From 2f944c9a4aec933ba64b3837ee18321c7e5f69b4 Mon Sep 17 00:00:00 2001
From: Xavier Leroy <xavier.leroy@college-de-france.fr>
Date: Sat, 28 Oct 2023 19:03:49 +0200
Subject: [PATCH] Change runtime/gen_primitives.sh to take the list of .c files
 as argument

This avoids writing the list twice, in the top-level Makefile and in the script.
The Dune build needs updating accordingly.

Also: move relevant comments from Makefile to runtime/gen_primitives.sh .
---
 Makefile                  | 22 ++--------------------
 runtime/gen_primitives.sh | 39 +++++++++++++++++++++++++--------------
 2 files changed, 27 insertions(+), 34 deletions(-)

diff --git a/Makefile b/Makefile
index 514996694db..d1bb156207d 100644
--- a/Makefile
+++ b/Makefile
@@ -1251,29 +1251,11 @@ 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; \
+  $(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 $^ $@
diff --git a/runtime/gen_primitives.sh b/runtime/gen_primitives.sh
index 9979732f38c..0fe676d496d 100755
--- a/runtime/gen_primitives.sh
+++ b/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