From 904004e79ad0739c7a8324d6b87708c49f08d464 Mon Sep 17 00:00:00 2001 From: Nigel Tao Date: Sun, 15 Nov 2020 20:56:04 +1100 Subject: [PATCH] Let non-strict JSON Pointer syntax allow "~t" --- example/jsonfindptrs/jsonfindptrs.cc | 25 +++++++++++++++++-------- example/jsonptr/jsonptr.cc | 16 +++++++++++----- internal/cgen/auxiliary/json.cc | 22 +++++++++++++--------- internal/cgen/data/data.go | 18 +++++++++--------- release/c/wuffs-unsupported-snapshot.c | 24 ++++++++++++++---------- std/json/decode_quirks.wuffs | 6 +++--- 6 files changed, 67 insertions(+), 44 deletions(-) diff --git a/example/jsonfindptrs/jsonfindptrs.cc b/example/jsonfindptrs/jsonfindptrs.cc index c973fc6e..bc766a93 100644 --- a/example/jsonfindptrs/jsonfindptrs.cc +++ b/example/jsonfindptrs/jsonfindptrs.cc @@ -176,11 +176,12 @@ static const char* g_usage = "\n" "The -strict-json-pointer-syntax flag restricts the output lines to\n" "exactly RFC 6901, with only two escape sequences: \"~0\" and \"~1\" for\n" - "\"~\" and \"/\". Without this flag, this program also lets \"~n\" and\n" - "\"~r\" escape the New Line and Carriage Return ASCII control characters,\n" - "which can work better with line oriented Unix tools that assume exactly\n" - "one value (i.e. one JSON Pointer string) per line. With this flag, it\n" - "fails if the input JSON's keys contain \"\\u000A\" or \"\\u000D\".\n" + "\"~\" and \"/\". Without this flag, this program also lets \"~n\",\n" + "\"~r\" and \"~t\" escape the New Line, Carriage Return and Horizontal\n" + "Tab ASCII control characters, which can work better with line oriented\n" + "(and tab separated) Unix tools that assume exactly one record (e.g. one\n" + "JSON Pointer string) per line. With this flag, it fails if the input\n" + "JSON's keys contain \"\\u0009\", \"\\u000A\" or \"\\u000D\".\n" "\n" "----\n" "\n" @@ -328,7 +329,7 @@ struct JsonValue : JsonVariant { bool // escape_needed(const std::string& s) { for (const char& c : s) { - if ((c == '~') || (c == '/') || (c == '\n') || (c == '\r')) { + if ((c == '~') || (c == '/') || (c == '\n') || (c == '\r') || (c == '\t')) { return true; } } @@ -359,6 +360,12 @@ escape(const std::string& s) { } e += "~r"; break; + case '\t': + if (g_flags.strict_json_pointer_syntax) { + return ""; + } + e += "~t"; + break; default: e += c; break; @@ -395,7 +402,8 @@ print_json_pointers(JsonValue& jvalue, uint32_t depth) { } else { std::string e = escape(kv.first); if (e.empty()) { - return "main: unsupported \"\\u000A\" or \"\\u000D\" in object key"; + return "main: unsupported \"\\u0009\", \"\\u000A\" or \"\\u000D\" in " + "object key"; } g_dst += e; } @@ -508,7 +516,8 @@ std::string // main1(int argc, char** argv) { TRY(parse_flags(argc, argv)); if (!g_flags.strict_json_pointer_syntax) { - g_quirks.push_back(WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_R_TILDE_N); + g_quirks.push_back( + WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_N_TILDE_R_TILDE_T); } FILE* in = stdin; diff --git a/example/jsonptr/jsonptr.cc b/example/jsonptr/jsonptr.cc index bd0c172f..d67e5eac 100644 --- a/example/jsonptr/jsonptr.cc +++ b/example/jsonptr/jsonptr.cc @@ -241,10 +241,11 @@ static const char* g_usage = "\n" "The -strict-json-pointer-syntax flag restricts the -query=STR string to\n" "exactly RFC 6901, with only two escape sequences: \"~0\" and \"~1\" for\n" - "\"~\" and \"/\". Without this flag, this program also lets \"~n\" and\n" - "\"~r\" escape the New Line and Carriage Return ASCII control characters,\n" - "which can work better with line oriented Unix tools that assume exactly\n" - "one value (i.e. one JSON Pointer string) per line.