Import ruby/prism@f0336b5

Author: eregon

src/main/c/yarp/include/prism.h

@@ -9,6 +9,7 @@
 #include "prism/defines.h"
 #include "prism/util/pm_buffer.h"
 #include "prism/util/pm_char.h"
+#include "prism/util/pm_integer.h"
 #include "prism/util/pm_memchr.h"
 #include "prism/util/pm_strncasecmp.h"
 #include "prism/util/pm_strpbrk.h"
@@ -24,6 +25,7 @@
 
 #include <assert.h>
 #include <errno.h>
+#include <math.h>
 #include <stdarg.h>
 #include <stdbool.h>
 #include <stdint.h>

src/main/c/yarp/include/prism/ast.h

@@ -15,6 +15,7 @@
 
 #include "prism/defines.h"
 #include "prism/util/pm_constant_pool.h"
+#include "prism/util/pm_integer.h"
 #include "prism/util/pm_string.h"
 
 #include <assert.h>
@@ -2841,6 +2842,13 @@ typedef struct pm_flip_flop_node {
 typedef struct pm_float_node {
     /** The embedded base node. */
     pm_node_t base;
+
+    /**
+     * FloatNode#value
+     *
+     * The value of the floating point number as a Float.
+     */
+    double value;
 } pm_float_node_t;
 
 /**
@@ -3712,6 +3720,13 @@ typedef struct pm_instance_variable_write_node {
 typedef struct pm_integer_node {
     /** The embedded base node. */
     pm_node_t base;
+
+    /**
+     * IntegerNode#value
+     *
+     * The value of the integer literal as a number.
+     */
+    pm_integer_t value;
 } pm_integer_node_t;
 
 /**

src/main/c/yarp/include/prism/defines.h

@@ -10,6 +10,7 @@
 #define PRISM_DEFINES_H
 
 #include <ctype.h>
+#include <math.h>
 #include <stdarg.h>
 #include <stddef.h>
 #include <stdint.h>
@@ -113,4 +114,15 @@
 #   endif
 #endif
 
+/**
+ * isinf on Windows is defined as accepting a float, but on POSIX systems it
+ * accepts a float, a double, or a long double. We want to mirror this behavior
+ * on windows.
+ */
+#ifdef _WIN32
+#   include <float.h>
+#   undef isinf
+#   define isinf(x) (sizeof(x) == sizeof(float) ? !_finitef(x) : !_finite(x))
+#endif
+
 #endif

src/main/c/yarp/include/prism/diagnostic.h

@@ -176,6 +176,7 @@ typedef enum {
     PM_ERR_EXPECT_STRING_CONTENT,
     PM_ERR_EXPECT_WHEN_DELIMITER,
     PM_ERR_EXPRESSION_BARE_HASH,
+    PM_ERR_FLOAT_PARSE,
     PM_ERR_FOR_COLLECTION,
     PM_ERR_FOR_IN,
     PM_ERR_FOR_INDEX,
@@ -305,6 +306,7 @@ typedef enum {
     PM_WARN_AMBIGUOUS_SLASH,
     PM_WARN_EQUAL_IN_CONDITIONAL,
     PM_WARN_END_IN_METHOD,
+    PM_WARN_FLOAT_OUT_OF_RANGE,
 
     // This is the number of diagnostic codes.
     PM_DIAGNOSTIC_ID_LEN,

src/main/c/yarp/include/prism/util/pm_buffer.h

@@ -129,6 +129,14 @@ void pm_buffer_append_varuint(pm_buffer_t *buffer, uint32_t value);
  */
 void pm_buffer_append_varsint(pm_buffer_t *buffer, int32_t value);
 
+/**
+ * Append a double to the buffer.
+ *
+ * @param buffer The buffer to append to.
+ * @param value The double to append.
+ */
+void pm_buffer_append_double(pm_buffer_t *buffer, double value);
+
 /**
  * The different types of escaping that can be performed by the buffer when
  * appending a slice of Ruby source code.

