Skip to content

Preserve native wrapper state #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
uvlad7 merged 9 commits into from
Aug 30, 2026
Merged

Preserve native wrapper state #4

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
adfc93e
Preserve native wrapper state
uvlad7 Aug 29, 2026
a48892c
Relocate copied selector pointers directly
uvlad7 Aug 29, 2026
54c57e3
Move internal includes to the top
uvlad7 Aug 30, 2026
c05040b
Copy selector context before relocation
uvlad7 Aug 30, 2026
063737d
Keep selector arena copying scoped
uvlad7 Aug 30, 2026
dd1d122
Guard uninitialized native wrappers
uvlad7 Aug 30, 2026
cdb0d6f
Keep failed selector copies uninitialized
uvlad7 Aug 30, 2026
65c8ded
Document selector zero initialization
uvlad7 Aug 30, 2026
48b7efa
Simplify options allocation
uvlad7 Aug 30, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
107 changes: 6 additions & 101 deletions ext/json_scanner/json_scanner.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,15 +1,12 @@
#include "json_scanner.h"
#include "scan_ctx.h"
#include "options.h"
#include "selector.h"

VALUE rb_mJsonScanner;
VALUE rb_cJsonScannerSelector;
VALUE rb_cJsonScannerOptions;
VALUE rb_eJsonScannerParseError;
#define BYTES_CONSUMED "bytes_consumed"
ID rb_iv_bytes_consumed;
#define SCAN_KWARGS_SIZE 9
ID scan_kwargs_table[SCAN_KWARGS_SIZE];

#include "options.h"

VALUE null_sym;
VALUE boolean_sym;
Expand All @@ -20,89 +17,6 @@ VALUE array_sym;

VALUE any_key_sym;

enum matcher_type
{
MATCHER_KEY,
MATCHER_INDEX,
MATCHER_ANY_KEY,
MATCHER_INDEX_RANGE,
// MATCHER_KEYS_LIST,
// MATCHER_KEY_REGEX,
};

enum path_type
{
PATH_KEY,
PATH_INDEX,
};

typedef struct
{
const char *val;
size_t len;
} hashkey_t;

typedef struct
{
const char *val;
size_t len;
int owned;
} path_key_t;

typedef struct
{
long start;
long end;
} range_t;

typedef struct
{
enum matcher_type type;
union
{
hashkey_t key;
long index;
range_t range;
} value;
} path_matcher_elem_t;

typedef struct
{
enum path_type type;
union
{
path_key_t key;
long index;
} value;
} path_elem_t;

typedef struct
{
path_matcher_elem_t *elems;
int len;
int matched_depth;
} paths_t;

typedef struct
{
int with_path;
int symbolize_path_keys;
int paths_len;
paths_t *paths;
int current_path_len;
int max_path_len;
path_elem_t *current_path;
// Easier to use a Ruby array for result than convert later
// must be supplied by the caller and RB_GC_GUARD-ed if it isn't on the stack
VALUE points_list;
VALUE roots_info_list;
// by depth
size_t *starts;
yajl_handle handle;
size_t yajl_bytes_consumed;
const unsigned char *json_text;
size_t json_text_len;
} scan_ctx;

static inline size_t scan_ctx_get_bytes_consumed(scan_ctx *ctx)
{
Expand Down Expand Up @@ -323,6 +237,7 @@ static void scan_ctx_init(scan_ctx *ctx, VALUE path_ary)
// Assign ctx->paths early so ruby_xfree(ctx->paths) will free the arena
// if a Ruby exception happens during the population loop below
ctx->paths = paths;
ctx->paths_arena_size = arena_size;
ctx->paths_len = 0;
ctx->current_path = NULL;
ctx->starts = NULL;
Expand Down Expand Up @@ -740,9 +655,6 @@ static int scan_on_end_array(void *ctx)
return true;
}

#include "selector.h"


static yajl_callbacks scan_callbacks = {
scan_on_null,
scan_on_boolean,
Expand Down Expand Up @@ -860,6 +772,8 @@ static VALUE scan(int argc, VALUE *argv, VALUE self)
{
free_ctx = false;
TypedData_Get_Struct(path_ary, scan_ctx, &selector_type, ctx);
if (!ctx->paths)
rb_raise(rb_eRuntimeError, "selector is not initialized");
}
else
{
Expand Down Expand Up @@ -955,13 +869,4 @@ Init_json_scanner(void)
string_sym = rb_id2sym(rb_intern("string"));
object_sym = rb_id2sym(rb_intern("object"));
array_sym = rb_id2sym(rb_intern("array"));
scan_kwargs_table[0] = rb_intern("with_path");
scan_kwargs_table[1] = rb_intern("verbose_error");
scan_kwargs_table[2] = rb_intern("allow_comments");
scan_kwargs_table[3] = rb_intern("dont_validate_strings");
scan_kwargs_table[4] = rb_intern("allow_trailing_garbage");
scan_kwargs_table[5] = rb_intern("allow_multiple_values");
scan_kwargs_table[6] = rb_intern("allow_partial_values");
scan_kwargs_table[7] = rb_intern("symbolize_path_keys");
scan_kwargs_table[8] = rb_intern("with_roots_info");
}
125 changes: 88 additions & 37 deletions ext/json_scanner/options.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,47 +13,47 @@ typedef struct
int symbolize_path_keys;
int with_roots_info;
} scan_options;

