From 8aac6767bbe43a71bf96b5274ff51a954b91e4c8 Mon Sep 17 00:00:00 2001 From: Nigel Tao Date: Wed, 12 Aug 2020 22:47:45 +1000 Subject: [PATCH] Add example/jsonfindptrs -query=STR flag --- example/jsonfindptrs/jsonfindptrs.cc | 11 +- internal/cgen/auxiliary/json.cc | 322 +++++++++++++++++++++--- internal/cgen/auxiliary/json.hh | 14 +- internal/cgen/data/data.go | 37 +-- release/c/wuffs-unsupported-snapshot.c | 336 ++++++++++++++++++++++--- 5 files changed, 638 insertions(+), 82 deletions(-) diff --git a/example/jsonfindptrs/jsonfindptrs.cc b/example/jsonfindptrs/jsonfindptrs.cc index c26ff4ed..47942798 100644 --- a/example/jsonfindptrs/jsonfindptrs.cc +++ b/example/jsonfindptrs/jsonfindptrs.cc @@ -106,6 +106,7 @@ static const char* g_usage = "\n" "Flags:\n" " -d=NUM -max-output-depth=NUM\n" + " -q=STR -query=STR\n" " -input-allow-comments\n" " -input-allow-extra-comma\n" " -input-allow-inf-nan-numbers\n" @@ -193,6 +194,7 @@ struct { char** remaining_argv; uint32_t max_output_depth; + char* query_c_string; bool strict_json_pointer_syntax; } g_flags = {0}; @@ -249,6 +251,12 @@ parse_flags(int argc, char** argv) { g_quirks.push_back(WUFFS_JSON__QUIRK_ALLOW_INF_NAN_NUMBERS); continue; } + if (!strncmp(arg, "q=", 2) || !strncmp(arg, "query=", 6)) { + while (*arg++ != '=') { + } + g_flags.query_c_string = arg; + continue; + } if (!strcmp(arg, "strict-json-pointer-syntax")) { g_flags.strict_json_pointer_syntax = true; continue; @@ -506,7 +514,8 @@ main1(int argc, char** argv) { return wuffs_aux::DecodeJson( Callbacks(), wuffs_aux::sync_io::FileInput(in), - wuffs_base__make_slice_u32(g_quirks.data(), g_quirks.size())) + wuffs_base__make_slice_u32(g_quirks.data(), g_quirks.size()), + (g_flags.query_c_string ? g_flags.query_c_string : "")) .error_message; } diff --git a/internal/cgen/auxiliary/json.cc b/internal/cgen/auxiliary/json.cc index 1613ff5f..bcd3908e 100644 --- a/internal/cgen/auxiliary/json.cc +++ b/internal/cgen/auxiliary/json.cc @@ -29,7 +29,7 @@ DecodeJsonResult::DecodeJsonResult(std::string&& error_message0, std::string // DecodeJsonCallbacks::AppendByteString(std::string&& val) { - return "wuffs_aux::JsonDecoder: unexpected JSON byte string"; + return "wuffs_aux::DecodeJson: unexpected JSON byte string"; } void // @@ -37,6 +37,10 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, sync_io::Input& input, IOBuffer& buffer) {} +const char DecodeJson_BadJsonPointer[] = + "wuffs_aux::DecodeJson: bad JSON Pointer"; +const char DecodeJson_NoMatch[] = "wuffs_aux::DecodeJson: no match"; + #define WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN \ while (tok_buf.meta.ri >= tok_buf.meta.wi) { \ if (tok_status.repr == nullptr) { \ @@ -48,17 +52,17 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, goto done; \ } else if (cursor_index != io_buf->meta.ri) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: bad cursor_index"; \ + "wuffs_aux::DecodeJson: internal error: bad cursor_index"; \ goto done; \ } else if (io_buf->meta.closed) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: io_buf is closed"; \ + "wuffs_aux::DecodeJson: internal error: io_buf is closed"; \ goto done; \ } \ io_buf->compact(); \ if (io_buf->meta.wi >= io_buf->data.len) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: io_buf is full"; \ + "wuffs_aux::DecodeJson: internal error: io_buf is full"; \ goto done; \ } \ cursor_index = io_buf->meta.ri; \ @@ -69,7 +73,7 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, } \ if (WUFFS_JSON__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE != 0) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: bad WORKBUF_LEN"; \ + "wuffs_aux::DecodeJson: internal error: bad WORKBUF_LEN"; \ goto done; \ } \ wuffs_base__slice_u8 work_buf = wuffs_base__empty_slice_u8(); \ @@ -80,16 +84,268 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, if ((io_buf->meta.ri < cursor_index) || \ ((io_buf->meta.ri - cursor_index) < token_len)) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: bad token indexes"; \ + "wuffs_aux::DecodeJson: internal error: bad