From 4e193599ac9128204b941e1707e29e2aa3f1b876 Mon Sep 17 00:00:00 2001 From: Nigel Tao Date: Wed, 15 Jul 2020 12:48:57 +1000 Subject: [PATCH] Add std/cbor --- doc/changelog.md | 1 + example/jsonptr/jsonptr.cc | 94 ++++- internal/cgen/base/token-public.h | 8 +- internal/cgen/data/data.go | 6 +- lang/builtin/builtin.go | 3 + release/c/wuffs-unsupported-snapshot.c | 505 ++++++++++++++++++++++++- script/print-json-token-debug-format.c | 53 ++- std/cbor/decode_cbor.wuffs | 108 ++++++ test/c/std/cbor.c | 133 +++++++ 9 files changed, 887 insertions(+), 24 deletions(-) create mode 100644 std/cbor/decode_cbor.wuffs create mode 100644 test/c/std/cbor.c diff --git a/doc/changelog.md b/doc/changelog.md index b320e7f2..8f045550 100644 --- a/doc/changelog.md +++ b/doc/changelog.md @@ -13,6 +13,7 @@ - Added `example/imageviewer`. - Added `example/jsonptr`. - Added `std/bmp`. +- Added `std/cbor`. - Added `std/gif.config_decoder`. - Added `std/json`. - Added `std/wbmp`. diff --git a/example/jsonptr/jsonptr.cc b/example/jsonptr/jsonptr.cc index b4768754..b5f86b66 100644 --- a/example/jsonptr/jsonptr.cc +++ b/example/jsonptr/jsonptr.cc @@ -121,6 +121,7 @@ for a C++ compiler $CXX, such as clang++ or g++. // code simply isn't compiled. #define WUFFS_CONFIG__MODULES #define WUFFS_CONFIG__MODULE__BASE +#define WUFFS_CONFIG__MODULE__CBOR #define WUFFS_CONFIG__MODULE__JSON // If building this program in an environment that doesn't easily accommodate @@ -152,6 +153,7 @@ static const char* g_usage = "Flags:\n" " -c -compact-output\n" " -d=NUM -max-output-depth=NUM\n" + " -i=FMT -input-format={json,cbor}\n" " -o=FMT -output-format={json,cbor}\n" " -q=STR -query=STR\n" " -s=NUM -spaces=NUM\n" @@ -316,7 +318,9 @@ in_dict_before_key() { uint32_t g_suppress_write_dst; bool g_wrote_to_dst; -wuffs_json__decoder g_dec; +wuffs_cbor__decoder g_cbor_decoder; +wuffs_json__decoder g_json_decoder; +wuffs_base__token_decoder* g_dec; // cbor_output_string_array is a 4 KiB buffer. For -output-format=cbor, strings // whose length are 4096 or less are written as a single definite-length @@ -597,6 +601,7 @@ struct { bool compact_output; bool fail_if_unsandboxed; + file_format input_format; bool input_json_extra_comma; uint32_t max_output_depth; file_format output_format; @@ -656,6 +661,14 @@ parse_flags(int argc, char** argv) { g_flags.fail_if_unsandboxed = true; continue; } + if (!strcmp(arg, "i=cbor") || !strcmp(arg, "input-format=cbor")) { + g_flags.input_format = file_format::cbor; + continue; + } + if (!strcmp(arg, "i=json") || !strcmp(arg, "input-format=json")) { + g_flags.input_format = file_format::json; + continue; + } if (!strcmp(arg, "input-json-extra-comma")) { g_flags.input_json_extra_comma = true; continue; @@ -747,18 +760,27 @@ initialize_globals(int argc, char** argv) { g_suppress_write_dst = g_query.next_fragment() ? 1 : 0; g_wrote_to_dst = false; - TRY(g_dec.initialize(sizeof__wuffs_json__decoder(), WUFFS_VERSION, 0) - .message()); + if (g_flags.input_format == file_format::json) { + TRY(g_json_decoder + .initialize(sizeof__wuffs_json__decoder(), WUFFS_VERSION, 0) + .message()); + g_dec = g_json_decoder.upcast_as__wuffs_base__token_decoder(); + } else { + TRY(g_cbor_decoder + .initialize(sizeof__wuffs_cbor__decoder(), WUFFS_VERSION, 0) + .message()); + g_dec = g_cbor_decoder.upcast_as__wuffs_base__token_decoder(); + } if (g_flags.input_json_extra_comma) { - g_dec.set_quirk_enabled(WUFFS_JSON__QUIRK_ALLOW_EXTRA_COMMA, true); + g_dec->set_quirk_enabled(WUFFS_JSON__QUIRK_ALLOW_EXTRA_COMMA, true); } // Consume an optional whitespace trailer. This isn't part of the JSON spec, // but it works better with line oriented Unix tools (such as "echo 123 | // jsonptr" where it's "echo", not "echo -n") or hand-edited JSON files which // can accidentally contain trailing whitespace. - g_dec.set_quirk_enabled(WUFFS_JSON__QUIRK_ALLOW_TRAILING_NEW_LINE, true); + g_dec->set_quirk_enabled(WUFFS_JSON__QUIRK_ALLOW_TRAILING_NEW_LINE, true); return nullptr; } @@ -937,11 +959,48 @@ write_number_cbor_u64(uint8_t base, uint64_t u) { const char* // write_number(uint64_t vbd, uint8_t* ptr, size_t len) { - if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT) { - if (g_flags.output_format == file_format::json) { + if (g_flags.output_format == file_format::json) { + if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT) { return write_dst(ptr, len); + } else if ((vbd & + WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED) && + (vbd & + WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN)) { + if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_IGNORE_FIRST_BYTE) { + if (len == 0) { + goto fail; + } + ptr++; + len--; + } + uint64_t u; + switch (len) { + case 1: + u = wuffs_base__load_u8__no_bounds_check(ptr); + break; + case 2: + u = wuffs_base__load_u16be__no_bounds_check(ptr); + break; + case 4: + u = wuffs_base__load_u32be__no_bounds_check(ptr); + break; + case 8: + u = wuffs_base__load_u64be__no_bounds_check(ptr); + break; + default: + goto fail; + } + uint8_t buf[WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL]; + size_t n = wuffs_base__render_number_u64( + wuffs_base__make_slice_u8(&buf[0], sizeof buf), u, + WUFFS_BASE__RENDER_NUMBER_XXX__DEFAULT_OPTIONS); + return write_dst(&buf[0], n); } + // From here on, (g_flags.output_format == file_format::cbor). + } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN) { + return write_dst(ptr, len); + } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT) { // First try to parse (ptr, len) as an integer. Something like // "1180591620717411303424" is a valid number (in the JSON sense) but will // overflow int64_t or uint64_t, so fall back to parsing it as a float64. @@ -973,9 +1032,23 @@ write_number(uint64_t vbd, uint8_t* ptr, size_t len) { } } +fail: return "main: internal error: unexpected write_number argument"; } +const char* // +write_inline_integer(uint64_t vbd, uint8_t* ptr, size_t len) { + if (g_flags.output_format == file_format::cbor) { + return write_dst(ptr, len); + } + + uint8_t buf[WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL]; + size_t n = wuffs_base__render_number_i64( + wuffs_base__make_slice_u8(&buf[0], sizeof buf), (int16_t)vbd, + WUFFS_BASE__RENDER_NUMBER_XXX__DEFAULT_OPTIONS); + return write_dst(&buf[0], n); +} + // ---- uint8_t // @@ -1332,6 +1405,11 @@ handle_token(wuffs_base__token t, bool start_of_token_chain) { TRY(write_number(vbd, g_src.data.ptr + g_curr_token_end_src_index - len, len)); goto after_value; + + case WUFFS_BASE__TOKEN__VBC__INLINE_INTEGER: + TRY(write_inline_integer( + vbd, g_src.data.ptr + g_curr_token_end_src_index - len, len)); + goto after_value; } // Return an error if we didn't match the (vbc, vbd) pair. @@ -1370,7 +1448,7 @@ main1(int argc, char** argv) { bool start_of_token_chain = true; while (true) { - wuffs_base__status status = g_dec.decode_tokens( + wuffs_base__status status = g_dec->decode_tokens( &g_tok, &g_src, wuffs_base__make_slice_u8(g_work_buffer_array, WORK_BUFFER_ARRAY_SIZE)); diff --git a/internal/cgen/base/token-public.h b/internal/cgen/base/token-public.h index 8dba5297..362cc0ef 100644 --- a/internal/cgen/base/token-public.h +++ b/internal/cgen/base/token-public.h @@ -63,6 +63,7 @@ wuffs_base__make_token(uint64_t repr) { #define WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT 3 #define WUFFS_BASE__TOKEN__VBC__LITERAL 4 #define WUFFS_BASE__TOKEN__VBC__NUMBER 5 +#define WUFFS_BASE__TOKEN__VBC__INLINE_INTEGER 6 // -------- @@ -137,7 +138,7 @@ wuffs_base__make_token(uint64_t repr) { // -------- // For a source string of "123" or "0x9A", it is valid for a tokenizer to -// return any one of: +// return any combination of: // - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT. // - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED. // - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED. @@ -156,11 +157,14 @@ wuffs_base__make_token(uint64_t repr) { // The number 300 might be represented as "\x01\x2C", "\x2C\x01\x00\x00" or // "300", which are big-endian, little-endian or text. For binary formats, the -// token length discriminates e.g. u16 little-endian vs u32 little-endian. +// token length (after adjusting for FORMAT_IGNORE_ETC) discriminates +// e.g. u16 little-endian vs u32 little-endian. #define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN 0x00100 #define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_ENDIAN 0x00200 #define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT 0x00400 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_IGNORE_FIRST_BYTE 0x01000 + // -------- // wuffs_base__token__value returns the token's high 46 bits, sign-extended. A diff --git a/internal/cgen/data/data.go b/internal/cgen/data/data.go index 1530fca8..f4629f08 100644 --- a/internal/cgen/data/data.go +++ b/internal/cgen/data/data.go @@ -569,7 +569,7 @@ const BaseTokenPublicH = "" + "" + "// --------\n\n#define WUFFS_BASE__TOKEN__LENGTH__MAX_INCL 0xFFFF\n\n#define WUFFS_BASE__TOKEN__VALUE__SHIFT 17\n#define WUFFS_BASE__TOKEN__VALUE_EXTENSION__SHIFT 17\n#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 42\n#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 17\n#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 38\n#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 17\n#define WUFFS_BASE__TOKEN__CONTINUED__SHIFT 16\n#define WUFFS_BASE__TOKEN__LENGTH__SHIFT 0\n\n" + "" + - "// --------\n\n#define WUFFS_BASE__TOKEN__VBC__FILLER 0\n#define WUFFS_BASE__TOKEN__VBC__STRUCTURE 1\n#define WUFFS_BASE__TOKEN__VBC__STRING 2\n#define WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT 3\n#define WUFFS_BASE__TOKEN__VBC__LITERAL 4\n#define WUFFS_BASE__TOKEN__VBC__NUMBER 5\n\n" + + "// --------\n\n#define WUFFS_BASE__TOKEN__VBC__FILLER 0\n#define WUFFS_BASE__TOKEN__VBC__STRUCTURE 1\n#define WUFFS_BASE__TOKEN__VBC__STRING 2\n#define WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT 3\n#define WUFFS_BASE__TOKEN__VBC__LITERAL 4\n#define WUFFS_BASE__TOKEN__VBC__NUMBER 5\n#define WUFFS_BASE__TOKEN__VBC__INLINE_INTEGER 6\n\n" + "" + "// --------\n\n#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_LINE 0x00001\n#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_BLOCK 0x00002\n\n" + "" + @@ -581,8 +581,8 @@ const BaseTokenPublicH = "" + "" + "// --------\n\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__UNDEFINED 0x00001\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__NULL 0x00002\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__FALSE 0x00004\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__TRUE 0x00008\n\n" + "" + - "// --------\n\n// For a source string of \"123\" or \"0x9A\", it is valid for a tokenizer to\n// return any one of:\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED.\n//\n// For a source string of \"+123\" or \"-0x9A\", only the first two are valid.\n//\n// For a source string of \"123.\", only the first one is valid.