src/main/c/yarp/include/prism/util/pm_integer.h

@@ -0,0 +1,104 @@
+/**
+ * @file pm_integer.h
+ *
+ * This module provides functions for working with arbitrary-sized integers.
+ */
+#ifndef PRISM_NUMBER_H
+#define PRISM_NUMBER_H
+
+#include "prism/defines.h"
+
+#include <assert.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+/**
+ * A node in the linked list of a pm_integer_t.
+ */
+typedef struct pm_integer_word {
+    /** A pointer to the next node in the list. */
+    struct pm_integer_word *next;
+
+    /** The value of the node. */
+    uint32_t value;
+} pm_integer_word_t;
+
+/**
+ * This structure represents an arbitrary-sized integer. It is implemented as a
+ * linked list of 32-bit integers, with the least significant digit at the head
+ * of the list.
+ */
+typedef struct {
+    /** The number of nodes in the linked list that have been allocated. */
+    size_t length;
+
+    /**
+     * The head of the linked list, embedded directly so that allocations do not
+     * need to be performed for small integers.
+     */
+    pm_integer_word_t head;
+
+    /**
+     * Whether or not the integer is negative. It is stored this way so that a
+     * zeroed pm_integer_t is always positive zero.
+     */
+    bool negative;
+} pm_integer_t;
+
+/**
+ * An enum controlling the base of an integer. It is expected that the base is
+ * already known before parsing the integer, even though it could be derived
+ * from the string itself.
+ */
+typedef enum {
+    /** The binary base, indicated by a 0b or 0B prefix. */
+    PM_INTEGER_BASE_BINARY,
+
+    /** The octal base, indicated by a 0, 0o, or 0O prefix. */
+    PM_INTEGER_BASE_OCTAL,
+
+    /** The decimal base, indicated by a 0d, 0D, or empty prefix. */
+    PM_INTEGER_BASE_DECIMAL,
+
+    /** The hexidecimal base, indicated by a 0x or 0X prefix. */
+    PM_INTEGER_BASE_HEXADECIMAL,
+
+    /**
+     * An unknown base, in which case pm_integer_parse will derive it based on
+     * the content of the string. This is less efficient and does more
+     * comparisons, so if callers know the base ahead of time, they should use
+     * that instead.
+     */
+    PM_INTEGER_BASE_UNKNOWN
+} pm_integer_base_t;
+
+/**
+ * Parse an integer from a string. This assumes that the format of the integer
+ * has already been validated, as internal validation checks are not performed
+ * here.
+ *
+ * @param integer The integer to parse into.
+ * @param base The base of the integer.
+ * @param start The start of the string.
+ * @param end The end of the string.
+ */
+PRISM_EXPORTED_FUNCTION void pm_integer_parse(pm_integer_t *integer, pm_integer_base_t base, const uint8_t *start, const uint8_t *end);
+
+/**
+ * Return the memory size of the integer.
+ *
+ * @param integer The integer to get the memory size of.
+ * @return The size of the memory associated with the integer.
+ */
+size_t pm_integer_memsize(const pm_integer_t *integer);
+
+/**
+ * Free the internal memory of an integer. This memory will only be allocated if
+ * the integer exceeds the size of a single node in the linked list.
+ *
+ * @param integer The integer to free.
+ */
+PRISM_EXPORTED_FUNCTION void pm_integer_free(pm_integer_t *integer);
+
+#endif

src/main/c/yarp/src/diagnostic.c

@@ -179,6 +179,7 @@ static const pm_diagnostic_data_t diagnostic_messages[PM_DIAGNOSTIC_ID_LEN] = {
     [PM_ERR_EXPECT_STRING_CONTENT]              = { "expected string content after opening string delimiter", PM_ERROR_LEVEL_FATAL },
     [PM_ERR_EXPECT_WHEN_DELIMITER]              = { "expected a delimiter after the predicates of a `when` clause", PM_ERROR_LEVEL_FATAL },
     [PM_ERR_EXPRESSION_BARE_HASH]               = { "unexpected bare hash in expression", PM_ERROR_LEVEL_FATAL },
+    [PM_ERR_FLOAT_PARSE]                        = { "could not parse the float '%.*s'", PM_ERROR_LEVEL_FATAL },
     [PM_ERR_FOR_COLLECTION]                     = { "expected a collection after the `in` in a `for` statement", PM_ERROR_LEVEL_FATAL },
     [PM_ERR_FOR_INDEX]                          = { "expected an index after `for`", PM_ERROR_LEVEL_FATAL },
     [PM_ERR_FOR_IN]                             = { "expected an `in` after the index in a `for` statement", PM_ERROR_LEVEL_FATAL },
@@ -307,6 +308,7 @@ static const pm_diagnostic_data_t diagnostic_messages[PM_DIAGNOSTIC_ID_LEN] = {
     [PM_WARN_AMBIGUOUS_SLASH]                   = { "ambiguous `/`; wrap regexp in parentheses or add a space after `/` operator", PM_WARNING_LEVEL_VERBOSE },
     [PM_WARN_EQUAL_IN_CONDITIONAL]              = { "found `= literal' in conditional, should be ==", PM_WARNING_LEVEL_DEFAULT },
     [PM_WARN_END_IN_METHOD]                     = { "END in method; use at_exit", PM_WARNING_LEVEL_DEFAULT },
+    [PM_WARN_FLOAT_OUT_OF_RANGE]                = { "Float %.*s%s out of range", PM_WARNING_LEVEL_VERBOSE }
 };
 
 static inline const char *