* module/language/tree-il/analyze.scm (validate-arity): Fix for
applicable structs. Applicable structs are procedures, but not every
struct has a first slot, and not every struct with a procedure in its
first slot is applicable. Besides, the approach in this patch gives
better errors.
* module/language/tree-il/effects.scm (make-effects-analyzer): The
analyzer will take an optional second argument, a lookup procedure of
type sym -> exp. This can let the analyzer dig into calls to
lexically bound procedures.
* module/language/tree-il/cse.scm (singly-valued-expression?, cse):
Allow CSE to propagate lexicals to tail positions, if the expression
is singly-valued.
* module/language/tree-il/peval.scm (peval): If we can lift one common
test, see if we can lift others as well.
* test-suite/tests/peval.test: Add a test.
* module/language/tree-il/peval.scm (peval): Optimize common tests in
chains of "if" expressions, like those generated by matchers.
* test-suite/tests/peval.test ("partial evaluation"): Add a test.
* module/language/tree-il/analyze.scm (proc-ref?)[special?]: New
procedure.
Return #t for any toplevel-ref of `_'.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string using gettext as top-level _"): New test.
* module/language/tree-il/analyze.scm (proc-ref?): Check for
SPECIAL-NAME in the <module-ref> and <lexical-ref> cases too.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string using gettext as module-ref _", "non-literal format
string using gettext as lexical _"): New tests.
* module/language/tree-il/peval.scm (<operand>, make-operand)
(make-bound-operands, peval): Avoid emitting needless aliases in
degenerate cases of let.
(visit-operand): If we visit an operand with a fresh counter and have
to abort, record that fact.
* test-suite/tests/peval.test ("partial evaluation"): Add a test.
* module/language/tree-il/cse.scm: New pass, some simple common
subexpression elimination with effects analysis.
* test-suite/tests/cse.test: New test.
* test-suite/Makefile.am:
* module/Makefile.am: Adapt.
* module/language/tree-il/peval.scm: Use effects analysis from (language
tree-il effects) instead of our own constant-expression?. Eagerly
mark assigned lexicals as non-copyable.
* module/language/tree-il/primitives.scm
(*interesting-primitive-names*): Add number? and char?. Add more
numeric predicates. Add character comparators. Add throw, error, and
scm-error.
(*primitive-accessors*): Remove struct-vtable. Though the vtable's
contents may change (through redefinition), its identity does not
change.
(*effect-free-primitives*): Put struct-vtable, number?, and char?
here.
(*multiply-valued-primitives*): Instead of listing singly-valued
primitives, list multiply-valued primitives.
(*bailout-primitives*): New list.
(*negatable-primitives*): New alist.
(*bailout-primitive-table*, *multiply-valued-primitive-table*)
(*negatable-primitive-table*): New tables.
(singly-valued-primitive?): Adapt to
use *multiply-valued-primitive-table*.
(bailout-primitive?, negate-primitive): New exported procedures.
* module/language/tree-il/primitives.scm (resolve-primitives!): Resolve
public module-refs to primitives.
* test-suite/tests/tree-il.test: New tests for primitive resolution.
* module/language/tree-il/analyze.scm (proc-ref?): Change to use less
false-if-exception and more variable-bound?. If a variable is present
in the local module but not bound, assume that it is gettext if it has
the right name. This is to allow for (define _ gettext).
* test-suite/tests/tree-il.test ("warnings"): Update (_ "foo") example.
* module/language/tree-il/peval.scm (peval): Add optimization to
hoist the inner procedure out of e.g.
(lambda args (apply (lambda ...) args))
This commit restores the ability to detect escape-only prompts at
compile-time.
* test-suite/tests/tree-il.test: Update test for prompt with a lambda,
and add a specific test for lambda application.
* module/language/tree-il/canonicalize.scm (canonicalize!): Fix a bug in
which the sense of `escape-only?' was reversed. We never saw this
though, because for other reasons, no prompts were being identified as
escape-only.
* module/language/tree-il/primitives.scm (*primitive-expand-table*):
Don't limit the call-with-prompt to <prompt> transition to lambda
expressions. Instead we can lexically bind the handler, and rely on
peval to propagate a lambda expression.
* module/language/tree-il/analyze.scm (proc-ref?): Wrap `variable-ref'
in `false-if-exception'.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string with forward declaration"): New test.
* module/language/tree-il/analyze.scm (gettext?): Rename to...
(proc-ref?): ... this. Add `proc' and `special-name' parameters.
(gettext?): Define in terms of `proc-ref?'.
(ngettext?): New procedure.
(const-fmt): Recognize `ngettext' calls.
(format-analysis)[<down>](check-format-args]: Check
constant-but-non-string 2nd argument in the (not (const-fmt ...))
case.
[check-simple-format-args]: Use `const-fmt'.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string using ngettext", "non-literal format string using
ngettext as N_"): New tests.
("simple-format")["unsupported, ngettext"]: New test.
