Merge pull request #12700 from xavierleroy/gen-prims

Generate runtime/prims.c with the correct types in primitive declarations

Changes

@@ -149,6 +149,10 @@ Working version
   `CAMLprim`, or remove them entirely if no longer used.
   (Xavier Leroy, review by David Allsopp)
 
+- #12700, continuing #11763 and trying to address #12660:
+  Use the correct types for primitives when generating the table of primitives
+  used by ocamlrun.
+  (Xavier Leroy, motivation, review and improvements by Antonin Décimo)
 
 ### Code generation and optimizations:
 

Makefile

@@ -1251,48 +1251,19 @@ runtime/ld.conf: $(ROOTDIR)/Makefile.config
 	$(V_GEN)echo "$(STUBLIBDIR)" > $@ && \
 	echo "$(LIBDIR)" >> $@
 
-# If primitives contain duplicated lines (e.g. because the code is defined
-# like
-# #ifdef X
-# CAMLprim value caml_foo() ...
-# #else
-# CAMLprim value caml_foo() ...
-# #endif), horrible things will happen: duplicated entries in Runtimedef ->
-# double registration in Symtable -> empty entry in the PRIM table ->
-# the bytecode interpreter is confused.
-# We sort the primitive file and remove duplicates to avoid this problem.
-
-# Warning: we use "sort | uniq" instead of "sort -u" because in the MSVC
-# port, the "sort" program in the path is Microsoft's and not cygwin's
-
-# Warning: POSIX sort is locale dependent, that's why we set LC_ALL explicitly.
-# Sort is unstable for "is_directory" and "isatty"
-# see http://pubs.opengroup.org/onlinepubs/9699919799/utilities/sort.html:
-# "using sort to process pathnames, it is recommended that LC_ALL .. set to C"
-
 # To speed up builds, we avoid changing "primitives" when files
 # containing primitives change but the primitives table does not
-runtime/primitives: \
-  $(shell runtime/gen_primitives.sh > runtime/primitives.new; \
+runtime/primitives: runtime/gen_primitives.sh \
+  $(shell runtime/gen_primitives.sh $(runtime_BYTECODE_C_SOURCES) \
+                    > runtime/primitives.new; \
                     cmp -s runtime/primitives runtime/primitives.new || \
                     echo runtime/primitives.new)
-	$(V_GEN)cp $^ $@
-
-runtime/prims.c : runtime/primitives
-	$(V_GEN)export LC_ALL=C; \
-	(echo '#include "caml/config.h"'; \
-	 echo 'typedef intnat value;'; \
-	 echo 'typedef value (*c_primitive)(void);'; \
-	 echo; \
-	 sed -e 's/.*/extern value &(void);/' $<; \
-	 echo; \
-	 echo 'const c_primitive caml_builtin_cprim[] = {'; \
-	 sed -e 's/.*/  &,/' $<; \
-	 echo '  0 };'; \
-	 echo; \
-	 echo 'const char * const caml_names_of_builtin_cprim[] = {'; \
-	 sed -e 's/.*/  "&",/' $<; \
-	 echo '  0 };') > $@
+	$(V_GEN)cp runtime/primitives.new $@
+
+runtime/prims.c: runtime/gen_primsc.sh runtime/primitives
+	$(V_GEN)runtime/gen_primsc.sh \
+                    runtime/primitives $(runtime_BYTECODE_C_SOURCES) \
+                    > $@
 
 runtime/caml/opnames.h : runtime/caml/instruct.h
 	$(V_GEN)tr -d '\r' < $< | \

runtime/dune

@@ -13,7 +13,7 @@
 ;**************************************************************************
 
 (rule
- (targets primitives)
+ (targets primitives prims.c)
  (mode    fallback)
  (deps
    ; matches the line structure of files in gen_primitives.sh
@@ -25,8 +25,9 @@
      runtime_events.c sync.c
    dynlink.c backtrace_byt.c backtrace.c afl.c bigarray.c prng.c)
  (action
-   (chdir ..
-     (with-stdout-to %{targets} (run %{dep:gen_primitives.sh})))))
+   (progn
+     (with-stdout-to primitives (run %{dep:gen_primitives.sh} %{deps}))
+     (with-stdout-to prims.c (run %{dep:gen_primsc.sh} primitives %{deps})))))
 
 (rule
  (targets libcamlrun.a)

runtime/floats.c

@@ -1100,16 +1100,6 @@ CAMLprim value caml_signbit_float(value f)
   return caml_signbit(Double_val(f));
 }
 
-CAMLprim value caml_neq_float(value f, value g)
-{
-  return Val_bool(Double_val(f) != Double_val(g));
-}
-
-#define DEFINE_NAN_CMP(op) (value f, value g) \
-{ \
-  return Val_bool(Double_val(f) op Double_val(g)); \
-}
-
 intnat caml_float_compare_unboxed(double f, double g)
 {
   /* If one or both of f and g is NaN, order according to the convention
@@ -1124,11 +1114,15 @@ intnat caml_float_compare_unboxed(double f, double g)
   return res;
 }
 
