Make rb_scan_args handle keywords more similar to Ruby methods (#2460)

Cfuncs that use rb_scan_args with the : entry suffer similar keyword
argument separation issues that Ruby methods suffer if the cfuncs
accept optional or variable arguments.

This makes the following changes to : handling.

* Treats as **kw, prompting keyword argument separation warnings
  if called with a positional hash.

* Do not look for an option hash if empty keywords are provided.
  For backwards compatibility, treat an empty keyword splat as a empty
  mandatory positional hash argument, but emit a a warning, as this
  behavior will be removed in Ruby 3.  The argument number check
  needs to be moved lower so it can correctly handle an empty
  positional argument being added.

* If the last argument is nil and it is necessary to treat it as an option
  hash in order to make sure all arguments are processed, continue to
  treat the last argument as the option hash. Emit a warning in this case,
  as this behavior will be removed in Ruby 3.

* If splitting the keyword hash into two hashes, issue a warning, as we
  will not be splitting hashes in Ruby 3.

* If the keyword argument is required to fill a mandatory positional
  argument, continue to do so, but emit a warning as this behavior will
  be going away in Ruby 3.

* If keyword arguments are provided and the last argument is not a hash,
  that indicates something wrong. This can happen if a cfunc is calling
  rb_scan_args multiple times, and providing arguments that were not
  passed to it from Ruby.  Callers need to switch to the new
  rb_scan_args_kw function, which allows passing of whether keywords
  were provided.

This commit fixes all warnings caused by the changes above.

It switches some function calls to *_kw versions with appropriate
kw_splat flags. If delegating arguments, RB_PASS_CALLED_KEYWORDS
is used.  If creating new arguments, RB_PASS_KEYWORDS is used if
the last argument is a hash to be treated as keywords.

In open_key_args in io.c, use rb_scan_args_kw.
In this case, the arguments provided come from another C
function, not Ruby.  The last argument may or may not be a hash,
so we can't set keyword argument mode.  However, if it is a
hash, we don't want to warn when treating it as keywords.

In Ruby files, make sure to appropriately use keyword splats
or literal keywords when calling Cfuncs that now issue keyword
argument separation warnings through rb_scan_args.  Also, make
sure not to pass nil in place of an option hash.

Work around Kernel#warn warnings due to problems in the Rubygems
override of the method.  There is an open pull request to fix
these issues in Rubygems, but part of the Rubygems tests for
their override fail on ruby-head due to rb_scan_args not
recognizing empty keyword splats, which this commit fixes.

Implementation wise, adding rb_scan_args_kw is kind of a pain,
because rb_scan_args takes a variable number of arguments.
In order to not duplicate all the code, the function internals need
to be split into two functions taking a va_list, and to avoid passing
in a ton of arguments, a single struct argument is used to handle
the variables previously local to the function.

1 files changed, 186 insertions, 60 deletions

diff --git a/class.c b/class.c
index 08e053967a..d0ee7a68fb 100644
--- a/class.c
+++ b/class.c
@@ -1943,85 +1943,172 @@ rb_get_kwargs(VALUE keyword_hash, const ID *table, int required, int optional, V
 #undef extract_kwarg
 }
 
-#undef rb_scan_args
-int
-rb_scan_args(int argc, const VALUE *argv, const char *fmt, ...)
+struct rb_scan_args_t {
+    int argc;
+    const VALUE *argv;
+    va_list vargs;
+    int f_var;
+    int f_hash;
+    int f_block;
+    int n_lead;
+    int n_opt;
+    int n_trail;
+    int n_mand;
+    int argi;
+    int last_idx;
+    VALUE hash;
+    VALUE last_hash;
+    VALUE *tmp_buffer;
+};
+
+static void
+rb_scan_args_parse(int kw_flag, int argc, const VALUE *argv, const char *fmt, struct rb_scan_args_t *arg)
 {
-    int i;
     const char *p = fmt;
-    VALUE *var;
-    va_list vargs;
-    int f_var = 0, f_hash = 0, f_block = 0;
-    int n_lead = 0, n_opt = 0, n_trail = 0, n_mand;
-    int argi = 0, last_idx = -1;
-    VALUE hash = Qnil, last_hash = 0;
+    VALUE *tmp_buffer = arg->tmp_buffer;
+    int keyword_given = 0;
+    int empty_keyword_given = 0;
+    int last_hash_keyword = 0;
+
+    memset(arg, 0, sizeof(*arg));
+    arg->last_idx = -1;
+    arg->hash = Qnil;
+
+    switch(kw_flag) {
+      case RB_SCAN_ARGS_PASS_CALLED_KEYWORDS:
+        if(!(keyword_given = rb_keyword_given_p())) {
+            empty_keyword_given = rb_empty_keyword_given_p();
+        }
+        break;
+      case RB_SCAN_ARGS_KEYWORDS:
+        keyword_given = 1;
+        break;
+      case RB_SCAN_ARGS_EMPTY_KEYWORDS:
+        empty_keyword_given = 1;
+        break;
+      case RB_SCAN_ARGS_LAST_HASH_KEYWORDS:
+        last_hash_keyword = 1;
+        break;
+    }
 