\n" + "\"~\" and \"/\". Without this flag, this program also lets \"~n\",\n" + "\"~r\" and \"~t\" escape the New Line, Carriage Return and Horizontal\n" + "Tab ASCII control characters, which can work better with line oriented\n" + "(and tab separated) Unix tools that assume exactly one record (e.g. one\n" + "JSON Pointer string) per line.\n" "\n" "----\n" "\n" @@ -660,6 +661,10 @@ class Query { if (*ptr != '\r') { break; } + } else if (*j == 't') { + if (*ptr != '\t') { + break; + } } else { break; } @@ -693,7 +698,7 @@ class Query { // Pointer. In particular, it must be valid UTF-8, and either be empty or // start with a '/'. Any '~' within must immediately be followed by either // '0' or '1'. If strict_json_pointer_syntax is false, a '~' may also be - // followed by either 'n' or 'r'. + // followed by either 'n', 'r' or 't'. static bool validate(char* query_c_string, size_t length, bool strict_json_pointer_syntax) { @@ -719,6 +724,7 @@ class Query { break; case 'n': case 'r': + case 't': if (strict_json_pointer_syntax) { return false; } diff --git a/internal/cgen/auxiliary/json.cc b/internal/cgen/auxiliary/json.cc index 22e799b5..640130fc 100644 --- a/internal/cgen/auxiliary/json.cc +++ b/internal/cgen/auxiliary/json.cc @@ -100,7 +100,7 @@ namespace { std::pair // DecodeJson_SplitJsonPointer(std::string& s, size_t i, - bool allow_tilde_r_tilde_n) { + bool allow_tilde_n_tilde_r_tilde_t) { std::string fragment; while (i < s.size()) { char c = s[i]; @@ -124,13 +124,17 @@ DecodeJson_SplitJsonPointer(std::string& s, fragment.push_back('/'); i++; continue; - } else if (allow_tilde_r_tilde_n) { - if (c == 'r') { + } else if (allow_tilde_n_tilde_r_tilde_t) { + if (c == 'n') { + fragment.push_back('\n'); + i++; + continue; + } else if (c == 'r') { fragment.push_back('\r'); i++; continue; - } else if (c == 'n') { - fragment.push_back('\n'); + } else if (c == 't') { + fragment.push_back('\t'); i++; continue; } @@ -353,12 +357,12 @@ DecodeJson(DecodeJsonCallbacks& callbacks, ret_error_message = "wuffs_aux::DecodeJson: out of memory"; goto done; } - bool allow_tilde_r_tilde_n = false; + bool allow_tilde_n_tilde_r_tilde_t = false; for (size_t i = 0; i < quirks.len; i++) { dec->set_quirk_enabled(quirks.ptr[i], true); if (quirks.ptr[i] == - WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_R_TILDE_N) { - allow_tilde_r_tilde_n = true; + WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_N_TILDE_R_TILDE_T) { + allow_tilde_n_tilde_r_tilde_t = true; } } @@ -380,7 +384,7 @@ DecodeJson(DecodeJsonCallbacks& callbacks, goto done; } std::pair split = DecodeJson_SplitJsonPointer( - json_pointer, i + 1, allow_tilde_r_tilde_n); + json_pointer, i + 1, allow_tilde_n_tilde_r_tilde_t); i = std::move(split.second); if (i == 0) { ret_error_message = DecodeJson_BadJsonPointer; diff --git a/internal/cgen/data/data.go b/internal/cgen/data/data.go index 23734172..d4a5cbb8 100644 --- a/internal/cgen/data/data.go +++ b/internal/cgen/data/data.go @@ -647,8 +647,8 @@ const AuxJsonCc = "" + "error_message = tok_status.message(); \\\n goto done; \\\n } \\\n if (WUFFS_JSON__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE != 0) { \\\n ret_error_message = \\\n \"wuffs_aux::DecodeJson: internal error: bad WORKBUF_LEN\"; \\\n goto done; \\\n } \\\n wuffs_base__slice_u8 work_buf = wuffs_base__empty_slice_u8(); \\\n tok_status = dec->decode_tokens(&tok_buf, io_buf, work_buf); \\\n } \\\n wuffs_base__token token = tok_buf.data.ptr[tok_buf.meta.ri++]; \\\n uint64_t token_len = token.length(); \\\n if ((io_buf->meta.ri < cursor_index) || " + " \\\n ((io_buf->meta.ri - cursor_index) < token_len)) { \\\n ret_error_message = \\\n \"wuffs_aux::DecodeJson: internal error: bad token indexes\"; \\\n goto done; \\\n } \\\n uint8_t* token_ptr = io_buf->data.ptr + cursor_index; \\\n (void)(token_ptr); \\\n cursor_index += token_len\n\n" + "" + - "// --------\n\nnamespace {\n\n// DecodeJson_SplitJsonPointer returns (\"bar\", 8) for (\"/foo/bar/b~1z/qux\", 5,\n// etc). It returns a 0 size_t when s has invalid JSON Pointer syntax.\n//\n// The string returned is unescaped. If calling it again, this time with i=8,\n// the \"b~1z\" substring would be returned as \"b/z\".\nstd::pair //\nDecodeJson_SplitJsonPointer(std::string& s,\n size_t i,\n bool allow_tilde_r_tilde_n) {\n std::string fragment;\n while (i < s.size()) {\n char c = s[i];\n if (c == '/') {\n break;\n } else if (c != '~') {\n fragment.push_back(c);\n i++;\n continue;\n }\n i++;\n if (i >= s.size()) {\n return std::make_pair(std::string(), 0);\n }\n c = s[i];\n if (c == '0') {\n fragment.push_back('~');\n i++;\n continue;\n } else if (c == '1') {\n fragment.push_back('/');\n i++;\n continue;\n } else if (allow_tilde_r_tilde_n) {\n if (c == 'r') {\n fragment.push_ba" + - "ck('\\r');\n i++;\n continue;\n } else if (c == 'n') {\n fragment.push_back('\\n');\n i++;\n continue;\n }\n }\n return std::make_pair(std::string(), 0);\n }\n return std::make_pair(std::move(fragment), i);\n}\n\n" + + "// --------\n\nnamespace {\n\n// DecodeJson_SplitJsonPointer returns (\"bar\", 8) for (\"/foo/bar/b~1z/qux\", 5,\n// etc). It returns a 0 size_t when s has invalid JSON Pointer syntax.\n//\n// The string returned is unescaped. If calling it again, this time with i=8,\n// the \"b~1z\" substring would be returned as \"b/z\".\nstd::pair //\nDecodeJson_SplitJsonPointer(std::string& s,\n size_t i,\n bool allow_tilde_n_tilde_r_tilde_t) {\n std::string fragment;\n while (i < s.size()) {\n char c = s[i];\n if (c == '/') {\n break;\n } else if (c != '~') {\n fragment.push_back(c);\n i++;\n continue;\n }\n i++;\n if (i >= s.size()) {\n return std::make_pair(std::string(), 0);\n }\n c = s[i];\n if (c == '0') {\n fragment.push_back('~');\n i++;\n continue;\n } else if (c == '1') {\n fragment.push_back('/');\n i++;\n continue;\n } else if (allow_tilde_n_tilde_r_tilde_t) {\n if (c == 'n') {\n " + + "fragment.push_back('\\n');\n i++;\n continue;\n } else if (c == 'r') {\n fragment.push_back('\\r');\n i++;\n continue;\n } else if (c == 't') {\n fragment.push_back('\\t');\n i++;\n continue;\n }\n }\n return std::make_pair(std::string(), 0);\n }\n return std::make_pair(std::move(fragment), i);\n}\n\n" + "" + "// --------\n\nstd::string //\nDecodeJson_WalkJsonPointerFragment(wuffs_base__token_buffer& tok_buf,\n wuffs_base__status& tok_status,\n wuffs_json__decoder::unique_ptr& dec,\n wuffs_base__io_buffer* io_buf,\n std::string& io_error_message,\n size_t& cursor_index,\n sync_io::Input& input,\n std::string& json_pointer_fragment) {\n std::string ret_error_message;\n while (true) {\n WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN;\n\n int64_t vbc = token.value_base_category();\n uint64_t vbd = token.value_base_detail();\n if (vbc == WUFFS_BASE__TOKEN__VBC__FILLER) {\n continue;\n } else if ((vbc != WUFFS_BASE__TOKEN__VBC__STRUCTURE) ||\n !(vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH)) {\n return DecodeJson_NoMatch;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__ST" + "RUCTURE__TO_LIST) {\n goto do_list;\n }\n goto do_dict;\n }\n\ndo_dict:\n // Alternate between these two things:\n // 1. Decode the next dict key (a string). If it matches the fragment, we're\n // done (success). If we've reached the dict's end (VBD__STRUCTURE__POP)\n // so that there was no next dict key, we're done (failure).\n // 2. Otherwise, skip the next dict value.\n while (true) {\n for (std::string str; true;) {\n WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN;\n\n int64_t vbc = token.value_base_category();\n uint64_t vbd = token.value_base_detail();\n switch (vbc) {\n case WUFFS_BASE__TOKEN__VBC__FILLER:\n continue;\n\n case WUFFS_BASE__TOKEN__VBC__STRUCTURE:\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) {\n goto fail;\n }\n return DecodeJson_NoMatch;\n\n case WUFFS_BASE__TOKEN__VBC__STRING: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP) {\n // No-op.\n } els" + @@ -658,13 +658,13 @@ const AuxJsonCc = "" + "\n\n int64_t vbc = token.value_base_category();\n uint64_t vbd = token.value_base_detail();\n if (vbc == WUFFS_BASE__TOKEN__VBC__FILLER) {\n continue;\n }\n\n // Undo the last part of WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN, so that\n // we're only peeking at the next token.\n tok_buf.meta.ri--;\n cursor_index -= token_len;\n\n if ((vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) &&\n (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__POP)) {\n return DecodeJson_NoMatch;\n }\n return \"\";\n } // check_that_a_value_follows\n\nfail:\n return \"wuffs_aux::DecodeJson: internal error: unexpected token\";\ndone:\n return ret_error_message;\n}\n\n} // namespace\n\n" + "" + "// --------\n\nDecodeJsonResult //\nDecodeJson(DecodeJsonCallbacks& callbacks,\n sync_io::Input& input,\n wuffs_base__slice_u32 quirks,\n std::string json_pointer) {\n // Prepare the wuffs_base__io_buffer and the resultant error_message.\n wuffs_base__io_buffer* io_buf = input.BringsItsOwnIOBuffer();\n wuffs_base__io_buffer fallback_io_buf = wuffs_base__empty_io_buffer();\n std::unique_ptr fallback_io_array(nullptr);\n if (!io_buf) {\n fallback_io_array = std::unique_ptr(new uint8_t[4096]);\n fallback_io_buf = wuffs_base__ptr_u8__writer(fallback_io_array.get(), 4096);\n io_buf = &fallback_io_buf;\n }\n // cursor_index is discussed at\n // https://nigeltao.github.io/blog/2020/jsonptr.html#the-cursor-index\n size_t cursor_index = 0;\n std::string ret_error_message;\n std::string io_error_message;\n\n do {\n // Prepare the low-level JSON decoder.\n wuffs_json__decoder::unique_ptr dec = wuffs_json__decoder::alloc();\n if (!dec) {\n ret_error_message = " + - "\"wuffs_aux::DecodeJson: out of memory\";\n goto done;\n }\n bool allow_tilde_r_tilde_n = false;\n for (size_t i = 0; i < quirks.len; i++) {\n dec->set_quirk_enabled(quirks.ptr[i], true);\n if (quirks.ptr[i] ==\n WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_R_TILDE_N) {\n allow_tilde_r_tilde_n = true;\n }\n }\n\n // Prepare the wuffs_base__tok_buffer. 