[ruby/prism] Command line options as a bitset

https://github.com/ruby/prism/commit/369ffbd57e

7 files changed, 117 insertions, 143 deletions

diff --git a/lib/prism/ffi.rb b/lib/prism/ffi.rb
index 347c35d26f..2aecd4df86 100644
--- a/lib/prism/ffi.rb
+++ b/lib/prism/ffi.rb
@@ -313,6 +313,24 @@ module Prism
       LibRubyParser.pm_parse_success_p(string.pointer, string.length, dump_options(options))
     end
 
+    # Return the value that should be dumped for the command_line option.
+    def dump_options_command_line(options)
+      command_line = options.fetch(:command_line, "")
+      raise ArgumentError, "command_line must be a string" unless command_line.is_a?(String)
+
+      command_line.each_char.inject(0) do |value, char|
+        case char
+        when "a" then value | 0b000001
+        when "e" then value | 0b000010
+        when "l" then value | 0b000100
+        when "n" then value | 0b001000
+        when "p" then value | 0b010000
+        when "x" then value | 0b100000
+        else raise ArgumentError, "invalid command_line option: #{char}"
+        end
+      end
+    end
+
     # Convert the given options into a serialized options string.
     def dump_options(options)
       template = +""
@@ -342,16 +360,7 @@ module Prism
       values << (options.fetch(:frozen_string_literal, false) ? 1 : 0)
 
       template << "C"
-      values << (options.fetch(:command_line_p, false) ? 1 : 0)
-
-      template << "C"
-      values << (options.fetch(:command_line_n, false) ? 1 : 0)
-
-      template << "C"
-      values << (options.fetch(:command_line_l, false) ? 1 : 0)
-
-      template << "C"
-      values << (options.fetch(:command_line_a, false) ? 1 : 0)
+      values << dump_options_command_line(options)
 
       template << "C"
       values << { nil => 0, "3.3.0" => 1, "3.4.0" => 0, "latest" => 0 }.fetch(options[:version])
diff --git a/prism/extension.c b/prism/extension.c
index ed56a6c513..81a60e44a9 100644
--- a/prism/extension.c
+++ b/prism/extension.c
@@ -29,10 +29,7 @@ ID rb_option_id_line;
 ID rb_option_id_frozen_string_literal;
 ID rb_option_id_version;
 ID rb_option_id_scopes;
-ID rb_option_id_command_line_p;
-ID rb_option_id_command_line_n;
-ID rb_option_id_command_line_l;
-ID rb_option_id_command_line_a;
+ID rb_option_id_command_line;
 
 /******************************************************************************/
 /* IO of Ruby code                                                            */
@@ -148,21 +145,32 @@ build_options_i(VALUE key, VALUE value, VALUE argument) {
             const char *version = check_string(value);
 
             if (!pm_options_version_set(options, version, RSTRING_LEN(value))) {
-                rb_raise(rb_eArgError, "invalid version: %"PRIsVALUE, value);
+                rb_raise(rb_eArgError, "invalid version: %" PRIsVALUE, value);
             }
         }
     } else if (key_id == rb_option_id_scopes) {
         if (!NIL_P(value)) build_options_scopes(options, value);
-    } else if (key_id == rb_option_id_command_line_p) {
-        if (!NIL_P(value)) pm_options_command_line_p_set(options, value == Qtrue);
-    } else if (key_id == rb_option_id_command_line_n) {
-        if (!NIL_P(value)) pm_options_command_line_n_set(options, value == Qtrue);
-    } else if (key_id == rb_option_id_command_line_l) {
-        if (!NIL_P(value)) pm_options_command_line_l_set(options, value == Qtrue);
-    } else if (key_id == rb_option_id_command_line_a) {
-        if (!NIL_P(value)) pm_options_command_line_a_set(options, value == Qtrue);
+    } else if (key_id == rb_option_id_command_line) {
+        if (!NIL_P(value)) {
+            const char *string = check_string(value);
+            uint8_t command_line = 0;
+
+            for (size_t index = 0; index < strlen(string); index++) {
+                switch (string[index]) {
+                    case 'a': command_line |= PM_OPTIONS_COMMAND_LINE_A; break;
+                    case 'e': command_line |= PM_OPTIONS_COMMAND_LINE_E; break;
+                    case 'l': command_line |= PM_OPTIONS_COMMAND_LINE_L; break;
+                    case 'n': command_line |= PM_OPTIONS_COMMAND_LINE_N; break;
+                    case 'p': command_line |= PM_OPTIONS_COMMAND_LINE_P; break;
+                    case 'x': command_line |= PM_OPTIONS_COMMAND_LINE_X; break;
+                    default: rb_raise(rb_eArgError, "invalid command line flag: '%c'", string[index]); break;
+                }
+            }
+
+            pm_options_command_line_set(options, command_line);
+        }
     } else {
-        rb_raise(rb_eArgError, "unknown keyword: %"PRIsVALUE, key);
+        rb_raise(rb_eArgError, "unknown keyword: %" PRIsVALUE, key);
     }
 
