* module/language/cps/specialize-numbers.scm (preserve-eq?): New
helper.
(sigbits-union): Use the new helper. Fixes bugs.gnu.org/38486.
Thanks to Zack Marvel for the bug report and Matt Wette for tracking
it down.
Fixes <https://bugs.gnu.org/33036>.
Reported by <calcium@disroot.org>.
* module/language/cps/types.scm (define-simple-type-inferrer):
Apply (logand (&type val) <>) uniformly. Previously, this was done only
in the false branch. Rename local variable to 'type*', to allow the
macro operand 'type' to be an arbitrary expression.
(*type-inferrers*)<null?>: Add &nil to the set of possible types.
(*type-inferrers*)<nil?>: Add &false and &null to the set the possible
types.
* module/language/cps/type-fold.scm (*branch-folders*)<null?>: Add &nil
to the set of possible types.
(*branch-folders*)<nil?>: Add &false and &null to the set the possible
types.
* test-suite/tests/compiler.test: Add tests.
Fixes <https://bugs.gnu.org/31474> and related bugs.
Reported by Jan Nieuwenhuizen <janneke@gnu.org>.
* module/language/cps/types.scm (next-power-of-two): Remove procedure.
(non-negative?, lognot*, saturate+, saturate-, logand-bounds)
(logsub-bounds, logior-bounds, logxor-bounds): New procedures. Use them
to improve and fix bugs in the range analysis of the type inferrers for
'logand', 'logsub', 'logior', 'ulogior', 'logxor', 'ulogxor', and
'lognot'.
* module/language/cps/compile-bytecode.scm (compile-function)
(emit-bytecode):
* module/language/cps/slot-allocation.scm (allocate-slots):
* module/language/cps/optimize.scm (cps-default-optimization-options):
Allow the "lazy vars" optimization, a form of slot precoloring, to be
disabled. It will be disabled at -O0 or -O1, to speed compilation
times.
* module/language/cps/cse.scm (compute-available-expressions):
(compute-equivalent-subexpressions): Improve algorithmic complexity of
CSE by pre-computing the labels whose reads are clobbered by a label's
writes.
* module/language/cps/cse.scm (compute-equivalent-subexpressions): Minor
optimization to reduce the size of equivalent expression keys, and to
avoid some work if an expression has no key.
* module/language/cps/slot-allocation.scm
(compute-reverse-control-flow-order): For graphs without back-edges,
use a simplified computation of reverse control flow order.
* module/language/cps/types.scm (logand): We were computing the wrong
ranges when either argument was negative; a terrible bug!
(logsub): Also fix range when A is negative and B is non-negative.
(ulogand): Tighten up range.
* libguile/Makefile.am (libguile_@GUILE_EFFECTIVE_VERSION@_la_SOURCES):
(DOT_X_FILES, DOT_DOC_FILES, noinst_HEADERS): Add syntax.c and
syntax.h.
* libguile/evalext.c (scm_self_evaluating_p):
* libguile/goops.c (class_syntax, scm_class_of, scm_goops_early_init):
* libguile/init.c (scm_init_guile):
* libguile/print.c (iprin1):
* libguile/tags.h (scm_tc7_syntax):
* module/oop/goops.scm (<syntax>):
* module/system/base/types.scm (%tc7-syntax, cell->object):
* module/system/vm/disassembler.scm (code-annotation): Wire up the new
data type.
* libguile/syntax.c:
* libguile/syntax.h: New files.
* module/ice-9/boot-9.scm: Move new definitions to (system syntax
internal).
* module/system/syntax.scm (print-syntax): New helper.
* module/system/vm/assembler.scm (statically-allocatable?)
(intern-constant, link-data): Arrange to be able to write syntax
objects into images.
* module/language/cps/types.scm (&syntax): New type. Remove
&hash-table; it was never detected, an internal binding, and we need
the bit to avoid going into bignum territory.
* module/language/cps/effects-analysis.scm (expression-effects): Prompts
cause &all-effects. I tried to limit this change to CSE but it was
actually LICM that was borked, so better to be conservative
* test-suite/tests/control.test ("escape-only continuations"): Add
test.
* doc/ref/vm.texi (Inlined Scheme Instructions): Add string-set!.
* libguile/vm-engine.c (string-set!): New opcode.
* module/ice-9/rdelim.scm (read-string): Reimplement in terms of a
geometrically growing list of strings, to reduce total heap usage when
reading big files.
* module/language/cps/compile-bytecode.scm (compile-function): Add
string-set! support.
* module/language/cps/types.scm (string-set!): Update for &u64 index.
* module/language/tree-il/compile-cps.scm (convert): Unbox index to
string-set!.
* module/system/vm/assembler.scm (system): Export string-set!.
This bug was introduced by 35a9059250.
* module/language/cps/specialize-numbers.scm (specialize-operations):
Check that both operands are real as a condition for
specialize-f64-comparison.
* test-suite/tests/numbers.test: Add test.
* module/language/cps/compile-bytecode.scm (compile-function): Check for
fallthrough after $kclause too; possible to need to jump if clause
tails are contified.
* module/language/cps/contification.scm (compute-contification-candidates):
Enable inter-clause contification.
* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/slot-allocation.scm ($allocation)
(lookup-nlocals, compute-frame-size, allocate-slots): Adapt to
have one frame size per function, for all clauses.