#define SCAN_OPTION_VALUE_MASK 1
#define SCAN_OPTION_SET_MASK (1 << 1)
#define SCAN_OPTION(options, field) ((options)->field & SCAN_OPTION_VALUE_MASK)
#define SCAN_OPTION_IS_SET(options, field) ((options)->field & SCAN_OPTION_SET_MASK)
#define SCAN_OPTION_SET(options, field, value) ((options)->field = ((value) & SCAN_OPTION_VALUE_MASK) | SCAN_OPTION_SET_MASK)
#define SCAN_OPTION_FALSE(options, field) \
(!SCAN_OPTION(options, field) && ((options)->field & SCAN_OPTION_SET_MASK))
#define OPTIONS_FIELDS(X) \
X(with_path) \
X(verbose_error) \
X(allow_comments) \
X(dont_validate_strings) \
X(allow_trailing_garbage) \
X(allow_multiple_values) \
X(allow_partial_values) \
X(symbolize_path_keys) \
X(with_roots_info)

static VALUE rb_cJsonScannerOptions;
enum
{
#define SCAN_KWARG_ENUM(field) SCAN_KWARG_##field,
OPTIONS_FIELDS(SCAN_KWARG_ENUM)
#undef SCAN_KWARG_ENUM
SCAN_KWARGS_SIZE,
};
static ID scan_kwargs_table[SCAN_KWARGS_SIZE];

static void scan_options_init(scan_options *options, VALUE kwargs)
{
options->with_path = 0;
options->verbose_error = 0;
options->allow_comments = 0;
options->dont_validate_strings = 0;
options->allow_trailing_garbage = 0;
options->allow_multiple_values = 0;
options->allow_partial_values = 0;
options->symbolize_path_keys = 0;
options->with_roots_info = 0;
memset(options, 0, sizeof(*options));
if (kwargs != Qnil)
{
VALUE kwargs_values[SCAN_KWARGS_SIZE];
rb_get_kwargs(kwargs, scan_kwargs_table, 0, SCAN_KWARGS_SIZE, kwargs_values);
if (kwargs_values[0] != Qundef)
SCAN_OPTION_SET(options, with_path, RTEST(kwargs_values[0]));
if (kwargs_values[1] != Qundef)
SCAN_OPTION_SET(options, verbose_error, RTEST(kwargs_values[1]));
if (kwargs_values[2] != Qundef)
SCAN_OPTION_SET(options, allow_comments, RTEST(kwargs_values[2]));
if (kwargs_values[3] != Qundef)
SCAN_OPTION_SET(options, dont_validate_strings, RTEST(kwargs_values[3]));
if (kwargs_values[4] != Qundef)
SCAN_OPTION_SET(options, allow_trailing_garbage, RTEST(kwargs_values[4]));
if (kwargs_values[5] != Qundef)
SCAN_OPTION_SET(options, allow_multiple_values, RTEST(kwargs_values[5]));
if (kwargs_values[6] != Qundef)
SCAN_OPTION_SET(options, allow_partial_values, RTEST(kwargs_values[6]));
if (kwargs_values[7] != Qundef)
SCAN_OPTION_SET(options, symbolize_path_keys, RTEST(kwargs_values[7]));
if (kwargs_values[8] != Qundef)
SCAN_OPTION_SET(options, with_roots_info, RTEST(kwargs_values[8]));
#define SCAN_OPTION_SET_KWARG(field) \
if (kwargs_values[SCAN_KWARG_##field] != Qundef) \
SCAN_OPTION_SET(options, field, RTEST(kwargs_values[SCAN_KWARG_##field]));
OPTIONS_FIELDS(SCAN_OPTION_SET_KWARG)
#undef SCAN_OPTION_SET_KWARG
}
}

Expand All @@ -68,12 +68,17 @@ static const rb_data_type_t options_type = {
.dfree = RUBY_DEFAULT_FREE,
.dsize = options_size,
},
.flags = RUBY_TYPED_FREE_IMMEDIATELY,
.flags = RUBY_TYPED_FREE_IMMEDIATELY
#ifdef RUBY_TYPED_FROZEN_SHAREABLE
| RUBY_TYPED_FROZEN_SHAREABLE
#endif
,
};

static VALUE options_alloc(VALUE self)
{
scan_options *options;
// TypedData_Make_Struct zeroes options.
return TypedData_Make_Struct(self, scan_options, &options_type, options);
}

