* module/language/cps/closure-conversion.scm (convert-one): Fix bug when
getting value of SCC whose free variables have been elided. Thanks to
abcdw for the report!
* test-suite/tests/compiler.test ("cse auxiliary definitions"): Remove
spurious newline.
("closure conversion"): New test.
* module/language/cps/contification.scm (compute-first-class-functions):
(compute-functions-called-by-label):
(compute-functions):
(compute-arities):
(compute-contification-candidates):
(compute-call-graph):
(compute-contification):
(apply-contification):
(contify): Given that the frontend will produce $callk now, allow it to
be contified if such callees are all called with the same continuation.
* module/language/tree-il/compile-cps.scm (module-call-stubs):
(module-call-label, convert, cps-convert/thunk): Arrange to call module
variables through out-of-line trampolines with unchecked arity. This
should speed up compile time in large files and reduce code size on hot
paths.
* module/language/cps/closure-conversion.scm: Use standard
compute-reachable-functions and intmap-select from utils to filter
reachable functions, allowing us to pick up callk. Adapt some uses to
expect callk for calls.
* module/language/cps/self-references.scm (resolve-self-references):
Subst the proc, if it's there.
* module/language/cps/split-rec.scm (compute-free-vars): Add a case for
callk.
* libguile/intrinsics.h:
* libguile/intrinsics.c (lookup_bound_public, lookup_bound_private): Two
new intrinsics.
(scm_bootstrap_intrinsics): Wire them up.
* libguile/jit.c (compile_call_scm_from_scmn_scmn):
(compile_call_scm_from_scmn_scmn_slow):
(COMPILE_X8_S24__N32__N32__C32): Add JIT support for new instruction
kind.
* libguile/vm-engine.c (call-scm<-scmn-scmn): New instruction, takes
arguments as non-immediate offsets, to avoid needless loads and register
pressure.
* module/language/cps/effects-analysis.scm: Add cases for new
primcalls.
* module/language/cps/compile-bytecode.scm (compile-function): Add new
primcalls.
* module/language/cps/reify-primitives.scm (cached-module-box): If the
variable is bound, call lookup-bound-public / lookup-bound-private as
appropriate instead of separately resolving the module, name, and doing
the bound check.
* module/language/tree-il/compile-bytecode.scm (emit-cached-module-box):
Use new instructions.
* module/system/vm/assembler.scm (define-scm<-scmn-scmn-intrinsic):
(lookup-bound-public, lookup-bound-private): Add assembler support.
* module/language/cps/reify-primitives.scm (cached-module-box): Include
public? in cache key, so we don't accidentally alias private and
exported names. Also include bound?, to avoid a window in which thread
A resolves and caches var V in preparation for setting it, but thread B
sees V for ref before it was initialized.
* module/language/cps.scm:
* module/language/cps/contification.scm:
* module/language/cps/cse.scm:
* module/language/cps/dce.scm:
* module/language/cps/simplify.scm:
* module/language/cps/slot-allocation.scm:
* module/language/cps/types.scm: Allow $kargs to follow $kfun. In that
case, the function must be well-known and callers are responsible for
calling with the appropriate arity.
* module/language/cps/compile-bytecode.scm: Emit "unchecked-arity" for
$kargs following $kfun.
* module/system/vm/assembler.scm: Adapt.
* module/language/tree-il/fix-letrec.scm (reorder-bindings):
(fix-letrec): Reorder definitions so that lambdas tend to stick
together, to avoid "complex" expressions interposing in lambda SCCs.
Instead of defining a separate module, given that "read" calls are quite
all over the place, we're just going to replace the boot "read" binding
with read.scm. This way, we'll be able to remove support for reader
options in the boot reader, as it will only ever be used for a finite
set of files.
* NEWS: Update.
* module/Makefile.am (ice-9/boot-9.go): Depend on read.scm.
(SOURCES):
* am/bootstrap.am (SOURCES): Don't build a ice-9/read.go, as we include
it.
* module/ice-9/boot-9.scm (read-syntax): Define here, as "include" now
uses it.
(read-hash-procedures, read-hash-procedure, read-hash-extend): New
procedures. Will replace C variants.
(read, read-syntax): Include read.scm to define these.
* module/ice-9/psyntax-pp.scm (include): Regenerate.
* module/ice-9/psyntax.scm (include): Use read-syntax, so we get better
source information.
* module/ice-9/read.scm (let*-values): New local definition, to avoid
loading srfi-11.
(%read): Use list->typed-array instead of u8-list->bytevector.
* module/language/scheme/spec.scm: Remove (ice-9 read) import;
read-syntax is there in the boot environment
* module/language/cps/types.scm (constant-type): Add case for EOF.
* module/language/tree-il/primitives.scm (*interesting-primitive-names*):
(*effect+exception-free-primitives*): Add case for eof-object?.
(eof-object?): Expand to eq? on the-eof-object.
* module/ice-9/boot-9.scm (%auto-compilation-options): Add
use-before-definition and non-idempotent-definition.