     if (ISDIGIT(*p)) {
-	n_lead = *p - '0';
+        arg->n_lead = *p - '0';
 	p++;
 	if (ISDIGIT(*p)) {
-	    n_opt = *p - '0';
+            arg->n_opt = *p - '0';
 	    p++;
 	}
     }
     if (*p == '*') {
-	f_var = 1;
+        arg->f_var = 1;
 	p++;
     }
     if (ISDIGIT(*p)) {
-	n_trail = *p - '0';
+        arg->n_trail = *p - '0';
 	p++;
     }
     if (*p == ':') {
-	f_hash = 1;
+        arg->f_hash = 1;
 	p++;
     }
     if (*p == '&') {
-	f_block = 1;
+        arg->f_block = 1;
 	p++;
     }
     if (*p != '\0') {
 	rb_fatal("bad scan arg format: %s", fmt);
     }
-    n_mand = n_lead + n_trail;
+    arg->n_mand = arg->n_lead + arg->n_trail;
 
-    if (argc < n_mand)
-	goto argc_error;
+    /* capture an option hash - phase 1: pop */
+    /* Ignore final positional hash if empty keywords given */
+    if (argc > 0 && !(arg->f_hash && empty_keyword_given)) {
+        VALUE last = argv[argc - 1];
+
+        if (arg->f_hash && arg->n_mand < argc) {
+            if (keyword_given) {
+                if (!RB_TYPE_P(last, T_HASH)) {
+                    rb_warn("Keyword flag set when calling rb_scan_args, but last entry is not a hash");
+                }
+                else {
+                    arg->hash = last;
+                }
+            }
+            else if (NIL_P(last)) {
+                /* For backwards compatibility, nil is taken as an empty
+                   option hash only if it is not ambiguous; i.e. '*' is
+                   not specified and arguments are given more than sufficient.
+                   This will be removed in Ruby 3. */
+                if (!arg->f_var && arg->n_mand + arg->n_opt < argc) {
+                    rb_warn("The last argument is nil, treating as empty keywords");
+                    argc--;
+                }
+            }
+            else {
+                arg->hash = rb_check_hash_type(last);
+            }
+
+            /* Ruby 3: Remove if branch, as it will not attempt to split hashes */
+            if (!NIL_P(arg->hash)) {
+                VALUE opts = rb_extract_keywords(&arg->hash);
+
+                if (!(arg->last_hash = arg->hash)) {
+                    if (!keyword_given && !last_hash_keyword) {
+                        /* Warn if treating positional as keyword, as in Ruby 3,
+                           this will be an error */
+                        rb_warn("The last argument is used as the keyword parameter");
+                    }
+                    argc--;
+                }
+                else {
+                    /* Warn if splitting either positional hash to keywords or keywords
+                       to positional hash, as in Ruby 3, no splitting will be done */
+                    rb_warn("The last argument is split into positional and keyword parameters");
+                    arg->last_idx = argc - 1;
+                }
+                arg->hash = opts ? opts : Qnil;
+            }
+        }
+        else if (arg->f_hash && keyword_given && arg->n_mand == argc) {
+            /* Warn if treating keywords as positional, as in Ruby 3, this will be an error */
+            rb_warn("The keyword argument is passed as the last hash parameter");
+        }
+    }
+    if (arg->f_hash && arg->n_mand == argc+1 && empty_keyword_given) {
+        VALUE *ptr = rb_alloc_tmp_buffer2(tmp_buffer, argc+1, sizeof(VALUE));
+        memcpy(ptr, argv, sizeof(VALUE)*argc);
+        ptr[argc] = rb_hash_new();
+        argc++;
+        *(&argv) = ptr;
+        rb_warn("The keyword argument is passed as the last hash parameter");
+    }
 