     return ST_CONTINUE;
@@ -697,21 +705,25 @@ parse_input(pm_string_t *input, const pm_options_t *options) {
  * Parse the given string and return a ParseResult instance. The options that
  * are supported are:
  *
- * * `filepath` - the filepath of the source being parsed. This should be a
- *       string or nil
+ * * `command_line` - either nil or a string of the various options that were
+ *       set on the command line. Valid values are combinations of "a", "l",
+ *       "n", "p", and "x".
  * * `encoding` - the encoding of the source being parsed. This should be an
- *       encoding or nil
- * * `line` - the line number that the parse starts on. This should be an
- *       integer or nil. Note that this is 1-indexed.
+ *       encoding or nil.
+ * * `filepath` - the filepath of the source being parsed. This should be a
+ *       string or nil.
  * * `frozen_string_literal` - whether or not the frozen string literal pragma
  *       has been set. This should be a boolean or nil.
- * * `version` - the version of prism that should be used to parse Ruby code. By
- *       default prism assumes you want to parse with the latest vesion of
- *       prism (which you can trigger with `nil` or `"latest"`). If you want to
- *       parse exactly as CRuby 3.3.0 would, then you can pass `"3.3.0"`.
+ * * `line` - the line number that the parse starts on. This should be an
+ *       integer or nil. Note that this is 1-indexed.
  * * `scopes` - the locals that are in scope surrounding the code that is being
  *       parsed. This should be an array of arrays of symbols or nil. Scopes are
  *       ordered from the outermost scope to the innermost one.
+ * * `version` - the version of Ruby syntax that prism should used to parse Ruby
+ *       code. By default prism assumes you want to parse with the latest vesion
+ *       of Ruby syntax (which you can trigger with `nil` or `"latest"`). You
+ *       may also restrict the syntax to a specific version of Ruby. The
+ *       supported values are `"3.3.0"` and `"3.4.0"`.
  */
 static VALUE
 parse(int argc, VALUE *argv, VALUE self) {
@@ -1244,10 +1256,7 @@ Init_prism(void) {
     rb_option_id_frozen_string_literal = rb_intern_const("frozen_string_literal");
     rb_option_id_version = rb_intern_const("version");
     rb_option_id_scopes = rb_intern_const("scopes");
-    rb_option_id_command_line_p = rb_intern_const("command_line_p");
-    rb_option_id_command_line_n = rb_intern_const("command_line_n");
-    rb_option_id_command_line_l = rb_intern_const("command_line_l");
-    rb_option_id_command_line_a = rb_intern_const("command_line_a");
+    rb_option_id_command_line = rb_intern_const("command_line");
 
     /**
      * The version of the prism library.
diff --git a/prism/options.c b/prism/options.c
index 54fd1bc014..ca0f150107 100644
--- a/prism/options.c
+++ b/prism/options.c
@@ -33,35 +33,11 @@ pm_options_frozen_string_literal_set(pm_options_t *options, bool frozen_string_l
 }
 