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT 0x00001\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED 0x00002\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED 0x00004\n\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF 0x00010\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF 0x00020\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN 0x00040\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN 0x00080\n\n// The number 300 might be represented as \"\\x01\\x2C\", \"\\x2C\\x01\\x00\\x00\" or\n// \"300\", which are big-endian, li" + - "ttle-endian or text. For binary formats, the\n// token length discriminates e.g. u16 little-endian vs u32 little-endian.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN 0x00100\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_ENDIAN 0x00200\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT 0x00400\n\n" + + "// --------\n\n// For a source string of \"123\" or \"0x9A\", it is valid for a tokenizer to\n// return any combination of:\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED.\n//\n// For a source string of \"+123\" or \"-0x9A\", only the first two are valid.\n//\n// For a source string of \"123.\", only the first one is valid.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT 0x00001\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED 0x00002\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED 0x00004\n\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF 0x00010\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF 0x00020\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN 0x00040\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN 0x00080\n\n// The number 300 might be represented as \"\\x01\\x2C\", \"\\x2C\\x01\\x00\\x00\" or\n// \"300\", which are big-en" + + "dian, little-endian or text. For binary formats, the\n// token length (after adjusting for FORMAT_IGNORE_ETC) discriminates\n// e.g. u16 little-endian vs u32 little-endian.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN 0x00100\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_ENDIAN 0x00200\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT 0x00400\n\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_IGNORE_FIRST_BYTE 0x01000\n\n" + "" + "// --------\n\n// wuffs_base__token__value returns the token's high 46 bits, sign-extended. A\n// negative value means an extended token, non-negative means a simple token.\nstatic inline int64_t //\nwuffs_base__token__value(const wuffs_base__token* t) {\n return ((int64_t)(t->repr)) >> WUFFS_BASE__TOKEN__VALUE__SHIFT;\n}\n\n// wuffs_base__token__value_extension returns a negative value if the token was\n// not an extended token.\nstatic inline int64_t //\nwuffs_base__token__value_extension(const wuffs_base__token* t) {\n return (~(int64_t)(t->repr)) >> WUFFS_BASE__TOKEN__VALUE_EXTENSION__SHIFT;\n}\n\n// wuffs_base__token__value_major returns a negative value if the token was not\n// a simple token.\nstatic inline int64_t //\nwuffs_base__token__value_major(const wuffs_base__token* t) {\n return ((int64_t)(t->repr)) >> WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT;\n}\n\n// wuffs_base__token__value_base_category returns a negative value if the token\n// was not a simple token.\nstatic inline int64_t //\nwuffs_base__token__value_base_cat" + "egory(const wuffs_base__token* t) {\n return ((int64_t)(t->repr)) >> WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__value_minor(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) & 0x1FFFFFF;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__value_base_detail(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT) & 0x1FFFFF;\n}\n\nstatic inline bool //\nwuffs_base__token__continued(const wuffs_base__token* t) {\n return t->repr & 0x10000;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__length(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__LENGTH__SHIFT) & 0xFFFF;\n}\n\n#ifdef __cplusplus\n\ninline int64_t //\nwuffs_base__token::value() const {\n return wuffs_base__token__value(this);\n}\n\ninline int64_t //\nwuffs_base__token::value_extension() const {\n return wuffs_base__token__value_extension(this);\n}\n\ninline int64_t //\nwuffs_base__token::value_major() const {\n retu" + diff --git a/lang/builtin/builtin.go b/lang/builtin/builtin.go index 764096a2..e70125c5 100644 --- a/lang/builtin/builtin.go +++ b/lang/builtin/builtin.go @@ -81,6 +81,7 @@ var Consts = [...]struct { {t.IDU32, "3", "TOKEN__VBC__UNICODE_CODE_POINT"}, {t.IDU32, "4", "TOKEN__VBC__LITERAL"}, {t.IDU32, "5", "TOKEN__VBC__NUMBER"}, + {t.IDU32, "6", "TOKEN__VBC__INLINE_INTEGER"}, // ---- @@ -139,6 +140,8 @@ var Consts = [...]struct { {t.IDU32, "0x00200", "TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_ENDIAN"}, {t.IDU32, "0x00400", "TOKEN__VBD__NUMBER__FORMAT_TEXT"}, + {t.IDU32, "0x01000", "TOKEN__VBD__NUMBER__FORMAT_IGNORE_FIRST_BYTE"}, + // ---- {t.IDU32, "0xFFFD", "UNICODE__REPLACEMENT_CHARACTER"}, diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index 261e08cd..26e6e0b1 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -2133,6 +2133,7 @@ wuffs_base__make_token(uint64_t repr) { #define WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT 3 #define WUFFS_BASE__TOKEN__VBC__LITERAL 4 #define WUFFS_BASE__TOKEN__VBC__NUMBER 5 +#define WUFFS_BASE__TOKEN__VBC__INLINE_INTEGER 6 // -------- @@ -2207,7 +2208,7 @@ wuffs_base__make_token(uint64_t repr) { // -------- // For a source string of "123" or "0x9A", it is valid for a tokenizer to -// return any one of: +// return any combination of: // - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT. // - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED. // - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED. @@ -2226,11 +2227,14 @@ wuffs_base__make_token(uint64_t repr) { // The number 300 might be represented as "\x01\x2C", "\x2C\x01\x00\x00" or // "300", which are big-endian, little-endian or text. For binary formats, the -// token length discriminates e.g. u16 little-endian vs u32 little-endian. +// token length (after adjusting for FORMAT_IGNORE_ETC) discriminates +// e.g. u16 little-endian vs u32 little-endian. #define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN 0x00100 #define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_ENDIAN 0x00200 #define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT 0x00400 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_IGNORE_FIRST_BYTE 0x01000 + // -------- // wuffs_base__token__value returns the token's high 46 bits, sign-extended. A @@ -5512,6 +5516,207 @@ extern "C" { // ---------------- Status Codes +extern const char wuffs_cbor__error__bad_input[]; + +// ---------------- Public Consts + +// ---------------- Struct Declarations + +typedef struct wuffs_cbor__decoder__struct wuffs_cbor__decoder; + +// ---------------- Public Initializer Prototypes + +// For any given "wuffs_foo__bar* self", "wuffs_foo__bar__initialize(self, +// etc)" should be called before any other "wuffs_foo__bar__xxx(self, etc)". +// +// Pass sizeof(*self) and WUFFS_VERSION for sizeof_star_self and wuffs_version. +// Pass 0 (or some combination of WUFFS_INITIALIZE__XXX) for options. + +wuffs_base__status WUFFS_BASE__WARN_UNUSED_RESULT +wuffs_cbor__decoder__initialize( + wuffs_cbor__decoder* self, + size_t sizeof_star_self, + uint64_t wuffs_version, + uint32_t options); + +size_t +sizeof__wuffs_cbor__decoder(); + +// ---------------- Allocs + +// These functions allocate and initialize Wuffs structs. They return NULL if +// memory allocation fails. If they return non-NULL, there is no need to call +// wuffs_foo__bar__initialize, but the caller is responsible for eventually +// calling free on the returned pointer. That pointer is effectively a C++ +// std::unique_ptr. + +wuffs_cbor__decoder* +wuffs_cbor__decoder__alloc(); + +static inline wuffs_base__token_decoder* +wuffs_cbor__decoder__alloc_as__wuffs_base__token_decoder() { + return (wuffs_base__token_decoder*)(wuffs_cbor__decoder__alloc()); +} + +// ---------------- Upcasts + +static inline wuffs_base__token_decoder* +wuffs_cbor__decoder__upcast_as__wuffs_base__token_decoder( + wuffs_cbor__decoder* p) { + return (wuffs_base__token_decoder*)p; +} + +// ---------------- Public Function Prototypes + +WUFFS_BASE__MAYBE_STATIC wuffs_base__empty_struct +wuffs_cbor__decoder__set_quirk_enabled( + wuffs_cbor__decoder* self, + uint32_t a_quirk, + bool a_enabled); + +WUFFS_BASE__MAYBE_STATIC wuffs_base__range_ii_u64 +wuffs_cbor__decoder__workbuf_len( + const wuffs_cbor__decoder* self); + +WUFFS_BASE__MAYBE_STATIC wuffs_base__status +wuffs_cbor__decoder__decode_tokens( + wuffs_cbor__decoder* self, + wuffs_base__token_buffer* a_dst, + wuffs_base__io_buffer* a_src, + wuffs_base__slice_u8 a_workbuf); + +// ---------------- Struct Definitions + +// These structs' fields, and the sizeof them, are private implementation +// details that aren't guaranteed to be stable across Wuffs versions. +// +// See https://en.wikipedia.org/wiki/Opaque_pointer#C + +#if defined(__cplusplus) || defined(WUFFS_IMPLEMENTATION) + +struct wuffs_cbor__decoder__struct { + // Do not access the private_impl's or private_data's fields directly. There + // is no API/ABI compatibility or safety guarantee if you do so. Instead, use + // the wuffs_foo__bar__baz functions. + // + // It is a struct, not a struct*, so that the outermost wuffs_foo__bar struct + // can be stack allocated when WUFFS_IMPLEMENTATION is defined. + + struct { + uint32_t magic; + uint32_t active_coroutine; + wuffs_base__vtable vtable_for__wuffs_base__token_decoder; + wuffs_base__vtable null_vtable; + + bool f_end_of_data; + + uint32_t p_decode_tokens[1]; + } private_impl; + + struct { + struct { + uint32_t v_depth; + uint32_t v_token_length; + uint8_t v_c; + uint64_t scratch; + } s_decode_tokens[1]; + } private_data; + +#ifdef __cplusplus +#if (__cplusplus >= 201103L) + using unique_ptr = std::unique_ptr; + + // On failure, the alloc_etc functions return nullptr. They don't throw. + + static inline unique_ptr + alloc() { + return unique_ptr(wuffs_cbor__decoder__alloc(), &free); + } + + static inline wuffs_base__token_decoder::unique_ptr + alloc_as__wuffs_base__token_decoder() { + return wuffs_base__token_decoder::unique_ptr( + wuffs_cbor__decoder__alloc_as__wuffs_base__token_decoder(), &free); + } +#endif // (__cplusplus >= 201103L) + +#if (__cplusplus >= 201103L) && !defined(WUFFS_IMPLEMENTATION) + // Disallow constructing or copying an object via standard C++ mechanisms, + // e.g. the "new" operator, as this struct is intentionally opaque. Its total + // size and field layout is not part of the public, stable, memory-safe API. + // Use malloc or memcpy and the sizeof__wuffs_foo__bar function instead, and + // call wuffs_foo__bar__baz methods (which all take a "this"-like pointer as + // their first argument) rather than tweaking bar.private_impl.qux fields. + // + // In C, we can just leave wuffs_foo__bar as an incomplete type (unless + // WUFFS_IMPLEMENTATION is #define'd). In C++, we define a complete type in + // order to provide convenience methods. These forward on "this", so that you + // can write "bar->baz(etc)" instead of "wuffs_foo__bar__baz(bar, etc)". + wuffs_cbor__decoder__struct() = delete; + wuffs_cbor__decoder__struct(const wuffs_cbor__decoder__struct&) = delete; + wuffs_cbor__decoder__struct& operator=( + const wuffs_cbor__decoder__struct&) = delete; + + // As above, the