-CAMLprim value caml_eq_float DEFINE_NAN_CMP(==)
-CAMLprim value caml_le_float DEFINE_NAN_CMP(<=)
-CAMLprim value caml_lt_float DEFINE_NAN_CMP(<)
-CAMLprim value caml_ge_float DEFINE_NAN_CMP(>=)
-CAMLprim value caml_gt_float DEFINE_NAN_CMP(>)
+#define FLOAT_CMP(op, f, g) \
+  return Val_bool(Double_val(f) op Double_val(g));
+
+CAMLprim value caml_neq_float(value f, value g) { FLOAT_CMP(!=, f, g) }
+CAMLprim value caml_eq_float(value f, value g) { FLOAT_CMP(==, f, g) }
+CAMLprim value caml_le_float(value f, value g) { FLOAT_CMP(<=, f, g) }
+CAMLprim value caml_lt_float(value f, value g) { FLOAT_CMP(<, f, g) }
+CAMLprim value caml_ge_float(value f, value g) { FLOAT_CMP(>=, f, g) }
+CAMLprim value caml_gt_float(value f, value g) { FLOAT_CMP(>, f, g) }
 
 CAMLprim value caml_float_compare(value vf, value vg)
 {

runtime/gen_primitives.sh

@@ -15,18 +15,29 @@
 #*                                                                        *
 #**************************************************************************
 
-# #8985: the meaning of character range a-z depends on the locale, so force C
-#        locale throughout.
+# If primitives contain duplicated lines (e.g. because the code is defined
+# like
+# #ifdef X
+# CAMLprim value caml_foo() ...
+# #else
+# CAMLprim value caml_foo() ...
+# #endif), horrible things will happen: duplicated entries in Runtimedef ->
+# double registration in Symtable -> empty entry in the PRIM table ->
+# the bytecode interpreter is confused.
+# We sort the primitive file and remove duplicates to avoid this problem.
+
+# Warning: we use "sort | uniq" instead of "sort -u" because in the MSVC
+# port, the "sort" program in the path is Microsoft's and not cygwin's
+
+# Warning: POSIX sort is locale dependent, that's why we set LC_ALL explicitly.
+# Sort is unstable for "is_directory" and "isatty"
+# see http://pubs.opengroup.org/onlinepubs/9699919799/utilities/sort.html:
+# "using sort to process pathnames, it is recommended that LC_ALL .. set to C"
+
+# #8985: in sed, the meaning of character range a-z depends on the locale,
+# so force C locale throughout.
+
 export LC_ALL=C
-(
-  for prim in \
-      alloc array compare extern floats gc_ctrl hash intern interp ints io \
-      lexing md5 meta memprof obj parsing signals str sys callback weak \
-      finalise domain platform fiber memory startup_aux runtime_events sync \
-      dynlink backtrace_byt backtrace afl \
-      bigarray prng
-  do
-      sed -n -e 's/^CAMLprim value \([a-z0-9_][a-z0-9_]*\).*/\1/p' \
-        "runtime/$prim.c"
-  done
-) | sort | uniq
+
+sed -n -e 's/^CAMLprim value \([a-z][a-z0-9_]*\).*$/\1/p' "$@" | \
+sort | uniq

runtime/gen_primsc.sh

@@ -0,0 +1,63 @@
+#!/bin/sh
+
+#**************************************************************************
+#*                                                                        *
+#*                                 OCaml                                  *
+#*                                                                        *
+#*            Xavier Leroy, Collège de France and Inria                   *
+#*                                                                        *
+#*   Copyright 2023 Institut National de Recherche en Informatique et     *
+#*     en Automatique.                                                    *
+#*                                                                        *
+#*   All rights reserved.  This file is distributed under the terms of    *
+#*   the GNU Lesser General Public License version 2.1, with the          *
+#*   special exception on linking described in the file LICENSE.          *
+#*                                                                        *
+#**************************************************************************
+
+# Build the runtime/prims.c file, with proper C declarations of the primitives
+
+export LC_ALL=C
+
+case $# in
+  0) echo "Usage: gen_primsc.sh <primitives file> <.c files>" 1>&2
+     exit 2;;
+  *) primitives="$1"; shift;;
+esac
+
+cat <<'EOF'
+/* Generated file, do not edit */
+
+#define CAML_INTERNALS
+#include "caml/mlvalues.h"
+#include "caml/prims.h"
+
+EOF
+
+# Extract the beginning of primitive definitions:
+# from 'CAMLprim' at beginning of line to the first closing parenthesis.
+# The first pattern below matches single-line definitions such as
+#    CAMLprim value foo(value x) {
+# The second pattern matches multi-line definitions such as
+#    CAMLprim value foo(value x,
+#                       value y)
+sed -n \
+  -e '/^CAMLprim value .*)/p' \
+  -e '/^CAMLprim value [^)]*$/,/)/p' \
+  "$@" |
+# Transform these definitions into "CAMLextern" declarations
+sed \
+  -e 's/^CAMLprim /CAMLextern /' \
+  -e 's/).*$/);/'
+
+# Generate the table of primitives
+echo
+echo 'const c_primitive caml_builtin_cprim[] = {'
+sed -e 's/.*/  (c_primitive) &,/' "$primitives"
+echo '  0 };'
+
+# Generate the table of primitive names
+echo
+echo 'const char * const caml_names_of_builtin_cprim[] = {'
+sed -e 's/.*/  "&",/' "$primitives"
+echo '  0 };'