This will allow individual primcall converters to define ad-hoc
conversion routines.
* module/language/tree-il/compile-cps.scm (convert-primcall/default)
(convert-indexed-getter, convert-indexed-setter)
(convert-indexed-getter/tag, convert-indexed-setter/untag)
(convert-scm-u64->scm-primcall, convert-scm-u64-scm-primcall)
(convert-scm-u64->f64-primcall, convert-scm-u64-f64-primcall)
(convert-scm-u64->u64-primcall, convert-scm-u64-u64-primcall)
(convert-scm-u64->s64-primcall, convert-scm-u64-s64-primcall)
(convert-*->u64-primcall, convert-scm->u64-primcall)
(convert-u64->scm-primcall): Define some primcall converter helpers.
(*primcall-converters*, define-primcall-converter)
(define-primcall-converters): Define converters for a number of
primcalls.
(convert-primcall*, convert-primcall): Interface to primcall
converters.
(convert): Pass most primcalls through convert-primcall, unless we
know already that they don't need instruction explosion or
boxing/unboxing.
* module/system/base/types/internal.scm (heap-tags): Rename complex and
fraction predicates to "compnum?" and "fracnum?", as they aren't the
same as the Scheme predicates "complex?" and "fraction?".
* module/system/vm/disassembler.scm (define-stack-effect-parser)
(define-clobber-parser):
* module/system/vm/frame.scm (compute-frame-sizes, compute-killv):
Fix bug introduced in dd8bf6a98c whereby
a call would clobber all locals, leaving the backtrace much less
useful than it should be.
Fixes a bug whereby "guild compile --target=arm-linux-gnueabihf
--from=elisp ... language/elisp/boot.el" would fail with an exception
from 'load-thunk-from-memory' while trying to load (language elisp spec)
from the guile-being-compiled instead of using that of the
guile-for-build.
The problem did not manifest with --from=scheme because (language scheme
spec) happened to be already loaded before we had changed %load-path.
* module/scripts/compile.scm (compile): Add calls to 'lookup-language'.
Previously 'load-thunk-from-memory' would often throw to 'system-error'
when passed an incorrect ELF file, leading to incorrect error messages.
* libguile/loader.c (load_thunk_from_memory): Reset 'errno' when
'check_elf_header' returns non-NULL.
* test-suite/tests/vm.test: New file.
* test-suite/Makefile.am (SCM_TESTS): Add it.
* libguile/posix.c (scm_crypt): Take 'scm_i_misc_mutex' right before
calling 'crypt'. Move 'SCM_SYSERROR' call after 'scm_dynwind_end'.
* test-suite/tests/posix.test ("crypt"): New test prefix.
* module/language/cps/dce.scm (compute-live-code): The primcall-param
refactor made it so that a vector-set! to a dead vector would
needlessly mark the vector as live. This commit fixes it and adds
scm-set!, etc.
* module/language/tree-il/compile-cps.scm (init-default-value): Use
undefined? predicate instead of reifying an unbound value in the CPS
term.
* module/language/cps/compile-bytecode.scm (compile-function): Add cases
for the special immediate predicates. Evidently we aren't emitting
these yet!
* module/language/cps/compile-bytecode.scm (compile-function): Allow a
'throw primcall in tail position.
* module/language/cps/prune-bailouts.scm (prune-bailouts): Continue
directly to the nearest tail continuation, so we don't cause
unreachable handle-interrupts / return 0 instructions to be emitted.
* module/language/cps/reify-primitives.scm (reify-clause): Residualized
'throw primcall continues directly to tail.
* module/language/cps/verify.scm (check-arities): Relax check for
'throw.
* module/language/cps/compile-bytecode.scm (compile-function): Add
support for emitting <= via < and jge / jnge.
* module/language/cps/effects-analysis.scm: Declare effects for <= and
f64-<=.
* module/language/cps/primitives.scm (*comparisons*): Add <=, f64-<=.
* module/language/cps/specialize-numbers.scm (specialize-operations):
Specialize <= to < for integer comparisons. Specialize to f64-<= for
f64 ops.
* module/language/cps/type-fold.scm (<=): Add folder.
* module/language/cps/types.scm (infer-<=): Add inferrer.
* module/language/tree-il/compile-cps.scm (canonicalize): Canonicalize
<= and >= to <=, so that nans are handled correctly.
* module/language/cps/type-fold.scm (compare-exact-ranges): Rename from
compare-integer-ranges.
(<, u64-<, s64-<, =, u64-=, s64-=): Separate the generic and unboxed
implementations.
* module/language/tree-il/compile-cps.scm (adapt-arity): Allow k to be
$kargs for the 1-valued case.
(convert): For single-valued continuations where the definition is
clearly single-valued, avoid making a needless $kreceive and extra
"rest" binding that will just be filled with () and have to be
eliminated later.
* module/language/tree-il/compile-cps.scm (adapt-arity): Allow k to be
$kargs for the 1-valued case.
(convert): For single-valued continuations where the definition is
clearly single-valued, avoid making a needless $kreceive and extra
"rest" binding that will just be filled with () and have to be
eliminated later.
* 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/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/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.
* 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.