token indexes"; \ goto done; \ } \ uint8_t* token_ptr = io_buf->data.ptr + cursor_index; \ cursor_index += token_len +namespace { + +// DecodeJson_SplitJsonPointer returns ("bar", 8) for ("/foo/bar/baz/qux", 5). +// It returns a 0 size_t when s has invalid JSON Pointer syntax. +std::pair // +DecodeJson_SplitJsonPointer(std::string& s, size_t i) { + std::string fragment; + while (i < s.size()) { + char c = s[i]; + if (c == '/') { + break; + } else if (c != '~') { + fragment.push_back(c); + i++; + continue; + } + i++; + if (i >= s.size()) { + return std::make_pair(std::string(), 0); + } + c = s[i]; + if (c == '0') { + fragment.push_back('~'); + i++; + continue; + } else if (c == '1') { + fragment.push_back('/'); + i++; + continue; + } + return std::make_pair(std::string(), 0); + } + return std::make_pair(std::move(fragment), i); +} + +std::string // +DecodeJson_DecodeBackslashX(std::string& str, + uint8_t* token_ptr, + size_t token_len) { + wuffs_base__slice_u8 encoded = + wuffs_base__make_slice_u8(token_ptr, token_len); + while (encoded.len > 0) { + uint8_t decoded[64]; + constexpr bool src_closed = true; + wuffs_base__transform__output o = wuffs_base__base_16__decode4( + wuffs_base__make_slice_u8(&decoded[0], sizeof decoded), encoded, + src_closed, WUFFS_BASE__BASE_16__DEFAULT_OPTIONS); + if (o.status.is_error()) { + return o.status.message(); + } else if ((o.num_dst > (sizeof decoded)) || (o.num_src > encoded.len)) { + return "wuffs_aux::DecodeJson: internal error: inconsistent base16 " + "decoding"; + } + str.append( // Convert from (uint8_t*). + static_cast(static_cast(&decoded[0])), o.num_dst); + encoded.ptr += o.num_src; + encoded.len -= o.num_src; + } + return ""; +} + +std::string // +DecodeJson_WalkJsonPointerFragment(wuffs_base__token_buffer& tok_buf, + wuffs_base__status& tok_status, + wuffs_json__decoder::unique_ptr& dec, + wuffs_base__io_buffer* io_buf, + std::string& io_error_message, + size_t& cursor_index, + sync_io::Input& input, + std::string& json_pointer_fragment) { + std::string ret_error_message; + while (true) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (vbc == WUFFS_BASE__TOKEN__VBC__FILLER) { + continue; + } else if ((vbc != WUFFS_BASE__TOKEN__VBC__STRUCTURE) || + !(vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH)) { + return DecodeJson_NoMatch; + } else if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_LIST) { + goto do_list; + } + goto do_dict; + } + +do_dict: + // Alternate between these two things: + // 1. Decode the next dict key (a string). If it matches the fragment, we're + // done (success). If we've reached the dict's end (VBD__STRUCTURE__POP) + // so that there was no next dict key, we're done (failure). + // 2. Otherwise, skip the next dict value. + while (true) { + for (std::string str; true;) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + switch (vbc) { + case WUFFS_BASE__TOKEN__VBC__FILLER: + continue; + + case WUFFS_BASE__TOKEN__VBC__STRUCTURE: + if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) { + goto fail; + } + return DecodeJson_NoMatch; + + case WUFFS_BASE__TOKEN__VBC__STRING: { + if (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP) { + // No-op. + } else if (vbd & + WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY) { + const char* ptr = // Convert from (uint8_t*). + static_cast(static_cast(token_ptr)); + str.append(ptr, token_len); + } else if ( + vbd & + WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) { + ret_error_message = + DecodeJson_DecodeBackslashX(str, token_ptr, token_len); + if (!ret_error_message.empty()) { + goto done; + } + } else { + goto fail; + } + break; + } + + case WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT: { + uint8_t u[WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL]; + size_t n = wuffs_base__utf_8__encode( + wuffs_base__make_slice_u8( + &u[0], WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL), + static_cast(vbd)); + const char* ptr = // Convert from (uint8_t*). + static_cast(static_cast(&u[0])); + str.append(ptr, n); + break; + } + + default: + goto fail; + } + + if (token.continued()) { + continue; + } + if (str == json_pointer_fragment) { + return ""; + } + goto skip_the_next_dict_value; + } + + skip_the_next_dict_value: + for (uint32_t skip_depth = 0; true;) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (token.continued() || (vbc == WUFFS_BASE__TOKEN__VBC__FILLER)) { + continue; + } else if (vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) { + if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) { + skip_depth++; + continue; + } + skip_depth--; + } + + if (skip_depth == 0) { + break; + } + } // skip_the_next_dict_value + } // do_dict + +do_list: + do { + wuffs_base__result_u64 result_u64 = wuffs_base__parse_number_u64( + wuffs_base__make_slice_u8( + static_cast(static_cast( + const_cast(json_pointer_fragment.data()))), + json_pointer_fragment.size()), + WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS); + if (!result_u64.status.is_ok()) { + return DecodeJson_NoMatch; + } + uint64_t remaining = result_u64.value; + if (remaining == 0) { + goto check_that_a_value_follows; + } + for (uint32_t skip_depth = 0; true;) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (token.continued() || (vbc == WUFFS_BASE__TOKEN__VBC__FILLER)) { + continue; + } else if (vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) { + if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) { + skip_depth++; + continue; + } + if (skip_depth == 0) { + return DecodeJson_NoMatch; + } + skip_depth--; + } + + if (skip_depth > 0) { + continue; + } + remaining--; + if (remaining == 0) { + goto check_that_a_value_follows; + } + } + } while (false); // do_list + +check_that_a_value_follows: + while (true) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (vbc == WUFFS_BASE__TOKEN__VBC__FILLER) { + continue; + } + + // Undo the last part of WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN, so that + // we're only peeking at the next token. + tok_buf.meta.ri--; + cursor_index -= token_len; + + if ((vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) && + (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__POP)) { + return DecodeJson_NoMatch; + } + return ""; + } // check_that_a_value_follows + +fail: + return "wuffs_aux::DecodeJson: internal error: unexpected token"; +done: + return ret_error_message; +} + +} // namespace + DecodeJsonResult // DecodeJson(DecodeJsonCallbacks&& callbacks, sync_io::Input&& input, - wuffs_base__slice_u32 quirks) { + wuffs_base__slice_u32 quirks, + std::string json_pointer) { // Prepare the wuffs_base__io_buffer and the resultant error_message. wuffs_base__io_buffer* io_buf = input.BringsItsOwnIOBuffer(); wuffs_base__io_buffer fallback_io_buf = wuffs_base__empty_io_buffer(); @@ -107,7 +363,7 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, // Prepare the low-level JSON decoder. wuffs_json__decoder::unique_ptr dec = wuffs_json__decoder::alloc(); if (!dec) { - ret_error_message = "wuffs_aux::JsonDecoder: out of memory"; + ret_error_message = "wuffs_aux::DecodeJson: out of memory"; goto done; } for (size_t i = 0; i < quirks.len; i++) { @@ -125,6 +381,27 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, uint32_t depth = 0; std::string str; + // Walk the (optional) JSON Pointer. + for (size_t i = 0; i < json_pointer.size();) { + if (json_pointer[i] != '/') { + ret_error_message = DecodeJson_BadJsonPointer; + goto done; + } + std::pair split = + DecodeJson_SplitJsonPointer(json_pointer, i + 1); + i = std::move(split.second); + if (i == 0) { + ret_error_message = DecodeJson_BadJsonPointer; + goto done; + } + ret_error_message = DecodeJson_WalkJsonPointerFragment( + tok_buf, tok_status, dec, io_buf, io_error_message, cursor_index, + input, split.first); + if (!ret_error_message.empty()) { + goto done; + } + } + // Loop, doing these two things: // 1. Get the next token. // 2. Process that token. @@ -162,29 +439,10 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, } else if ( vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) { - wuffs_base__slice_u8 