* module/language/tree-il/analyze.scm (<use-before-def-info>): New
analysis info.
(make-use-before-definition-analysis): New function.
(goops-toplevel-definition): Move down.
(unbound-variable-analysis, macro-use-before-definition): Remove, as
they are subsumed by use-before-def. There are some deprecated
bindings though.
(make-analyzer): Rework to allow for use-before-def analysis to handle
multiple
* module/system/base/message.scm (%warning-types): Add handlers for the
new warning types.
* test-suite/tests/tree-il.test: Add tests.
* doc/ref/api-evaluation.texi (Compilation): Update.
* module/language/cps/cse.scm (intset-intersect*): New helper. Use it
to replace manual uses.
(lset-unionq, meet-constants, adjoin-constant, set-constants): New
helpers.
(compute-consts): New function, to compute constants at each label,
using not only definitions but flow.
(<analysis>): Add consts to analysis.
(elide-predecessor, prune-branch, forward-branch, compute-out-edges)
(propagate-analysis, eliminate-common-subexpressions-in-fun): Plumb
consts through the algorithm.
(fold-branch): Fold an eq-constant? using the flow-determined constant
info. Finally allows compile-bytecode to fold to switch statements!
* module/language/cps/optimize.scm (optimize-first-order-cps): Move
branch chain optimization before the final CSE/DCE pass.
* module/language/cps/compile-bytecode.scm (compile-function): Expect
eq-constant? instead of eq-null?, etc.
* module/language/cps/effects-analysis.scm: Likewise.
* module/language/cps/reify-primitives.scm (reify-primitives): For
eq-constant?, reify a $const unless the constant is an immediate whose
encoding fits in 16 bits.
* module/language/cps/type-fold.scm (materialize-constant): Helper to
make a constant from a type, min, and max.
(fold-eq-constant?): New helper.
(eq-constant?): New folder.
(undefined?): Define specifically.
(define-nullish-predicate-folder): Renamd from
define-special-immediate-predicate-folder. Use only for null?, false,
and nil?.
(*branch-reducers*): New mechanism. Reduce eq? to eq-constant? if
possible.
(local-type-fold): Refactor to use materialize-constant, and to allow
reducing branches.
* module/language/cps/types.scm (constant-type): Return three values
instead of a type entry.
(constant-type-entry): New function that returns a type entry. Adapt
callers.
(infer-constant-comparison): New helper.
(eq-constant?): New inferrer.
(undefined?): New inferrer.
* module/language/tree-il/compile-bytecode.scm (eq-constant?): Fix
truncate-bits signed arg.
(define-immediate-type-predicate): Adapt to visit-immediate-tags
change.
* module/language/tree-il/compile-cps.scm (convert): Convert eq? to
constant to eq-constant?. Advantaged is that it gets fixnums and
chars in addition to special immediates.
* module/language/tree-il/cps-primitives.scm (define-immediate-type-predicate):
Adapt to allow #f as pred.
* module/system/base/types/internal.scm (immediate-tags): Use #f as pred
for false, nil, etc.
(immediate-bits->scm): Adapt.
* module/system/vm/assembler.scm (emit-eq-null?, emit-eq-nil?)
(emit-eq-false?, emit-eq-true?, emit-unspecified?, emit-eof-object?):
Remove specialized emitters.
* module/system/vm/assembler.scm (define-immediate-tag=?-macro-assembler):
Allow for pred to be #f.
* module/system/vm/disassembler.scm (define-immediate-tag-annotation):
Adapt to pred being #f.
* module/language/tree-il/compile-bytecode.scm (eq?): Define
eq-immediate? as immediate emitter.
(canonicalize): Don't fuss so much about eq?; only if commutation is
needed. (Perhaps a more generic commutation pass is needed.)
(compile-closure): Add support for emit/immediate for branches.
* doc/ref/vm.texi (Instruction Set, Constant Instructions): Document new
instruction.
* libguile/instructions.c (FOR_EACH_INSTRUCTION_WORD_TYPE): New first
word kind with zi16 operand.
* libguile/jit.c (compile_make_immediate, compile_make_immediate_slow):
New compilers.
(COMPILE_X8_S8_ZI16): New operand kind.
* libguile/vm-engine.c (make-immediate): New instruction.
* module/language/bytecode.scm:
* module/system/vm/assembler.scm (encode-X8_S8_ZI16<-/shuffle):
(signed-bits, load-constant): Support the new instruction kind.
* module/system/vm/disassembler.scm (disassemblers)
(sign-extended-immediate, code-annotation): Support for zi16
operands.
Fixes a bug whereby, for example, "guild compile --target=i686-linux-gnu"
running on x86_64 would generate invalid code for 'bytevector-u32-native-set!'
because 'target-most-positive-fixnum' was called from the top-level
when (language tree-il compile-cps) was loaded.
Consequently, the .go files under prebuilt/ would be invalid, leading to
build failures on 32-bit platforms.
This issue became apparent with cb8cabe85f.
* module/language/tree-il/compile-cps.scm (bytevector-ref-converter)[tag]:
Turn into a lambda so that 'target-most-positive-fixnum' is called in
the right context.