* module/language/cps/types.scm (define-binary-result!): Arithmetic
where one argument is a flonum may produce a complex.
* test-suite/tests/compiler.test: Add test.
* module/language/cps/specialize-numbers.scm (compute-significant-bits):
Always revisit predecessors after first visit. Avoids situation where
predecessor of an unvisited phi var could default to 0 significant
bits and never be revisited. Fixes (format #f "~2f" 9.9).
* libguile/vm-engine.c (fluid-set!): Fix name of opcode to correspond
with name of Tree-IL primitive. Fixes compilation of fluid-set! to
actually use the fluid-set! opcode.
* doc/ref/vm.texi (Dynamic Environment Instructions): Update.
* module/language/cps/compile-bytecode.scm (compile-function): Add
fluid-set! case.
* module/system/vm/assembler.scm: Update export name for
emit-fluid-set!.
* module/language/cps/slot-allocation.scm
(add-prompt-control-flow-edges): Fix to add links from prompt bodies
to handlers, even in cases where the handler can reach the body but
the body can't reach the handler.
* test-suite/tests/compiler.test ("prompt body slot allocation"): Add
test case.
* doc/ref/vm.texi (Inlined Atomic Instructions): New section.
* libguile/vm-engine.c (VM_VALIDATE_ATOMIC_BOX, make-atomic-box)
(atomic-box-ref, atomic-box-set!, atomic-box-swap!)
(atomic-box-compare-and-swap!): New instructions.
* libguile/vm.c: Include atomic and atomics-internal.h.
(vm_error_not_a_atomic_box): New function.
* module/ice-9/atomic.scm: Register primitives with the compiler.
* module/language/cps/compile-bytecode.scm (compile-function): Add
support for atomic ops.
* module/language/cps/effects-analysis.scm: Add comment about why no
effects analysis needed.
* module/language/cps/reify-primitives.scm (primitive-module): Add case
for (ice-9 atomic).
* module/language/tree-il/primitives.scm (*effect-free-primitives*):
(*effect+exception-free-primitives*): Add atomic-box?.
* module/system/vm/assembler.scm: Add new instructions.
* test-suite/tests/atomic.test: Test with compilation and
interpretation.
* module/language/tree-il/primitives.scm
(character-comparison-expander): Expand out char<? and friends to <,
unboxing the char arguments.
* module/language/cps/types.scm:
* module/language/cps/effects-analysis.scm: Remove mention of char<? and
friends as we won't see them any more. Also fixes#24318.
* module/language/cps/specialize-numbers.scm (truncate-u64): New helper,
truncates a SCM value.
(specialize-u64-binop): Add ulogxor case.
(sigbits-union, sigbits-intersect, sigbits-intersect3)
(next-power-of-two, range->sigbits, inferred-sigbits)
(significant-bits-handlers, define-significant-bits-handler):
(significant-bits-handler, compute-significant-bits): Add facility to
compute the bits in a value that are significant.
(specialize-operations): Unbox in more cases, when only u64 bits are
significant. Unbox logxor. Elide logand where it has no effect.
* module/language/cps/rotate-loops.scm (rotate-loops-in-function): Don't
attempt to rotate a loop whose header is a $prompt. Fixes use of
`continue' in `while'. Thanks to Nala Ginrut for the report :)
* doc/ref/vm.texi (Top-Level Environment Instructions): Update
documentation.
* libguile/_scm.h (SCM_OBJCODE_MINOR_VERSION): Bump, sadly.
* module/system/vm/assembler.scm (*bytecode-minor-version*): Bump.
* libguile/vm-engine.c (define!): Change to store variable in dst slot.
* module/language/tree-il/compile-cps.scm (convert):
* module/language/cps/compile-bytecode.scm (compile-function): Adapt to
define! change.
* module/language/cps/effects-analysis.scm (current-module): Fix define!
effects. Incidentally here was the bug: in Guile 2.2 you can't have
effects on different object kinds in one instruction, without
reverting to &unknown-memory-kinds.
* test-suite/tests/compiler.test ("regression tests"): Add a test.
* module/language/cps/types.scm (type-entry-saturating-union): Fix range
saturation in the negative direction. Previously we were artificially
truncating negative range ends to zero.
* module/language/cps/optimize.scm (cps-default-optimization-options):
Add missing #:simplify? option. Otherwise the simplification pass was
running even at -O0.
* module/language/cps/simplify.scm (compute-eta-reductions): Eta-reduce
branches as well, so that passing a constant to a branch will fold to
the true or false branch, provided that the test variable was just
used in the branch.
* module/language/cps/reify-primitives.scm (primitive-module): Add cases
for SRFI-4 primitives. These primitives are only treated as such for
the purposes of Tree-IL primitive expansion; perhaps the right fix is
elsewhere, but it's here for now.
* module/language/cps/types.scm (&min/0, &min/s64, &max/s64, &max/size)
(&max/u64, &max/vector): New clamped variable range accessors. Use
them in type inferrers.
* module/language/cps/utils.scm (intmap-map): Use transient intmap-add!
on an empty intmap to build the result instead of intmap-replace! on
the argument. Avoids spooky action-at-a-distance mutation of the
argument if it happens to be a transient -- although the intmap-fold
will correctly traverse a snapshot of the argument and the result will
be correct, the argument value would be modified in place, causing
strange results to calling code that passes in a transient.