size of the struct is not part of the public API, and unless + // WUFFS_IMPLEMENTATION is #define'd, this struct type T should be heap + // allocated, not stack allocated. Its size is not intended to be known at + // compile time, but it is unfortunately divulged as a side effect of + // defining C++ convenience methods. Use "sizeof__T()", calling the function, + // instead of "sizeof T", invoking the operator. To make the two values + // different, so that passing the latter will be rejected by the initialize + // function, we add an arbitrary amount of dead weight. + uint8_t dead_weight[123000000]; // 123 MB. +#endif // (__cplusplus >= 201103L) && !defined(WUFFS_IMPLEMENTATION) + + inline wuffs_base__status WUFFS_BASE__WARN_UNUSED_RESULT + initialize( + size_t sizeof_star_self, + uint64_t wuffs_version, + uint32_t options) { + return wuffs_cbor__decoder__initialize( + this, sizeof_star_self, wuffs_version, options); + } + + inline wuffs_base__token_decoder* + upcast_as__wuffs_base__token_decoder() { + return (wuffs_base__token_decoder*)this; + } + + inline wuffs_base__empty_struct + set_quirk_enabled( + uint32_t a_quirk, + bool a_enabled) { + return wuffs_cbor__decoder__set_quirk_enabled(this, a_quirk, a_enabled); + } + + inline wuffs_base__range_ii_u64 + workbuf_len() const { + return wuffs_cbor__decoder__workbuf_len(this); + } + + inline wuffs_base__status + decode_tokens( + wuffs_base__token_buffer* a_dst, + wuffs_base__io_buffer* a_src, + wuffs_base__slice_u8 a_workbuf) { + return wuffs_cbor__decoder__decode_tokens(this, a_dst, a_src, a_workbuf); + } + +#endif // __cplusplus +}; // struct wuffs_cbor__decoder__struct + +#endif // defined(__cplusplus) || defined(WUFFS_IMPLEMENTATION) + +#ifdef __cplusplus +} // extern "C" +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +// ---------------- Status Codes + // ---------------- Public Consts // ---------------- Struct Declarations @@ -16008,6 +16213,302 @@ wuffs_bmp__decoder__workbuf_len( #endif // !defined(WUFFS_CONFIG__MODULES) || defined(WUFFS_CONFIG__MODULE__BMP) +#if !defined(WUFFS_CONFIG__MODULES) || defined(WUFFS_CONFIG__MODULE__CBOR) + +// ---------------- Status Codes Implementations + +const char wuffs_cbor__error__bad_input[] = "#cbor: bad input"; +const char wuffs_cbor__error__internal_error_inconsistent_i_o[] = "#cbor: internal error: inconsistent I/O"; + +// ---------------- Private Consts + +// ---------------- Private Initializer Prototypes + +// ---------------- Private Function Prototypes + +// ---------------- VTables + +const wuffs_base__token_decoder__func_ptrs +wuffs_cbor__decoder__func_ptrs_for__wuffs_base__token_decoder = { + (wuffs_base__status(*)(void*, + wuffs_base__token_buffer*, + wuffs_base__io_buffer*, + wuffs_base__slice_u8))(&wuffs_cbor__decoder__decode_tokens), + (wuffs_base__empty_struct(*)(void*, + uint32_t, + bool))(&wuffs_cbor__decoder__set_quirk_enabled), + (wuffs_base__range_ii_u64(*)(const void*))(&wuffs_cbor__decoder__workbuf_len), +}; + +// ---------------- Initializer Implementations + +wuffs_base__status WUFFS_BASE__WARN_UNUSED_RESULT +wuffs_cbor__decoder__initialize( + wuffs_cbor__decoder* self, + size_t sizeof_star_self, + uint64_t wuffs_version, + uint32_t options){ + if (!self) { + return wuffs_base__make_status(wuffs_base__error__bad_receiver); + } + if (sizeof(*self) != sizeof_star_self) { + return wuffs_base__make_status(wuffs_base__error__bad_sizeof_receiver); + } + if (((wuffs_version >> 32) != WUFFS_VERSION_MAJOR) || + (((wuffs_version >> 16) & 0xFFFF) > WUFFS_VERSION_MINOR)) { + return wuffs_base__make_status(wuffs_base__error__bad_wuffs_version); + } + + if ((options & WUFFS_INITIALIZE__ALREADY_ZEROED) != 0) { + // The whole point of this if-check is to detect an uninitialized *self. + // We disable the warning on GCC. Clang-5.0 does not have this warning. +#if !defined(__clang__) && defined(__GNUC__) +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wmaybe-uninitialized" +#endif + if (self->private_impl.magic != 0) { + return wuffs_base__make_status(wuffs_base__error__initialize_falsely_claimed_already_zeroed); + } +#if !defined(__clang__) && defined(__GNUC__) +#pragma GCC diagnostic pop +#endif + } else { + if ((options & WUFFS_INITIALIZE__LEAVE_INTERNAL_BUFFERS_UNINITIALIZED) == 0) { + memset(self, 0, sizeof(*self)); + options |= WUFFS_INITIALIZE__ALREADY_ZEROED; + } else { + memset(&(self->private_impl), 0, sizeof(self->private_impl)); + } + } + + self->private_impl.magic = WUFFS_BASE__MAGIC; + self->private_impl.vtable_for__wuffs_base__token_decoder.vtable_name = + wuffs_base__token_decoder__vtable_name; + self->private_impl.vtable_for__wuffs_base__token_decoder.function_pointers = + (const void*)(&wuffs_cbor__decoder__func_ptrs_for__wuffs_base__token_decoder); + return wuffs_base__make_status(NULL); +} + +wuffs_cbor__decoder* +wuffs_cbor__decoder__alloc() { + wuffs_cbor__decoder* x = + (wuffs_cbor__decoder*)(calloc(sizeof(wuffs_cbor__decoder), 1)); + if (!x) { + return NULL; + } + if (wuffs_cbor__decoder__initialize( + x, sizeof(wuffs_cbor__decoder), WUFFS_VERSION, WUFFS_INITIALIZE__ALREADY_ZEROED).repr) { + free(x); + return NULL; + } + return x; +} + +size_t +sizeof__wuffs_cbor__decoder() { + return sizeof(wuffs_cbor__decoder); +} + +// ---------------- Function Implementations + +// -------- func cbor.decoder.set_quirk_enabled + +WUFFS_BASE__MAYBE_STATIC wuffs_base__empty_struct +wuffs_cbor__decoder__set_quirk_enabled( + wuffs_cbor__decoder* self, + uint32_t a_quirk, + bool a_enabled) { + return wuffs_base__make_empty_struct(); +} + +// -------- func