(bytevector-set-converter)[integer-unboxer]: Likewise.
* module/language/cps/cse.scm (eliminate-common-subexpressions-in-fun):
I think it's possible to get an orphan loop, with predecessors
after successors in the original RPO. Handle that here.
* bootstrap/Makefile.am (GUILE_OPTIMIZATIONS): Change to just -O1.
* module/language/tree-il/spec.scm (choose-compiler): Use CPS for -O2
and higher.
* module/system/base/optimize.scm (available-optimizations): CPS for -O2
and higher, but -Oresolve-primitives now at -O1 also.
* module/language/cps/cse.scm (propagate-analysis): New helper.
(eliminate-common-subexpressions-in-fun): Recompute avail and bool set
in response to simplifications in predecessor CFG. Allows much better
compilation of pattern-matching idioms!
* module/language/cps/cse.scm (forward-cont, forward-branch)
(compute-avail-and-bool-edge): New helpers.
(add-equivalent-expression!): Allow idempotent adds; can happen now
when revisiting a cont after changes to its predecessors.
(fold-branch): New helper.
(eliminate-common-subexpressions-in-fun): Allow for reductions to
branch predecessors. In that case, revisit the branch, as the CFG
will have changed.
* module/language/cps/cse.scm (elide-predecessor, prune-branch)
(prune-successors, term-successors): New helpers.
(eliminate-common-subexpressions-in-fun): When we modify the CFG,
update the analysis. Also, thread the substs map through CSE so that
closures in high-level CPS can take advantage of eliminated variables.
(fold-renumbered-functions): Take multiple seeds.
(eliminate-common-subexpressions): Thread var substs map through CSE.
* module/language/cps/cse.scm (<analysis>): New data type, grouping
available expression analysis, predecessor map, etc.
(eliminate-common-subexpressions-in-fun): Instead of having a static
analysis, thread it through the CSE pass so that we can update the CFG
as we go.
* module/language/cps/cse.scm (compute-available-expressions): Take a
clobber map instead of an effects map.
(compute-singly-referenced): Remove unused function.
(eliminate-common-subexpressions-in-fun): Keep a preds map. Use it
add entries to the equiv-set and var-substs at expression
continuations instead of at the expression terms themselves.
* module/language/cps/cse.scm (compute-truthy-expressions): Operate on a
single function.
(eliminate-common-subexpressions-in-fun): Instead of computing a set
of labels to eliminate, go ahead and do the elimination as we go.
(fold-renumbered-functions): Can just use a single seed now.
(eliminate-common-subexpressions): Simplify to just fold over
functions, building up renamed output as we go.
* module/language/cps/cse.scm (fold-renumbered-functions): New helper.
(compute-equivalent-expressions): Use new helper.
(compute-equivalent-expressions-in-fun): Lift to top-level.
(eliminate-common-subexpressions): Adapt.
* module/language/tree-il/optimize.scm (make-optimizer): New procedure,
to compute an optimizer given options, lazily loading optimization
modules.
(optimize): Use make-optimizer.
(make-lowerer): Use make-optimizer, so as to only load needed
optimizations. Speeds up bootstrap times.
* module/language/tree-il/primitives.scm (expand-eq): Just expand out to
binary comparisons. Also expand eq?, which was missing. Leave
strength reduction to peval.
(character-comparison-expander): Move down, as it depends on <, <=,
and so on.
* module/language/tree-il/peval.scm (peval): Robustly reduce equal? and
eqv?.
* test-suite/tests/peval.test ("partial evaluation"): Expect fixnum
comparison to reduce to eq?.
("eqv?", "equal?"): A new battery of tests.
* test-suite/tests/tree-il.test ("primitives"): Remove reduction tests.
* module/language/tree-il.scm (with-lexicals): New public helper.
* .dir-locals.el (with-lexicals): Add indentation rule.
* module/language/tree-il/compile-bytecode.scm (canonicalize): Use
with-lexicals.
* module/language/tree-il/compile-cps.scm (canonicalize): Use
with-lexicals from tree-il.
* module/language/tree-il/primitives.scm (chained-comparison-expander):
Remove duplicate expander definitions for <, <=, and so on.
* module/language/tree-il/primitives.scm (maybe-simplify-to-eq): Avoid
inadvertent code duplication by using with-lexicals.
(expand-chained-comparisons): Likewise.
(call-with-prompt): Simplify to use with-lexicals.
* module/language/cps/optimize.scm (define-optimizer):
(optimize-higher-order-cps, optimize-first-order-cps):
(make-cps-lowerer):
* module/language/tree-il/optimize.scm (optimize, make-lowerer): In an
embarrassing bug, after parsing optimization arguments, we were
aconsing them instead of the expected cons*. This meant the bootstrap
was running all Tree-IL optimizations! Change to have optimizers not
have defaults and use alists after parsing.
* module/language/tree-il/compile-bytecode.scm (compile-closure):
for-value-at and for-values-at take indexes instead of environments to
denote destination.