-    va_start(vargs, fmt);
+    arg->argc = argc;
+    arg->argv = argv;
+}
 
-    /* capture an option hash - phase 1: pop */
-    if (f_hash && n_mand < argc) {
-	VALUE last = argv[argc - 1];
-
-	if (NIL_P(last)) {
-	    /* nil is taken as an empty option hash only if it is not
-	       ambiguous; i.e. '*' is not specified and arguments are
-	       given more than sufficient */
-	    if (!f_var && n_mand + n_opt < argc)
-		argc--;
-	}
-	else {
-	    hash = rb_check_hash_type(last);
-	    if (!NIL_P(hash)) {
-		VALUE opts = rb_extract_keywords(&hash);
-		if (!(last_hash = hash)) argc--;
-		else last_idx = argc - 1;
-		hash = opts ? opts : Qnil;
-	    }
-	}
+static int
+rb_scan_args_assign(struct rb_scan_args_t *arg, va_list vargs)
+{
+    int argi = 0;
+    int i;
+    VALUE *var;
+
+    if (arg->argc < arg->n_mand) {
+        return 1;
     }
+
     /* capture leading mandatory arguments */
-    for (i = n_lead; i-- > 0; ) {
+    for (i = arg->n_lead; i-- > 0; ) {
 	var = va_arg(vargs, VALUE *);
-	if (var) *var = (argi == last_idx) ? last_hash : argv[argi];
+        if (var) *var = (argi == arg->last_idx) ? arg->last_hash : arg->argv[argi];
 	argi++;
     }
     /* capture optional arguments */
-    for (i = n_opt; i-- > 0; ) {
+    for (i = arg->n_opt; i-- > 0; ) {
 	var = va_arg(vargs, VALUE *);
-	if (argi < argc - n_trail) {
-	    if (var) *var = (argi == last_idx) ? last_hash : argv[argi];
+        if (argi < arg->argc - arg->n_trail) {
+            if (var) *var = (argi == arg->last_idx) ? arg->last_hash : arg->argv[argi];
 	    argi++;
 	}
 	else {
@@ -2029,15 +2116,15 @@ rb_scan_args(int argc, const VALUE *argv, const char *fmt, ...)
 	}
     }
     /* capture variable length arguments */
-    if (f_var) {
-	int n_var = argc - argi - n_trail;
+    if (arg->f_var) {
+        int n_var = arg->argc - argi - arg->n_trail;
 
 	var = va_arg(vargs, VALUE *);
 	if (0 < n_var) {
 	    if (var) {
-		int f_last = (last_idx + 1 == argc - n_trail);
-		*var = rb_ary_new4(n_var-f_last, &argv[argi]);
-		if (f_last) rb_ary_push(*var, last_hash);
+                int f_last = (arg->last_idx + 1 == arg->argc - arg->n_trail);
+                *var = rb_ary_new4(n_var - f_last, &arg->argv[argi]);
+                if (f_last) rb_ary_push(*var, arg->last_hash);
 	    }
 	    argi += n_var;
 	}
@@ -2046,18 +2133,18 @@ rb_scan_args(int argc, const VALUE *argv, const char *fmt, ...)
 	}
     }
     /* capture trailing mandatory arguments */
-    for (i = n_trail; i-- > 0; ) {
+    for (i = arg->n_trail; i-- > 0; ) {
 	var = va_arg(vargs, VALUE *);
-	if (var) *var = (argi == last_idx) ? last_hash : argv[argi];
+        if (var) *var = (argi == arg->last_idx) ? arg->last_hash : arg->argv[argi];
 	argi++;
     }
     /* capture an option hash - phase 2: assignment */
-    if (f_hash) {
+    if (arg->f_hash) {
 	var = va_arg(vargs, VALUE *);
-	if (var) *var = hash;
+        if (var) *var = arg->hash;
     }
     /* capture iterator block */
-    if (f_block) {
+    if (arg->f_block) {
 	var = va_arg(vargs, VALUE *);
 	if (rb_block_given_p()) {
 	    *var = rb_block_proc();
@@ -2066,14 +2153,53 @@ rb_scan_args(int argc, const VALUE *argv, const char *fmt, ...)
 	    *var = Qnil;
 	}
     }
-    va_end(vargs);
 
-    if (argi < argc) {
-      argc_error:
-	rb_error_arity(argc, n_mand, f_var ? UNLIMITED_ARGUMENTS : n_mand + n_opt);
+    if (argi < arg->argc) return 1;
+
+    return 0;
+}
+
+#undef rb_scan_args
+int
+rb_scan_args(int argc, const VALUE *argv, const char *fmt, ...)
+{
+    int error;
+    va_list vargs;
+    VALUE tmp_buffer = 0;
+    struct rb_scan_args_t arg;
+    arg.tmp_buffer = &tmp_buffer;
+    rb_scan_args_parse(RB_SCAN_ARGS_PASS_CALLED_KEYWORDS, argc, argv, fmt, &arg);
+    va_start(vargs,fmt);
+    error = rb_scan_args_assign(&arg, vargs);
+    va_end(vargs);
+    if (tmp_buffer) {
+        rb_free_tmp_buffer(&tmp_buffer);
     }
+    if (error) {
+        rb_error_arity(arg.argc, arg.n_mand, arg.f_var ? UNLIMITED_ARGUMENTS : arg.n_mand + arg.n_opt);
+    }
+    return arg.argc;
+}
 
-    return argc;
+int
+rb_scan_args_kw(int kw_flag, int argc, const VALUE *argv, const char *fmt, ...)
+{
+    int error;
+    va_list vargs;
+    VALUE tmp_buffer = 0;
+    struct rb_scan_args_t arg;
+    arg.tmp_buffer = &tmp_buffer;
+    rb_scan_args_parse(kw_flag, argc, argv, fmt, &arg);
+    va_start(vargs,fmt);
+    error = rb_scan_args_assign(&arg, vargs);
+    va_end(vargs);
+    if (tmp_buffer) {
+        rb_free_tmp_buffer(&tmp_buffer);
+    }
+    if (error) {
+        rb_error_arity(arg.argc, arg.n_mand, arg.f_var ? UNLIMITED_ARGUMENTS : arg.n_mand + arg.n_opt);
+    }
+    return arg.argc;
 }
 
 int