Expand All @@ -91,6 +96,50 @@ static VALUE options_m_initialize(int argc, VALUE *argv, VALUE self)
return self;
}

static VALUE options_m_initialize_copy(VALUE self, VALUE other)
{
scan_options *options, *other_options;
rb_call_super(1, &other);
TypedData_Get_Struct(self, scan_options, &options_type, options);
TypedData_Get_Struct(other, scan_options, &options_type, other_options);
*options = *other_options;
return self;
}

#define DEFINE_OPTIONS_ACCESSORS(field) \
static VALUE options_m_##field(VALUE self) \
{ \
scan_options *options; \
TypedData_Get_Struct(self, scan_options, &options_type, options); \
if (!SCAN_OPTION_IS_SET(options, field)) \
return Qnil; \
return SCAN_OPTION(options, field) ? Qtrue : Qfalse; \
} \
static VALUE options_m_##field##_set(VALUE self, VALUE value) \
{ \
scan_options *options; \
rb_check_frozen(self); \
TypedData_Get_Struct(self, scan_options, &options_type, options); \
SCAN_OPTION_SET(options, field, RTEST(value)); \
return value; \
}
OPTIONS_FIELDS(DEFINE_OPTIONS_ACCESSORS)

static VALUE options_m_equal(VALUE self, VALUE other)
{
scan_options *options, *other_options;
if (!rb_obj_is_kind_of(other, rb_cJsonScannerOptions))
return Qfalse;
TypedData_Get_Struct(self, scan_options, &options_type, options);
TypedData_Get_Struct(other, scan_options, &options_type, other_options);
#define OPTIONS_EQUAL(field) \
if (options->field != other_options->field) \
return Qfalse;
OPTIONS_FIELDS(OPTIONS_EQUAL)
#undef OPTIONS_EQUAL
return Qtrue;
}

static VALUE options_m_inspect(VALUE self)
{
VALUE res;
Expand All @@ -100,15 +149,7 @@ static VALUE options_m_inspect(VALUE self)
#define APPEND_OPTION_INSPECT(field) \
if (SCAN_OPTION_IS_SET(options, field)) \
rb_str_catf(res, #field ": %s, ", SCAN_OPTION(options, field) ? "true" : "false");
APPEND_OPTION_INSPECT(with_path)
APPEND_OPTION_INSPECT(verbose_error)
APPEND_OPTION_INSPECT(allow_comments)
APPEND_OPTION_INSPECT(dont_validate_strings)
APPEND_OPTION_INSPECT(allow_trailing_garbage)
APPEND_OPTION_INSPECT(allow_multiple_values)
APPEND_OPTION_INSPECT(allow_partial_values)
APPEND_OPTION_INSPECT(symbolize_path_keys)
APPEND_OPTION_INSPECT(with_roots_info)
OPTIONS_FIELDS(APPEND_OPTION_INSPECT)
#undef APPEND_OPTION_INSPECT
if (RSTRING_END(res)[-1] == ' ')
rb_str_resize(res, RSTRING_LEN(res) - 2);
Expand All @@ -121,7 +162,17 @@ static void init_options(VALUE json_scanner_module)
rb_cJsonScannerOptions = rb_define_class_under(json_scanner_module, "Options", rb_cObject);
rb_define_alloc_func(rb_cJsonScannerOptions, options_alloc);
rb_define_method(rb_cJsonScannerOptions, "initialize", options_m_initialize, -1);
rb_define_method(rb_cJsonScannerOptions, "initialize_copy", options_m_initialize_copy, 1);
#define DEFINE_OPTIONS_ACCESSOR_METHODS(field) \
rb_define_method(rb_cJsonScannerOptions, #field, options_m_##field, 0); \
rb_define_method(rb_cJsonScannerOptions, #field "=", options_m_##field##_set, 1);
OPTIONS_FIELDS(DEFINE_OPTIONS_ACCESSOR_METHODS)
#undef DEFINE_OPTIONS_ACCESSOR_METHODS
rb_define_method(rb_cJsonScannerOptions, "==", options_m_equal, 1);
rb_define_method(rb_cJsonScannerOptions, "inspect", options_m_inspect, 0);
#define INIT_SCAN_KWARG(field) scan_kwargs_table[SCAN_KWARG_##field] = rb_intern(#field);
OPTIONS_FIELDS(INIT_SCAN_KWARG)
#undef INIT_SCAN_KWARG
}

#endif /* JSON_SCANNER_OPTIONS_H */
Loading
Loading