cbor.decoder.workbuf_len + +WUFFS_BASE__MAYBE_STATIC wuffs_base__range_ii_u64 +wuffs_cbor__decoder__workbuf_len( + const wuffs_cbor__decoder* self) { + if (!self) { + return wuffs_base__utility__empty_range_ii_u64(); + } + if ((self->private_impl.magic != WUFFS_BASE__MAGIC) && + (self->private_impl.magic != WUFFS_BASE__DISABLED)) { + return wuffs_base__utility__empty_range_ii_u64(); + } + + return wuffs_base__utility__empty_range_ii_u64(); +} + +// -------- func cbor.decoder.decode_tokens + +WUFFS_BASE__MAYBE_STATIC wuffs_base__status +wuffs_cbor__decoder__decode_tokens( + wuffs_cbor__decoder* self, + wuffs_base__token_buffer* a_dst, + wuffs_base__io_buffer* a_src, + wuffs_base__slice_u8 a_workbuf) { + if (!self) { + return wuffs_base__make_status(wuffs_base__error__bad_receiver); + } + if (self->private_impl.magic != WUFFS_BASE__MAGIC) { + return wuffs_base__make_status( + (self->private_impl.magic == WUFFS_BASE__DISABLED) + ? wuffs_base__error__disabled_by_previous_error + : wuffs_base__error__initialize_not_called); + } + if (!a_dst || !a_src) { + self->private_impl.magic = WUFFS_BASE__DISABLED; + return wuffs_base__make_status(wuffs_base__error__bad_argument); + } + if ((self->private_impl.active_coroutine != 0) && + (self->private_impl.active_coroutine != 1)) { + self->private_impl.magic = WUFFS_BASE__DISABLED; + return wuffs_base__make_status(wuffs_base__error__interleaved_coroutine_calls); + } + self->private_impl.active_coroutine = 0; + wuffs_base__status status = wuffs_base__make_status(NULL); + + uint32_t v_depth = 0; + uint32_t v_token_length = 0; + uint8_t v_c = 0; + + wuffs_base__token* iop_a_dst = NULL; + wuffs_base__token* io0_a_dst WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + wuffs_base__token* io1_a_dst WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + wuffs_base__token* io2_a_dst WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + if (a_dst) { + io0_a_dst = a_dst->data.ptr; + io1_a_dst = io0_a_dst + a_dst->meta.wi; + iop_a_dst = io1_a_dst; + io2_a_dst = io0_a_dst + a_dst->data.len; + if (a_dst->meta.closed) { + io2_a_dst = iop_a_dst; + } + } + const uint8_t* iop_a_src = NULL; + const uint8_t* io0_a_src WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + const uint8_t* io1_a_src WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + const uint8_t* io2_a_src WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + if (a_src) { + io0_a_src = a_src->data.ptr; + io1_a_src = io0_a_src + a_src->meta.ri; + iop_a_src = io1_a_src; + io2_a_src = io0_a_src + a_src->meta.wi; + } + + uint32_t coro_susp_point = self->private_impl.p_decode_tokens[0]; + if (coro_susp_point) { + v_depth = self->private_data.s_decode_tokens[0].v_depth; + v_token_length = self->private_data.s_decode_tokens[0].v_token_length; + v_c = self->private_data.s_decode_tokens[0].v_c; + } + switch (coro_susp_point) { + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_0; + + if (self->private_impl.f_end_of_data) { + status = wuffs_base__make_status(wuffs_base__note__end_of_data); + goto ok; + } + label__outer__continue:; + while (true) { + while (true) { + if (((uint64_t)(io2_a_dst - iop_a_dst)) <= 0) { + status = wuffs_base__make_status(wuffs_base__suspension__short_write); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(1); + goto label__outer__continue; + } + if (((uint64_t)(io2_a_src - iop_a_src)) <= 0) { + if (a_src && a_src->meta.closed) { + status = wuffs_base__make_status(wuffs_cbor__error__bad_input); + goto exit; + } + status = wuffs_base__make_status(wuffs_base__suspension__short_read); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(2); + goto label__outer__continue; + } + v_c = wuffs_base__load_u8be__no_bounds_check(iop_a_src); + if ((24 <= (v_c & 31)) && ((v_c & 31) <= 27)) { + v_token_length = (1 + (((uint32_t)(1)) << (v_c & 3))); + if (((uint64_t)(io2_a_src - iop_a_src)) < ((uint64_t)(v_token_length))) { + if (a_src && a_src->meta.closed) { + status = wuffs_base__make_status(wuffs_cbor__error__bad_input); + goto exit; + } + status = wuffs_base__make_status(wuffs_base__suspension__short_read); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(3); + goto label__outer__continue; + } + self->private_data.s_decode_tokens[0].scratch = v_token_length; + WUFFS_BASE__COROUTINE_SUSPENSION_POINT(4); + if (self->private_data.s_decode_tokens[0].scratch > ((uint64_t)(io2_a_src - iop_a_src))) { + self->private_data.s_decode_tokens[0].scratch -= ((uint64_t)(io2_a_src - iop_a_src)); + iop_a_src = io2_a_src; + status = wuffs_base__make_status(wuffs_base__suspension__short_read); + goto suspend; + } + iop_a_src += self->private_data.s_decode_tokens[0].scratch; + } else { + (iop_a_src += 1, wuffs_base__make_empty_struct()); + } + if (v_c < 32) { + v_c &= 31; + if (v_c < 24) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)((12582912 | ((uint32_t)(v_c))))) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + goto label__goto_parsed_a_leaf_value__break; + } else if (v_c < 28) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)(10490116)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)((1 + (((uint32_t)(1)) << (v_c - 24))))) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + goto label__goto_parsed_a_leaf_value__break; + } + } + if (iop_a_src > io1_a_src) { + (iop_a_src--, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(wuffs_cbor__error__bad_input); + goto exit; + } + status = wuffs_base__make_status(wuffs_cbor__error__internal_error_inconsistent_i_o); + goto exit; + } + label__goto_parsed_a_leaf_value__break:; + if (v_depth == 0) { + goto label__outer__break; + } + } + label__outer__break:; + self->private_impl.f_end_of_data = true; + + goto ok; + ok: + self->private_impl.p_decode_tokens[0] = 0; + goto exit; + } + + goto suspend; + suspend: + self->private_impl.p_decode_tokens[0] = wuffs_base__status__is_suspension(&status) ? coro_susp_point : 0; + self->private_impl.active_coroutine = wuffs_base__status__is_suspension(&status) ? 