256 tokens is 2KiB.\n wuffs_base__token tok_array[256];\n wuffs_base__token_buffer tok_buf =\n wuffs_base__slice_token__writer(wuffs_base__make_slice_token(\n &tok_array[0], (sizeof(tok_array) / sizeof(tok_array[0]))));\n wuffs_base__status tok_status = wuffs_base__make_status(nullptr);\n\n // Prepare other state.\n uint32_t depth = 0;\n std::string str;\n\n // Walk the (optional) JSON Pointer.\n for (size_t i = 0; i < json_pointer.size();) {\n if (json_pointer[i] != '/') {\n ret_error_message = DecodeJson_BadJsonPointer;\n goto done;\n }\n std::pair spl" + - "it = DecodeJson_SplitJsonPointer(\n json_pointer, i + 1, allow_tilde_r_tilde_n);\n i = std::move(split.second);\n if (i == 0) {\n ret_error_message = DecodeJson_BadJsonPointer;\n goto done;\n }\n ret_error_message = DecodeJson_WalkJsonPointerFragment(\n tok_buf, tok_status, dec, io_buf, io_error_message, cursor_index,\n input, split.first);\n if (!ret_error_message.empty()) {\n goto done;\n }\n }\n\n // Loop, doing these two things:\n // 1. Get the next token.\n // 2. Process that token.\n while (true) {\n WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN;\n\n int64_t vbc = token.value_base_category();\n uint64_t vbd = token.value_base_detail();\n switch (vbc) {\n case WUFFS_BASE__TOKEN__VBC__FILLER:\n continue;\n\n case WUFFS_BASE__TOKEN__VBC__STRUCTURE: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) {\n ret_error_message = callbacks.Push(static_cast(vbd));\n i" + - "f (!ret_error_message.empty()) {\n goto done;\n }\n depth++;\n continue;\n }\n ret_error_message = callbacks.Pop(static_cast(vbd));\n depth--;\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__STRING: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP) {\n // No-op.\n } else if (vbd &\n WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY) {\n const char* ptr = // Convert from (uint8_t*).\n static_cast(static_cast(token_ptr));\n str.append(ptr, token_len);\n } else {\n goto fail;\n }\n if (token.continued()) {\n continue;\n }\n ret_error_message = callbacks.AppendTextString(std::move(str));\n str.clear();\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT: {\n uin" + - "t8_t u[WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL];\n size_t n = wuffs_base__utf_8__encode(\n wuffs_base__make_slice_u8(\n &u[0], WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL),\n static_cast(vbd));\n const char* ptr = // Convert from (uint8_t*).\n static_cast(static_cast(&u[0]));\n str.append(ptr, n);\n if (token.continued()) {\n continue;\n }\n goto fail;\n }\n\n case WUFFS_BASE__TOKEN__VBC__LITERAL: {\n ret_error_message =\n (vbd & WUFFS_BASE__TOKEN__VBD__LITERAL__NULL)\n ? callbacks.AppendNull()\n : callbacks.AppendBool(vbd &\n WUFFS_BASE__TOKEN__VBD__LITERAL__TRUE);\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__NUMBER: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT) {\n if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__" + - "CONTENT_INTEGER_SIGNED) {\n wuffs_base__result_i64 r = wuffs_base__parse_number_i64(\n wuffs_base__make_slice_u8(token_ptr, token_len),\n WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS);\n if (r.status.is_ok()) {\n ret_error_message = callbacks.AppendI64(r.value);\n goto parsed_a_value;\n }\n }\n if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT) {\n wuffs_base__result_f64 r = wuffs_base__parse_number_f64(\n wuffs_base__make_slice_u8(token_ptr, token_len),\n WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS);\n if (r.status.is_ok()) {\n ret_error_message = callbacks.AppendF64(r.value);\n goto parsed_a_value;\n }\n }\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_" + - "bit_representation__from_u64_to_f64(\n 0xFFF0000000000000ul));\n goto parsed_a_value;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0x7FF0000000000000ul));\n goto parsed_a_value;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0xFFFFFFFFFFFFFFFFul));\n goto parsed_a_value;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0x7FFFFFFFFFFFFFFFul));\n goto parsed_a_value;\n }\n goto fail;\n }\n }\n\n fail:\n " + - " ret_error_message =\n \"wuffs_aux::DecodeJson: internal error: unexpected token\";\n goto done;\n\n parsed_a_value:\n if (!ret_error_message.empty() || (depth == 0)) {\n goto done;\n }\n }\n } while (false);\n\ndone:\n DecodeJsonResult result(\n std::move(ret_error_message),\n wuffs_base__u64__sat_add(io_buf->meta.pos, cursor_index));\n callbacks.Done(result, input, *io_buf);\n return result;\n}\n\n#undef WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN\n\n} // namespace wuffs_aux\n\n#endif // !defined(WUFFS_CONFIG__MODULES) ||\n // defined(WUFFS_CONFIG__MODULE__AUX__JSON)\n" + + "\"wuffs_aux::DecodeJson: out of memory\";\n goto done;\n }\n bool allow_tilde_n_tilde_r_tilde_t = false;\n for (size_t i = 0; i < quirks.len; i++) {\n dec->set_quirk_enabled(quirks.ptr[i], true);\n if (quirks.ptr[i] ==\n WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_N_TILDE_R_TILDE_T) {\n allow_tilde_n_tilde_r_tilde_t = true;\n }\n }\n\n // Prepare the wuffs_base__tok_buffer. 256 tokens is 2KiB.\n wuffs_base__token tok_array[256];\n wuffs_base__token_buffer tok_buf =\n wuffs_base__slice_token__writer(wuffs_base__make_slice_token(\n &tok_array[0], (sizeof(tok_array) / sizeof(tok_array[0]))));\n wuffs_base__status tok_status = wuffs_base__make_status(nullptr);\n\n // Prepare other state.\n uint32_t depth = 0;\n std::string str;\n\n // Walk the (optional) JSON Pointer.\n for (size_t i = 0; i < json_pointer.size();) {\n if (json_pointer[i] != '/') {\n ret_error_message = DecodeJson_BadJsonPointer;\n goto done;\n }\n std::pair<" + + "std::string, size_t> split = DecodeJson_SplitJsonPointer(\n json_pointer, i + 1, allow_tilde_n_tilde_r_tilde_t);\n i = std::move(split.second);\n if (i == 0) {\n ret_error_message = DecodeJson_BadJsonPointer;\n goto done;\n }\n ret_error_message = DecodeJson_WalkJsonPointerFragment(\n tok_buf, tok_status, dec, io_buf, io_error_message, cursor_index,\n input, split.first);\n if (!ret_error_message.empty()) {\n goto done;\n }\n }\n\n // Loop, doing these two things:\n // 1. Get the next token.\n // 2. Process that token.\n while (true) {\n WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN;\n\n int64_t vbc = token.value_base_category();\n uint64_t vbd = token.value_base_detail();\n switch (vbc) {\n case WUFFS_BASE__TOKEN__VBC__FILLER:\n continue;\n\n case WUFFS_BASE__TOKEN__VBC__STRUCTURE: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) {\n ret_error_message = callbacks.Push(static_cas" + + "t(vbd));\n if (!ret_error_message.empty()) {\n goto done;\n }\n depth++;\n continue;\n }\n ret_error_message = callbacks.Pop(static_cast(vbd));\n depth--;\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__STRING: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP) {\n // No-op.