encoded = - wuffs_base__make_slice_u8(token_ptr, token_len); - while (encoded.len > 0) { - uint8_t decoded[64]; - constexpr bool src_closed = true; - wuffs_base__transform__output o = wuffs_base__base_16__decode4( - wuffs_base__make_slice_u8(&decoded[0], sizeof decoded), - encoded, src_closed, WUFFS_BASE__BASE_16__DEFAULT_OPTIONS); - if (o.status.is_error()) { - ret_error_message = o.status.message(); - goto done; - } else if ((o.num_dst > (sizeof decoded)) || - (o.num_src > encoded.len)) { - ret_error_message = - "wuffs_aux::JsonDecoder: internal error: inconsistent " - "base16 decoding"; - goto done; - } - str.append( // Convert from (uint8_t*). - static_cast(static_cast(&decoded[0])), - o.num_dst); - encoded.ptr += o.num_src; - encoded.len -= o.num_src; + ret_error_message = + DecodeJson_DecodeBackslashX(str, token_ptr, token_len); + if (!ret_error_message.empty()) { + goto done; } } else { goto fail; @@ -271,7 +529,7 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, fail: ret_error_message = - "wuffs_aux::JsonDecoder: internal error: unexpected token"; + "wuffs_aux::DecodeJson: internal error: unexpected token"; goto done; parsed_a_value: diff --git a/internal/cgen/auxiliary/json.hh b/internal/cgen/auxiliary/json.hh index 413c45d4..89998d02 100644 --- a/internal/cgen/auxiliary/json.hh +++ b/internal/cgen/auxiliary/json.hh @@ -78,15 +78,27 @@ class DecodeJsonCallbacks { IOBuffer& buffer); }; +extern const char DecodeJson_BadJsonPointer[]; +extern const char DecodeJson_NoMatch[]; + // DecodeJson calls callbacks based on the JSON-formatted data in input. // // On success, the returned error_message is empty and cursor_position counts // the number of bytes consumed. On failure, error_message is non-empty and // cursor_position is the location of the error. That error may be a content // error (invalid JSON) or an input error (e.g. network failure). +// +// json_pointer is a query in the JSON Pointer (RFC 6901) syntax. The callbacks +// run for the input's sub-node that matches the query. DecodeJson_NoMatch is +// returned if no matching sub-node was found. The empty query matches the +// input's root node, consistent with JSON Pointer semantics. +// +// The JSON Pointer implementation is greedy: duplicate keys are not rejected +// but only the first match for each '/'-separated fragment is followed. DecodeJsonResult DecodeJson( DecodeJsonCallbacks&& callbacks, sync_io::Input&& input, - wuffs_base__slice_u32 quirks = wuffs_base__empty_slice_u32()); + wuffs_base__slice_u32 quirks = wuffs_base__empty_slice_u32(), + std::string json_pointer = std::string()); } // namespace wuffs_aux diff --git a/internal/cgen/data/data.go b/internal/cgen/data/data.go index d0fbc0f6..97a3686d 100644 --- a/internal/cgen/data/data.go +++ b/internal/cgen/data/data.go @@ -648,26 +648,33 @@ const AuxCborHh = "" + "" const AuxJsonCc = "" + - "// ---------------- Auxiliary - JSON\n\n#if !defined(WUFFS_CONFIG__MODULES) || defined(WUFFS_CONFIG__MODULE__AUX__JSON)\n\n#include \n\nnamespace wuffs_aux {\n\nDecodeJsonResult::DecodeJsonResult(std::string&& error_message0,\n uint64_t cursor_position0)\n : error_message(std::move(error_message0)),\n cursor_position(cursor_position0) {}\n\nstd::string //\nDecodeJsonCallbacks::AppendByteString(std::string&& val) {\n return \"wuffs_aux::JsonDecoder: unexpected JSON byte string\";\n}\n\nvoid //\nDecodeJsonCallbacks::Done(DecodeJsonResult& result,\n sync_io::Input& input,\n IOBuffer& buffer) {}\n\n#define WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN \\\n while (tok_buf.meta.ri >= tok_buf.meta.wi) { \\\n if (tok_status.repr == nullptr) { \\\n } else if (tok_status.repr == wuffs_base__suspension__short_write) { \\\n tok_buf.compact(); " + - " \\\n } else if (tok_status.repr == wuffs_base__suspension__short_read) { \\\n if (!io_error_message.empty()) { \\\n ret_error_message = std::move(io_error_message); \\\n goto done; \\\n } else if (cursor_index != io_buf->meta.ri) { \\\n ret_error_message = \\\n \"wuffs_aux::JsonDecoder: internal error: bad