1 : 0; + self->private_data.s_decode_tokens[0].v_depth = v_depth; + self->private_data.s_decode_tokens[0].v_token_length = v_token_length; + self->private_data.s_decode_tokens[0].v_c = v_c; + + goto exit; + exit: + if (a_dst) { + a_dst->meta.wi = ((size_t)(iop_a_dst - a_dst->data.ptr)); + } + if (a_src) { + a_src->meta.ri = ((size_t)(iop_a_src - a_src->data.ptr)); + } + + if (wuffs_base__status__is_error(&status)) { + self->private_impl.magic = WUFFS_BASE__DISABLED; + } + return status; +} + +#endif // !defined(WUFFS_CONFIG__MODULES) || defined(WUFFS_CONFIG__MODULE__CBOR) + #if !defined(WUFFS_CONFIG__MODULES) || defined(WUFFS_CONFIG__MODULE__CRC32) // ---------------- Status Codes Implementations diff --git a/script/print-json-token-debug-format.c b/script/print-json-token-debug-format.c index a2a5277c..7c306e87 100644 --- a/script/print-json-token-debug-format.c +++ b/script/print-json-token-debug-format.c @@ -82,6 +82,7 @@ // code simply isn't compiled. #define WUFFS_CONFIG__MODULES #define WUFFS_CONFIG__MODULE__BASE +#define WUFFS_CONFIG__MODULE__CBOR #define WUFFS_CONFIG__MODULE__JSON // If building this program in an environment that doesn't easily accommodate @@ -91,8 +92,14 @@ // Wuffs allows either statically or dynamically allocated work buffers. This // program exercises static allocation. +#if WUFFS_CBOR__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE > \ + WUFFS_JSON__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE +#define WORK_BUFFER_ARRAY_SIZE \ + WUFFS_CBOR__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE +#else #define WORK_BUFFER_ARRAY_SIZE \ WUFFS_JSON__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE +#endif #if WORK_BUFFER_ARRAY_SIZE > 0 uint8_t g_work_buffer_array[WORK_BUFFER_ARRAY_SIZE]; #else @@ -113,7 +120,9 @@ wuffs_base__token g_tok_buffer_array[TOKEN_BUFFER_ARRAY_SIZE]; wuffs_base__io_buffer g_src; wuffs_base__token_buffer g_tok; -wuffs_json__decoder g_dec; +wuffs_cbor__decoder g_cbor_decoder; +wuffs_json__decoder g_json_decoder; +wuffs_base__token_decoder* g_dec; wuffs_base__status g_dec_status; #define TRY(error_msg) \ @@ -149,12 +158,18 @@ read_src() { // ---- +typedef enum file_format_enum { + FILE_FORMAT_JSON, + FILE_FORMAT_CBOR, +} file_format; + struct { int remaining_argc; char** remaining_argv; bool all_tokens; bool human_readable; + file_format input_format; bool quirks; } g_flags = {0}; @@ -188,6 +203,14 @@ parse_flags(int argc, char** argv) { g_flags.human_readable = true; continue; } + if (!strcmp(arg, "i=cbor") || !strcmp(arg, "input-format=cbor")) { + g_flags.input_format = FILE_FORMAT_CBOR; + continue; + } + if (!strcmp(arg, "i=json") || !strcmp(arg, "input-format=json")) { + g_flags.input_format = FILE_FORMAT_JSON; + continue; + } if (!strcmp(arg, "q") || !strcmp(arg, "quirks")) { g_flags.quirks = true; continue; @@ -208,7 +231,7 @@ const char* g_vbc_names[16] = { "3:UnicodeCodePoint", // "4:Literal.........", // "5:Number..........", // - "6:Reserved........", // + "6:Inline_Integer..", // "7:Reserved........", // "8:Reserved........", // "9:Reserved........", // @@ -241,10 +264,22 @@ main1(int argc, char** argv) { wuffs_base__make_slice_token(g_tok_buffer_array, TOKEN_BUFFER_ARRAY_SIZE), wuffs_base__empty_token_buffer_meta()); - wuffs_base__status init_status = wuffs_json__decoder__initialize( - &g_dec, sizeof__wuffs_json__decoder(), WUFFS_VERSION, 0); - if (!wuffs_base__status__is_ok(&init_status)) { - return wuffs_base__status__message(&init_status); + if (g_flags.input_format == FILE_FORMAT_JSON) { + wuffs_base__status init_status = wuffs_json__decoder__initialize( + &g_json_decoder, sizeof__wuffs_json__decoder(), WUFFS_VERSION, 0); + if (!wuffs_base__status__is_ok(&init_status)) { + return wuffs_base__status__message(&init_status); + } + g_dec = wuffs_json__decoder__upcast_as__wuffs_base__token_decoder( + &g_json_decoder); + } else { + wuffs_base__status init_status = wuffs_cbor__decoder__initialize( + &g_cbor_decoder, sizeof__wuffs_cbor__decoder(), WUFFS_VERSION, 0); + if (!wuffs_base__status__is_ok(&init_status)) { + return wuffs_base__status__message(&init_status); + } + g_dec = wuffs_cbor__decoder__upcast_as__wuffs_base__token_decoder( + &g_cbor_decoder); } if (g_flags.quirks) { @@ -269,14 +304,14 @@ main1(int argc, char** argv) { }; uint32_t i; for (i = 0; quirks[i]; i++) { - wuffs_json__decoder__set_quirk_enabled(&g_dec, quirks[i], true); + wuffs_base__token_decoder__set_quirk_enabled(g_dec, quirks[i], true); } } uint64_t pos = 0; while (true) { - wuffs_base__status status = wuffs_json__decoder__decode_tokens( - &g_dec, &g_tok, &g_src, + wuffs_base__status status = wuffs_base__token_decoder__decode_tokens( + g_dec, &g_tok, &g_src, wuffs_base__make_slice_u8(g_work_buffer_array, WORK_BUFFER_ARRAY_SIZE)); while (g_tok.meta.ri < g_tok.meta.wi) { diff --git a/std/cbor/decode_cbor.wuffs b/std/cbor/decode_cbor.wuffs new file mode 100644 index 00000000..7ec6084e --- /dev/null +++ b/std/cbor/decode_cbor.wuffs @@ -0,0 +1,108 @@ +// Copyright 2020 The Wuffs Authors. +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// https://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +pub status "#bad input" + +pri status "#internal error: inconsistent I/O" + +pub struct decoder? implements base.token_decoder( + end_of_data : base.bool, + + util : base.utility, +)( +) + +pub func decoder.set_quirk_enabled!