\n } else if (vbd &\n WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY) {\n const char* ptr = // Convert from (uint8_t*).\n static_cast(static_cast(token_ptr));\n str.append(ptr, token_len);\n } else {\n goto fail;\n }\n if (token.continued()) {\n continue;\n }\n ret_error_message = callbacks.AppendTextString(std::move(str));\n str.clear();\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__UNI" + + "CODE_CODE_POINT: {\n uint8_t u[WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL];\n size_t n = wuffs_base__utf_8__encode(\n wuffs_base__make_slice_u8(\n &u[0], WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL),\n static_cast(vbd));\n const char* ptr = // Convert from (uint8_t*).\n static_cast(static_cast(&u[0]));\n str.append(ptr, n);\n if (token.continued()) {\n continue;\n }\n goto fail;\n }\n\n case WUFFS_BASE__TOKEN__VBC__LITERAL: {\n ret_error_message =\n (vbd & WUFFS_BASE__TOKEN__VBD__LITERAL__NULL)\n ? callbacks.AppendNull()\n : callbacks.AppendBool(vbd &\n WUFFS_BASE__TOKEN__VBD__LITERAL__TRUE);\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__NUMBER: {\n if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT) {\n if (vbd & " + + "WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED) {\n wuffs_base__result_i64 r = wuffs_base__parse_number_i64(\n wuffs_base__make_slice_u8(token_ptr, token_len),\n WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS);\n if (r.status.is_ok()) {\n ret_error_message = callbacks.AppendI64(r.value);\n goto parsed_a_value;\n }\n }\n if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT) {\n wuffs_base__result_f64 r = wuffs_base__parse_number_f64(\n wuffs_base__make_slice_u8(token_ptr, token_len),\n WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS);\n if (r.status.is_ok()) {\n ret_error_message = callbacks.AppendF64(r.value);\n goto parsed_a_value;\n }\n }\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF) {\n ret_error_message = callbacks.AppendF64(\n " + + " wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0xFFF0000000000000ul));\n goto parsed_a_value;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0x7FF0000000000000ul));\n goto parsed_a_value;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0xFFFFFFFFFFFFFFFFul));\n goto parsed_a_value;\n } else if (vbd & WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN) {\n ret_error_message = callbacks.AppendF64(\n wuffs_base__ieee_754_bit_representation__from_u64_to_f64(\n 0x7FFFFFFFFFFFFFFFul));\n goto parsed_a_value;\n }\n goto fail;" + + "\n }\n }\n\n fail:\n ret_error_message =\n \"wuffs_aux::DecodeJson: internal error: unexpected token\";\n goto done;\n\n parsed_a_value:\n if (!ret_error_message.empty() || (depth == 0)) {\n goto done;\n }\n }\n } while (false);\n\ndone:\n DecodeJsonResult result(\n std::move(ret_error_message),\n wuffs_base__u64__sat_add(io_buf->meta.pos, cursor_index));\n callbacks.Done(result, input, *io_buf);\n return result;\n}\n\n#undef WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN\n\n} // namespace