Fixes <http://bugs.gnu.org/10846>.
Reported by Bruno Haible <bruno@clisp.org>.
* module/language/tree-il/analyze.scm (gettext?): New procedure.
(const-fmt): Add `env' parameter; update callers. Use `gettext?'.
(format-analysis)[check-simple-format-args]: Actually support
gettextized format strings.
* test-suite/tests/tree-il.test ("warnings")["format"]("non-literal
format string using gettext"): Use `gettext' as the procedure name.
("non-literal format string using gettext as _"): New test.
["simple-format"]("unsupported, gettext"): New test.
* module/language/tree-il/analyze.scm (format-string-argument-count):
Add support for ~h.
* test-suite/tests/tree-il.test ("warnings")["format"]("~h", "~:h with
locale object", "~:h without locale object"): New tests.
* module/language/tree-il/peval.scm (peval):
* module/language/tree-il/primitives.scm (dynamic-wind): When you make a
gensym that just has to be compared against other gensyms, it will be
unique if the prefix doesn't end in something that can be interpreted
as a number. There's no reason to make that character something
difficult like " ". So change to use a dash in that case.
* module/ice-9/psyntax-pp.scm: Regenerate. More readable now.
* module/language/tree-il/analyze.scm
(format-analysis)[check-simple-format-args]: New procedure. Use it.
Add support for applications of <module-ref>.
* module/system/base/message.scm (%warning-types): Handle the `format
simple-format' warning.
* module/language/scheme/spec.scm (scheme)[make-default-environment]:
Use `simple-format' as the default `format'.
* test-suite/tests/tree-il.test ("warnings")["format"]: Explicitly use
(@ (ice-9 format) format) where needed.
("simple-format"): New test prefix.
* module/language/tree-il/peval.scm (peval): Fold (values
'singly-valued-expression) to 'singly-valued-expression in contexts
that expect multiple values, in addition to those that expect single
values.
* module/language/tree-il/compile-glil.scm (flatten-lambda-case): Fix
miscompilation of `values' in a push context with RA.
* test-suite/tests/tree-il.test: Add low-level test for this
miscompilation.
* module/language/tree-il/peval.scm (singly-valued-expression?): New
helper.
(truncate-values): Use the helper.
(make-operand): Minor refactor.
(set-operand-residual-value!): Try to undo the effects of (values
FOO), if the continuation will check itself for the correct number of
values.
(peval): Fold helpers into fold-constant. Add a constant-expression?
case for (values FOO). Add a new context: "values", for contexts in
which multiple values are allowed, either because of being in a tail
context relative to a function, or because of let-values. "value" is
now for single values. Don't visit operands for "values", as their
binding form truncates to one value. Add a case to fold (values ...)
forms. Fix folding of (lambda), to process the cases in values
context instead of tail context (which could have been "value", which
would cause the procedure to truncate).
Reported by Cédric Cellier <rixed@happyleptic.org>.
* module/language/tree-il/peval.scm (truncate-values): New procedure.
(make-operand): Call `truncate-values' SOURCE.
* test-suite/tests/tree-il.test ("partial evaluation"): New tests for
multiple value truncation.
* module/language/tree-il/compile-glil.scm (flatten-lambda-case): Clear
lexical stack slots at the end of a non-tail let, letrec, or fix.
Fixes http://debbugs.gnu.org/9900.
* test-suite/tests/gc.test ("gc"): Add test.
* module/language/tree-il/peval.scm (peval): The <dynwind> compiler will
copy the winder and unwinder values, so make sure that they are
constant, and if not, create lexical bindings. Fixes
http://debbugs.gnu.org/9844.
* test-suite/tests/tree-il.test ("partial evaluation"): Add a couple
<dynwind> tests.
* module/language/tree-il/peval.scm (peval): Since constant-expression?
is used to determine whether to copy values, return #f if any lexical
is assigned.
* module/language/tree-il/analyze.scm (analyze-lexicals): When stepping
into a non-tail form, we know that labels allocation will be invalid,
so use an empty labels set. Fixes http://debbugs.gnu.org/9769.
* test-suite/tests/tree-il.test ("labels allocation"): Add a test.
* module/language/tree-il/peval.scm (peval): Factor constant folding out
to a helper. Use it in the accessor case in addition to the normal
effect-free-primitive case.
* test-suite/tests/tree-il.test: Add a test.
* module/language/tree-il/peval.scm (peval): Add special handlers for
memq and memv, as inline.scm used to have. This is important for
`case' clauses. It is very ugly, though.
* test-suite/tests/tree-il.test ("partial evaluation"): Add tests.
* module/language/tree-il/primitives.scm (*primitive-accessors*): New
set of primitives: those that access mutable memory, but are otherwise
pure. Include bytevector references here.
(accessor-primitive?): New public predicate.
* module/language/tree-il/peval.scm (peval): Refactor to distinguish
constructor-primitive? from accessor-primitive?.