(quirk: base.u32, enabled: base.bool) { +} + +pub func decoder.workbuf_len() base.range_ii_u64 { + return this.util.empty_range_ii_u64() +} + +pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader, workbuf: slice base.u8) { + var depth : base.u32[..= 1024] + var token_length : base.u32[..= 9] + var c : base.u8 + + if this.end_of_data { + return base."@end of data" + } + + while.outer true { + while.goto_parsed_a_leaf_value true {{ + if args.dst.available() <= 0 { + yield? base."$short write" + continue.outer + } + if args.src.available() <= 0 { + if args.src.is_closed() { + return "#bad input" + } + yield? base."$short read" + continue.outer + } + c = args.src.peek_u8() + if (0x18 <= (c & 0x1F)) and ((c & 0x1F) <= 0x1B) { + token_length = 1 + ((1 as base.u32) << (c & 0x03)) + if args.src.available() < (token_length as base.u64) { + if args.src.is_closed() { + return "#bad input" + } + yield? base."$short read" + continue.outer + } + args.src.skip_u32?(n: token_length) + } else { + args.src.skip_u32_fast!(actual: 1, worst_case: 1) + } + + if c < 0x20 { + c &= 0x1F + if c < 0x18 { + args.dst.write_simple_token_fast!( + value_major: 0, + value_minor: (base.TOKEN__VBC__INLINE_INTEGER << 21) | + (c as base.u32), + continued: 0, + length: 1) + break.goto_parsed_a_leaf_value + } else if c < 0x1C { + args.dst.write_simple_token_fast!( + value_major: 0, + value_minor: (base.TOKEN__VBC__NUMBER << 21) | + base.TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED | + base.TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN | + base.TOKEN__VBD__NUMBER__FORMAT_IGNORE_FIRST_BYTE, + continued: 0, + length: 1 + ((1 as base.u32) << (c - 0x18))) + break.goto_parsed_a_leaf_value + } + } + + if args.src.can_undo_byte() { + args.src.undo_byte!() + return "#bad input" + } + return "#internal error: inconsistent I/O" + }} endwhile.goto_parsed_a_leaf_value + + // We've just parsed a leaf (non-container) value: literal (null, + // false, true), number or string. + if depth == 0 { + break.outer + } + } endwhile.outer + + this.end_of_data = true +} diff --git a/test/c/std/cbor.c b/test/c/std/cbor.c new file mode 100644 index 00000000..e903a637 --- /dev/null +++ b/test/c/std/cbor.c @@ -0,0 +1,133 @@ +// Copyright 2020 The Wuffs Authors. +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// https://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +// ---------------- + +/* +This test program is typically run indirectly, by the "wuffs test" or "wuffs +bench" commands. These commands take an optional "-mimic" flag to check that +Wuffs' output mimics (i.e. exactly matches) other libraries' output, such as +giflib for GIF, libpng for PNG, etc. + +To manually run this test: + +for CC in clang gcc; do + $CC -std=c99 -Wall -Werror cbor.c && ./a.out + rm -f a.out +done + +Each edition should print "PASS", amongst other information, and exit(0). + +Add the "wuffs mimic cflags" (everything after the colon below) to the C +compiler flags (after the .c file) to run the mimic tests. + +To manually run the benchmarks, replace "-Wall -Werror" with "-O3" and replace +the first "./a.out" with "./a.out -bench". Combine these changes with the +"wuffs mimic cflags" to run the mimic benchmarks. +*/ + +// !! wuffs mimic cflags: -DWUFFS_MIMIC + +// Wuffs ships as a "single file C library" or "header file library" as per +// https://github.com/nothings/stb/blob/master/docs/stb_howto.txt +// +// To use that single file as a "foo.c"-like implementation, instead of a +// "foo.h"-like header, #define WUFFS_IMPLEMENTATION before #include'ing or +// compiling it. +#define WUFFS_IMPLEMENTATION + +// Defining the WUFFS_CONFIG__MODULE* macros are optional, but it lets users of +// release/c/etc.c whitelist which parts of Wuffs to build. That file contains +// the entire Wuffs standard library, implementing a variety of codecs and file +// formats. Without this macro definition, an optimizing compiler or linker may +// very well discard Wuffs code for unused codecs, but listing the Wuffs +// modules we use makes that process explicit. Preprocessing means that such +// code simply isn't compiled. +#define WUFFS_CONFIG__MODULES +#define WUFFS_CONFIG__MODULE__BASE +#define WUFFS_CONFIG__MODULE__CBOR + +// If building this program in an environment that doesn't easily accommodate +// relative includes, you can use the script/inline-c-relative-includes.go +// program to generate a stand-alone C file. +#include "../../../release/c/wuffs-unsupported-snapshot.c" +#include "../testlib/testlib.c" +#ifdef WUFFS_MIMIC +// No mimic library. +#endif + +// ---------------- CBOR Tests + +const char* // +test_wuffs_cbor_decode_interface() { + CHECK_FOCUS(__func__); + + // TODO. + + return NULL; +} + +// ---------------- Mimic Tests + +#ifdef WUFFS_MIMIC + +// No mimic tests. + +#endif // WUFFS_MIMIC + +// ---------------- CBOR Benches + +// No CBOR benches. + +// ---------------- Mimic Benches + +#ifdef WUFFS_MIMIC + +// No mimic benches. + +#endif // WUFFS_MIMIC + +// ---------------- Manifest + +proc g_tests[] = { + + test_wuffs_cbor_decode_interface, + +#ifdef WUFFS_MIMIC + +// No mimic tests. + +#endif // WUFFS_MIMIC + + NULL, +}; + +proc g_benches[] = { + +// No CBOR benches. + +#ifdef WUFFS_MIMIC + +// No mimic benches. + +#endif // WUFFS_MIMIC + + NULL, +}; + +int // +main(int argc, char** argv) { + g_proc_package_name = "std/cbor"; + return test_main(argc, argv, g_tests, g_benches); +}