 /**
- * Sets the -p command line option on the given options struct.
+ * Sets the command line option on the given options struct.
  */
 PRISM_EXPORTED_FUNCTION void
-pm_options_command_line_p_set(pm_options_t *options, bool command_line_p) {
-    options->command_line_p = command_line_p;
-}
-
-/**
- * Sets the -n command line option on the given options struct.
- */
-PRISM_EXPORTED_FUNCTION void
-pm_options_command_line_n_set(pm_options_t *options, bool command_line_n) {
-    options->command_line_n = command_line_n;
-}
-
-/**
- * Sets the -l command line option on the given options struct.
- */
-PRISM_EXPORTED_FUNCTION void
-pm_options_command_line_l_set(pm_options_t *options, bool command_line_l) {
-    options->command_line_l = command_line_l;
-}
-
-/**
- * Sets the -a command line option on the given options struct.
- */
-PRISM_EXPORTED_FUNCTION void
-pm_options_command_line_a_set(pm_options_t *options, bool command_line_a) {
-    options->command_line_a = command_line_a;
+pm_options_command_line_set(pm_options_t *options, uint8_t command_line) {
+    options->command_line = command_line;
 }
 
 /**
@@ -226,10 +202,7 @@ pm_options_read(pm_options_t *options, const char *data) {
     }
 
     options->frozen_string_literal = (*data++) ? true : false;
-    options->command_line_p = (*data++) ? true : false;
-    options->command_line_n = (*data++) ? true : false;
-    options->command_line_l = (*data++) ? true : false;
-    options->command_line_a = (*data++) ? true : false;
+    options->command_line = (uint8_t) *data++;
     options->version = (pm_options_version_t) *data++;
 
     uint32_t scopes_count = pm_options_read_u32(data);
diff --git a/prism/options.h b/prism/options.h
index 4a60fc266a..ce979656a5 100644
--- a/prism/options.h
+++ b/prism/options.h
@@ -76,21 +76,49 @@ typedef struct {
      */
     pm_options_version_t version;
 
+    /** A bitset of the various options that were set on the command line. */
+    uint8_t command_line;
+
     /** Whether or not the frozen string literal option has been set. */
     bool frozen_string_literal;
+} pm_options_t;
 
-    /** Whether or not the -p command line option has been set. */
-    bool command_line_p;
+/**
+ * A bit representing whether or not the command line -a option was set. -a
+ * splits the input line $_ into $F.
+ */
+static const uint8_t PM_OPTIONS_COMMAND_LINE_A = 0x1;
 
-    /** Whether or not the -n command line option has been set. */
-    bool command_line_n;
+/**
+ * A bit representing whether or not the command line -e option was set. -e
+ * allow the user to specify a script to be executed. This is necessary for
+ * prism to know because certain warnings are not generated when -e is used.
+ */
+static const uint8_t PM_OPTIONS_COMMAND_LINE_E = 0x2;
 
-    /** Whether or not the -l command line option has been set. */
-    bool command_line_l;
+/**
+ * A bit representing whether or not the command line -l option was set. -l
+ * chomps the input line by default.
+ */
+static const uint8_t PM_OPTIONS_COMMAND_LINE_L = 0x4;
 
-    /** Whether or not the -a command line option has been set. */
-    bool command_line_a;
-} pm_options_t;
+/**
+ * A bit representing whether or not the command line -n option was set. -n
+ * wraps the script in a while gets loop.
+ */
+static const uint8_t PM_OPTIONS_COMMAND_LINE_N = 0x8;
+
+/**
+ * A bit representing whether or not the command line -p option was set. -p
+ * prints the value of $_ at the end of each loop.
+ */
+static const uint8_t PM_OPTIONS_COMMAND_LINE_P = 0x10;
+
+/**
+ * A bit representing whether or not the command line -x option was set. -x
+ * searches the input file for a shebang that matches the current Ruby engine.
+ */
+static const uint8_t PM_OPTIONS_COMMAND_LINE_X = 0x20;
 