wuffs_aux\n\n#endif // !defined(WUFFS_CONFIG__MODULES) ||\n // defined(WUFFS_CONFIG__MODULE__AUX__JSON)\n" + "" const AuxJsonHh = "" + diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index 1ab7a05a..8c6dfa47 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -7616,7 +7616,7 @@ extern const char wuffs_json__error__unsupported_recursion_depth[]; #define WUFFS_JSON__QUIRK_EXPECT_TRAILING_NEW_LINE_OR_EOF 1225364498 -#define WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_R_TILDE_N 1225364499 +#define WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_N_TILDE_R_TILDE_T 1225364499 #define WUFFS_JSON__QUIRK_REPLACE_INVALID_UNICODE 1225364500 @@ -29425,7 +29425,7 @@ namespace { std::pair // DecodeJson_SplitJsonPointer(std::string& s, size_t i, - bool allow_tilde_r_tilde_n) { + bool allow_tilde_n_tilde_r_tilde_t) { std::string fragment; while (i < s.size()) { char c = s[i]; @@ -29449,13 +29449,17 @@ DecodeJson_SplitJsonPointer(std::string& s, fragment.push_back('/'); i++; continue; - } else if (allow_tilde_r_tilde_n) { - if (c == 'r') { + } else if (allow_tilde_n_tilde_r_tilde_t) { + if (c == 'n') { + fragment.push_back('\n'); + i++; + continue; + } else if (c == 'r') { fragment.push_back('\r'); i++; continue; - } else if (c == 'n') { - fragment.push_back('\n'); + } else if (c == 't') { + fragment.push_back('\t'); i++; continue; } @@ -29678,12 +29682,12 @@ DecodeJson(DecodeJsonCallbacks& callbacks, ret_error_message = "wuffs_aux::DecodeJson: out of memory"; goto done; } - bool allow_tilde_r_tilde_n = false; + bool allow_tilde_n_tilde_r_tilde_t = false; for (size_t i = 0; i < quirks.len; i++) { dec->set_quirk_enabled(quirks.ptr[i], true); if (quirks.ptr[i] == - WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_R_TILDE_N) { - allow_tilde_r_tilde_n = true; + WUFFS_JSON__QUIRK_JSON_POINTER_ALLOW_TILDE_N_TILDE_R_TILDE_T) { + allow_tilde_n_tilde_r_tilde_t = true; } } @@ -29705,7 +29709,7 @@ DecodeJson(DecodeJsonCallbacks& callbacks, goto done; } std::pair split = DecodeJson_SplitJsonPointer( - json_pointer, i + 1, allow_tilde_r_tilde_n); + json_pointer, i + 1, allow_tilde_n_tilde_r_tilde_t); i = std::move(split.second); if (i == 0) { ret_error_message = DecodeJson_BadJsonPointer; diff --git a/std/json/decode_quirks.wuffs b/std/json/decode_quirks.wuffs index 34b0ab0b..f4030333 100644 --- a/std/json/decode_quirks.wuffs +++ b/std/json/decode_quirks.wuffs @@ -206,12 +206,12 @@ pub const QUIRK_ALLOW_TRAILING_FILLER : base.u32 = 0x4909_9400 | 0x11 // "application/json-seq". pub const QUIRK_EXPECT_TRAILING_NEW_LINE_OR_EOF : base.u32 = 0x4909_9400 | 0x12 -// When this quirk is enabled, JSON Pointer strings containing "~r" or "~n", -// which would otherwise be invalid, are unescaped as "\r" or "\n". +// When this quirk is enabled, JSON Pointer strings containing "~n", "~r" or +// "~t", which would otherwise be invalid, are unescaped as "\n", "\r" or "\t". // // This quirk isn't used by Wuffs' std/json package per se, but it is used by // the wuffs_aux::DecodeJson function. -pub const QUIRK_JSON_POINTER_ALLOW_TILDE_R_TILDE_N : base.u32 = 0x4909_9400 | 0x13 +pub const QUIRK_JSON_POINTER_ALLOW_TILDE_N_TILDE_R_TILDE_T : base.u32 = 0x4909_9400 | 0x13 // When this quirk is enabled, invalid UTF-8 inside a JSON string is accepted. // Each byte of invalid UTF-8 is equivalent to "\uFFFD", the Unicode