes.scm:
* module/language/cps/reify-primitives.scm:
* module/language/cps/renumber.scm:
* module/language/cps/rotate-loops.scm:
* module/language/cps/self-references.scm:
* module/language/cps/simplify.scm:
* module/language/cps/slot-allocation.scm:
* module/language/cps/specialize-numbers.scm:
* module/language/cps/specialize-primcalls.scm:
* module/language/cps/split-rec.scm:
* module/language/cps/type-checks.scm:
* module/language/cps/type-fold.scm:
* module/language/cps/types.scm:
* module/language/cps/utils.scm:
* module/language/cps/verify.scm:
* module/language/tree-il/compile-cps.scm: Adapt all users.
* libguile/weak-table.c (scm_t_weak_table); Add last_gc_no member.
* libguile/weak-table.c (vacuum_weak_table): Only vacuum if we haven't
done so since the last GC.
(scm_c_weak_table_ref, scm_c_weak_table_put_x, scm_c_weak_table_remove_x)
(scm_c_weak_table_fold): Vacuum the weak table if needed.
(scm_weak_table_clear_x): Update last_gc_no flag, as no more vacuuming
will be needed.
This change should make weak tables work better with libgc, as the weak
components that need mark functions are smaller, so they don't overflow
the mark queue. Also this prevents the need to move disappearing
links.
* libguile/weak-table.c (scm_t_weak_entry): Change to be a hash table
chain entry.
(struct weak_entry_data, do_read_weak_entry, read_weak_entry): Read
out the key and value directly.
(GC_move_disappearing_link, move_disappearing_links, move_weak_entry):
Remove.
(scm_t_weak_table): Rename "entries" member to "buckets", and "size" to
"n_buckets".
(hash_to_index, entry_distance, rob_from_rich, give_to_poor): Remove.
(mark_weak_key_entry, mark_weak_value_entry): Mark a single link, and
the next link.
(mark_doubly_weak_entry): New kind.
(allocate_entry): Allocate a single entry.
(add_entry): New helper.
(resize_table): Reimplement more like normal hash tables.
(vacuum_weak_table): Adapt to new implementation.
(weak_table_ref, weak_table_put_x, weak_table_remove_x): Adapt.
(make_weak_table): Adapt.
(scm_weak_table_clear_x): Actually unregister the links to prevent a
memory leak.
(scm_c_weak_table_fold): Collect items in an alist, then fold outside
the lock.
(scm_weak_table_prehistory): Initialize doubly_weak_gc_kind.
* module/ice-9/arrays.scm (array-print-prefix): New private function.
* libguile/arrays.c (scm_i_print_array): Reuse (array-print-prefix) from
(ice-9 arrays). Make sure to release the array handle.
* module/ice-9/pretty-print.scm (truncated-print): Support
bitvectors.
Don't try to guess the array prefix but call array-print-prefix from
(ice-9 arrays) instead.
Fix call to print-sequence to support non-zero lower bound arrays.
* test-suite/tests/arrays.test: Test that arrays print properly.
* test-suite/tests/print.test: Test truncated-print with bitvectors,
non-zero lower bound arrays.
* module/ice-9/arrays.scm (array-copy): New function, export.
* module/Makefile.am: Install (ice-9 arrays).
* doc/ref/api-data.texi: Add documentation for (ice-9 arrays).
* libguile/quicksort.i.c: Use signed bounds throughout.
* libguile/sort.c (scm_restricted_vector_sort_x): Fix error calls. Fix
calls to quicksort.
* test-suite/tests/sort.test: Actually test that the sorted results
match the original data. Test cases for non-zero base index arrays for
sort, sort!, and stable-sort!.
* benchmark-suite/benchmarks/uniform-vector-read.bm:
Remove; uniform-vector-read! and uniform-vector-write were deprecated
in 2.0 and are have been removed in 2.1.
* benchmark-suite/benchmarks/bytevector-io.bm: New benchmark.
* benchmark-suite/Makefile.am: Run the new benchmark.
* libguile/array-handle.c (initialize_vector_handle): Set both element
pointers to NULL if the vector is empty.
* libguile/array-map.c (racp): Ignore immutability if destination is
empty.
* test-suite/tests/sort.test: Check empty/mutable/immutable vectors with
sort!.
* test-suite/tests/array-map.test: Check array-copy! with
empty/immutable destination.
* module/system/vm/assembler.scm (standard-prelude, opt-prelude):
(kw-prelude): Emit new instructions in function preludes. Now all
branches are via the new instructions. Remove exports for old
branches.
* module/language/cps/compile-bytecode.scm (compile-function): Rename
the binary* helper back to binary, update uses, and remove logtest
branch as we no longer put logtest in test context.
* module/language/cps/primitives.scm (*comparisons*): Remove logtest.
* module/language/cps/type-fold.scm: Remove logtest folder.
(logbit?): Fold to logand.
* module/language/cps/types.scm (logtest): Update to be a type inferrer
and not a predicate inferrer.
* module/language/tree-il/peval.scm (peval): Transform logtest and
logbit? to (zero? (logand _ _)).
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/effects-analysis.scm:
* module/language/cps/primitives.scm (*comparisons*):
* module/language/cps/type-fold.scm:
* module/language/cps/types.scm: Remove compiler support for u64-scm
comparisons, as this is now inlined.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/cse.scm (compute-equivalent-subexpressions):
* module/language/cps/effects-analysis.scm:
* module/language/cps/primitives.scm (*macro-instruction-arities*):
* module/language/cps/specialize-numbers.scm (compute-specializable-vars):
* module/language/cps/types.scm: Add new variants of u64->scm and
s64->scm that can't be replaced by CSE's auxiliary definitions, so we
can sink unlikely allocations to side branches. This is a hack until
we can get allocation sinking working
* module/language/tree-il/compile-cps.scm (canonicalize, convert):
Simplify handling of branching primcalls so that `convert' only ever
sees branching primcalls in a test context.
* module/system/base/types/internal.scm (heap-tags): Rename number to
heap-number.
* module/system/vm/assembler.scm: Adapt for emit-heap-number?.
* module/system/base/types.scm (%tc7-heap-number): Rename from
%tc7-number.
* module/language/tree-il/compile-cps.scm (convert): Fold test of
constants directly to their branches without reifying booleans, to
simplify "and" chains early.