 /**
  * Set the filepath option on the given options struct.
@@ -125,36 +153,12 @@ PRISM_EXPORTED_FUNCTION void pm_options_encoding_set(pm_options_t *options, cons
 PRISM_EXPORTED_FUNCTION void pm_options_frozen_string_literal_set(pm_options_t *options, bool frozen_string_literal);
 
 /**
- * Sets the -p command line option on the given options struct.
- *
- * @param options The options struct to set the -p command line option on.
- * @param command_line_p The -p command line option to set.
- */
-PRISM_EXPORTED_FUNCTION void pm_options_command_line_p_set(pm_options_t *options, bool command_line_p);
-
-/**
- * Sets the -n command line option on the given options struct.
- *
- * @param options The options struct to set the -n command line option on.
- * @param command_line_n The -n command line option to set.
- */
-PRISM_EXPORTED_FUNCTION void pm_options_command_line_n_set(pm_options_t *options, bool command_line_n);
-
-/**
- * Sets the -l command line option on the given options struct.
- *
- * @param options The options struct to set the -l command line option on.
- * @param command_line_l The -l command line option to set.
- */
-PRISM_EXPORTED_FUNCTION void pm_options_command_line_l_set(pm_options_t *options, bool command_line_l);
-
-/**
- * Sets the -a command line option on the given options struct.
+ * Sets the command line option on the given options struct.
  *
- * @param options The options struct to set the -a command line option on.
- * @param command_line_a The -a command line option to set.
+ * @param options The options struct to set the command line option on.
+ * @param command_line The command_line value to set.
  */
-PRISM_EXPORTED_FUNCTION void pm_options_command_line_a_set(pm_options_t *options, bool command_line_a);
+PRISM_EXPORTED_FUNCTION void pm_options_command_line_set(pm_options_t *options, uint8_t command_line);
 
 /**
  * Set the version option on the given options struct by parsing the given
diff --git a/prism/parser.h b/prism/parser.h
index b790489d7d..80521e4ad9 100644
--- a/prism/parser.h
+++ b/prism/parser.h
@@ -703,6 +703,9 @@ struct pm_parser {
     /** The version of prism that we should use to parse. */
     pm_options_version_t version;
 
+    /** The command line flags given from the options. */
+    uint8_t command_line;
+
     /** Whether or not we're at the beginning of a command. */
     bool command_start;
 
@@ -736,30 +739,6 @@ struct pm_parser {
      * a true value.
      */
     bool frozen_string_literal;
-
-    /**
-     * Whether or not -p was present on the command line that invoked the
-     * parser. -p prints the value of $_ at the end of each loop.
-     */
-    bool command_line_p;
-
-    /**
-     * Whether or not -n was present on the command line that invoked the
-     * parser. -n wraps the script in a while gets loop.
-     */
-    bool command_line_n;
-
-    /**
-     * Whether or not -l was present on the command line that invoked the
-     * parser. -l chomps the input line by default.
-     */
-    bool command_line_l;
-
-    /**
-     * Whether or not -a was present on the command line that invoked the
-     * parser. -a splits the input line $_ into $F.
-     */
-    bool command_line_a;
 };
 
 #endif
diff --git a/prism/prism.c b/prism/prism.c
index d6a8d50680..7394786162 100644
--- a/prism/prism.c
+++ b/prism/prism.c
@@ -18273,7 +18273,7 @@ parse_expression(pm_parser_t *parser, pm_binding_power_t binding_power, bool acc
  */
 static pm_statements_node_t *
 wrap_statements(pm_parser_t *parser, pm_statements_node_t *statements) {
-    if (parser->command_line_p) {
+    if (parser->command_line & PM_OPTIONS_COMMAND_LINE_P) {
         pm_arguments_node_t *arguments = pm_arguments_node_create(parser);
         pm_arguments_node_arguments_append(
             arguments,
@@ -18287,8 +18287,8 @@ wrap_statements(pm_parser_t *parser, pm_statements_node_t *statements) {
         ));
     }
 
-    if (parser->command_line_n) {
-        if (parser->command_line_a) {
+    if (parser->command_line & PM_OPTIONS_COMMAND_LINE_N) {
+        if (parser->command_line & PM_OPTIONS_COMMAND_LINE_A) {
             pm_arguments_node_t *arguments = pm_arguments_node_create(parser);
             pm_arguments_node_arguments_append(
                 arguments,
@@ -18313,7 +18313,7 @@ wrap_statements(pm_parser_t *parser, pm_statements_node_t *statements) {
             (pm_node_t *) pm_global_variable_read_node_synthesized_create(parser, pm_parser_constant_id_constant(parser, "$/", 2))
         );
 
-        if (parser->command_line_l) {
+        if (parser->command_line & PM_OPTIONS_COMMAND_LINE_L) {
             pm_keyword_hash_node_t *keywords = pm_keyword_hash_node_create(parser);
             pm_keyword_hash_node_elements_append(keywords, (pm_node_t *) pm_assoc_node_create(
                 parser,
@@ -18367,7 +18367,7 @@ parse_program(pm_parser_t *parser) {
 
     // At the top level, see if we need to wrap the statements in a program
     // node with a while loop based on the options.
-    if (parser->command_line_p || parser->command_line_n) {
+    if (parser->command_line & (PM_OPTIONS_COMMAND_LINE_P | PM_OPTIONS_COMMAND_LINE_N)) {
         statements = wrap_statements(parser, statements);
     }
 
@@ -18421,6 +18421,7 @@ pm_parser_init(pm_parser_t *parser, const uint8_t *source, size_t size, const pm
         .current_string = PM_STRING_EMPTY,
         .start_line = 1,
         .explicit_encoding = NULL,
+        .command_line = 0,
         .command_start = true,
         .recovering = false,
         .encoding_changed = false,
@@ -18428,11 +18429,7 @@ pm_parser_init(pm_parser_t *parser, const uint8_t *source, size_t size, const pm
         .in_keyword_arg = false,
         .current_param_name = 0,
         .semantic_token_seen = false,
-        .frozen_string_literal = false,
-        .command_line_p = options != NULL && options->command_line_p,
-        .command_line_n = options != NULL && options->command_line_n,
-        .command_line_l = options != NULL && options->command_line_l,
-        .command_line_a = options != NULL && options->command_line_a
+        .frozen_string_literal = false
     };
 
     // Initialize the constant pool. We're going to completely guess as to the
@@ -18478,6 +18475,9 @@ pm_parser_init(pm_parser_t *parser, const uint8_t *source, size_t size, const pm
             parser->frozen_string_literal = true;
         }
 
+        // command_line option
+        parser->command_line = options->command_line;
+
         // version option
         parser->version = options->version;
 
diff --git a/test/prism/command_line_test.rb b/test/prism/command_line_test.rb
index f83110a612..f5438bd0ad 100644
--- a/test/prism/command_line_test.rb
+++ b/test/prism/command_line_test.rb
@@ -5,7 +5,7 @@ require_relative "test_helper"
 module Prism
   class CommandLineTest < TestCase
     def test_command_line_p
-      program = Prism.parse("1", command_line_p: true).value
+      program = Prism.parse("1", command_line: "p").value
       statements = program.statements.body
 
       assert_equal 2, statements.length
@@ -14,7 +14,7 @@ module Prism
     end
 
     def test_command_line_n
-      program = Prism.parse("1", command_line_n: true).value
+      program = Prism.parse("1", command_line: "n").value
       statements = program.statements.body
 
       assert_equal 1, statements.length
@@ -30,7 +30,7 @@ module Prism
     end
 
     def test_command_line_a
-      program = Prism.parse("1", command_line_n: true, command_line_a: true).value
+      program = Prism.parse("1", command_line: "na").value
       statements = program.statements.body
 
       assert_equal 1, statements.length
@@ -42,7 +42,7 @@ module Prism
     end
 
     def test_command_line_l
-      program = Prism.parse("1", command_line_n: true, command_line_l: true).value
+      program = Prism.parse("1", command_line: "nl").value
       statements = program.statements.body
 
       assert_equal 1, statements.length