cursor_index\"; \\\n goto done; \\\n } else if (io_buf->meta.closed) { \\\n ret_error_message = \\\n \"wuffs_aux::JsonDecoder: internal error: io_buf is closed\"; \\\n goto done; \\\n } \\\n io_buf->com" + - "pact(); \\\n if (io_buf->meta.wi >= io_buf->data.len) { \\\n ret_error_message = \\\n \"wuffs_aux::JsonDecoder: internal error: io_buf is full\"; \\\n goto done; \\\n } \\\n cursor_index = io_buf->meta.ri; \\\n io_error_message = input.CopyIn(io_buf); \\\n } else { \\\n ret_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::JsonDecoder: 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::JsonDecoder: internal error: bad token indexes\"; \\\n goto done; \\\n } " + - " \\\n uint8_t* token_ptr = io_buf->data.ptr + cursor_index; \\\n cursor_index += token_len\n\nDecodeJsonResult //\nDecodeJson(DecodeJsonCallbacks&& callbacks,\n sync_io::Input&& input,\n wuffs_base__slice_u32 quirks) {\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 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::Jso" + - "nDecoder: out of memory\";\n goto done;\n }\n for (size_t i = 0; i < quirks.len; i++) {\n dec->set_quirk_enabled(quirks.ptr[i], true);\n }\n\n // Prepare the wuffs_base__tok_buffer.\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 // 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 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 if (\n vbd &\n WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) {\n wuffs_base__slice_u8 encoded =\n wuffs_base__make_slice_u8(token_ptr, token_len);\n " + - " while (encoded.len > 0) {\n uint8_t decoded[64];\n constexpr bool src_closed = true;\n wuffs_base__transform__output o = wuffs_base__base_16__decode4(\n wuffs_base__make_slice_u8(&decoded[0], sizeof decoded),\n encoded, src_closed, WUFFS_BASE__BASE_16__DEFAULT_OPTIONS);\n if (o.status.is_error()) {\n ret_error_message = o.status.message();\n goto done;\n } else if ((o.num_dst > (sizeof decoded)) ||\n (o.num_src > encoded.len)) {\n ret_error_message =\n \"wuffs_aux::JsonDecoder: internal error: inconsistent \"\n \"base16 decoding\";\n goto done;\n }\n str.append( // Convert from (uint8_t*).\n static_cast(static_cast(&decoded[0])),\n o.num_dst);\n encoded.ptr += o.num_src;\n encoded.len -= o.num_src;\n " + - " }\n } else {\n goto fail;\n }\n if (token.continued()) {\n continue;\n }\n ret_error_message =\n (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CHAIN_MUST_BE_UTF_8)\n ? callbacks.AppendTextString(std::move(str))\n : callbacks.AppendByteString(std::move(str));\n str.clear();\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__UNICODE_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_ba" + - "se__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::JsonDecoder: 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 // !def" + - "ined(WUFFS_CONFIG__MODULES) ||\n // defined(WUFFS_CONFIG__MODULE__AUX__JSON)\n" + + "// ---------------- Auxiliary - JSON\n\n#if !defined(WUFFS_CONFIG__MODULES) || defined(WUFFS_CONFIG__MODULE__AUX__JSON)\n\n#include \n\nnamespace wuffs_aux {\n\nDecodeJsonResult::DecodeJsonResult(std::string&& error_message0,\n uint64_t cursor_position0)\n : error_message(std::move(error_message0)),\n cursor_position(cursor_position0) {}\n\nstd::string //\nDecodeJsonCallbacks::AppendByteString(std::string&& val) {\n return \"wuffs_aux::DecodeJson: unexpected JSON byte string\";\n}\n\nvoid //\nDecodeJsonCallbacks::Done(DecodeJsonResult& result,\n sync_io::Input& input,\n IOBuffer& buffer) {}\n\nconst char DecodeJson_BadJsonPointer[] =\n \"wuffs_aux::DecodeJson: bad JSON Pointer\";\nconst char DecodeJson_NoMatch[] = \"wuffs_aux::DecodeJson: no match\";\n\n#define WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN \\\n while (tok_buf.meta.ri >= tok_buf.meta.wi) { \\\n if (tok_status.repr == nullptr) " + + "{ \\\n } else if (tok_status.repr == wuffs_base__suspension__short_write) { \\\n tok_buf.compact(); \\\n } else if (tok_status.repr == wuffs_base__suspension__short_read) { \\\n if (!io_error_message.empty()) { \\\n ret_error_message = std::move(io_error_message); \\\n goto done; \\\n } else if (cursor_index != io_buf->meta.ri) { \\\n ret_error_message = \\\n \"wuffs_aux::DecodeJson: internal error: bad cursor_index\"; \\\n goto done; \\\n } else if (io_buf->meta.closed) { \\\n ret_error_message = \\\n \"wuffs_aux::DecodeJson: internal error: io_buf is closed\"; \\\n go" + + "to done; \\\n } \\\n io_buf->compact(); \\\n if (io_buf->meta.wi >= io_buf->data.len) { \\\n ret_error_message = \\\n \"wuffs_aux::DecodeJson: internal error: io_buf is full\"; \\\n goto done; \\\n } \\\n cursor_index = io_buf->meta.ri; \\\n io_error_message = input.CopyIn(io_buf); \\\n } else { \\\n ret_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: i" + + "nternal error: bad token indexes\"; \\\n goto done; \\\n } \\\n uint8_t* token_ptr = io_buf->data.ptr + cursor_index; \\\n cursor_index += token_len\n\nnamespace {\n\n// DecodeJson_SplitJsonPointer returns (\"bar\", 8) for (\"/foo/bar/baz/qux\", 5).\n// It returns a 0 size_t when s has invalid JSON Pointer syntax.\nstd::pair //\nDecodeJson_SplitJsonPointer(std::string& s, size_t i) {\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 }\n return std::make" + + "_pair(std::string(), 0);\n }\n return std::make_pair(std::move(fragment), i);\n}\n\nstd::string //\nDecodeJson_DecodeBackslashX(std::string& str,\n uint8_t* token_ptr,\n size_t token_len) {\n wuffs_base__slice_u8 encoded =\n wuffs_base__make_slice_u8(token_ptr, token_len);\n while (encoded.len > 0) {\n uint8_t decoded[64];\n constexpr bool src_closed = true;\n wuffs_base__transform__output o = wuffs_base__base_16__decode4(\n wuffs_base__make_slice_u8(&decoded[0], sizeof decoded), encoded,\n src_closed, WUFFS_BASE__BASE_16__DEFAULT_OPTIONS);\n if (o.status.is_error()) {\n return o.status.message();\n } else if ((o.num_dst > (sizeof decoded)) || (o.num_src > encoded.len)) {\n return \"wuffs_aux::DecodeJson: internal error: inconsistent base16 \"\n \"decoding\";\n }\n str.append( // Convert from (uint8_t*).\n static_cast(static_cast(&decoded[0])), o.num_dst);\n encoded.ptr += o.num_src;\n " + + "encoded.len -= o.num_src;\n }\n return \"\";\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__STRUCTURE__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 } 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 if (\n vbd &\n WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) {\n ret_error_message =\n DecodeJson_DecodeBackslashX(str, token_ptr, token_len);\n if (!ret_error_message.empty()) {\n goto done;\n }\n } else {\n goto fail;\n }\n break;\n }\n\n case WUFFS_BASE__TOKEN__VBC__UNICODE_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 break;\n }\n\n default:\n goto fail;\n }\n\n if (token.continued()) {\n continue;\n }\n if (str == json_pointer_fragment) {\n return \"\";\n }\n goto skip_the_next_dict_value;\n }\n\n skip_the_next_dict_value:\n for (uint32_t skip_depth = 0; 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 (token.continued() || (vbc == WUFFS_BASE__TOKEN__VBC__FILLER)) {\n continue;\n } else if (vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) {\n skip_depth++;\n continue;\n }\n skip_depth--;\n }\n\n if (skip_depth == 0) {\n break;\n }\n } // skip_the_next_dict_value\n } // do_dict\n\ndo_list:\n " + + "do {\n wuffs_base__result_u64 result_u64 = wuffs_base__parse_number_u64(\n wuffs_base__make_slice_u8(\n static_cast(static_cast(\n const_cast(json_pointer_fragment.data()))),\n json_pointer_fragment.size()),\n WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS);\n if (!result_u64.status.is_ok()) {\n return DecodeJson_NoMatch;\n }\n uint64_t remaining = result_u64.value;\n if (remaining == 0) {\n goto check_that_a_value_follows;\n }\n for (uint32_t skip_depth = 0; 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 (token.continued() || (vbc == WUFFS_BASE__TOKEN__VBC__FILLER)) {\n continue;\n } else if (vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) {\n if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) {\n skip_depth++;\n continue;\n }\n if (skip_depth == 0) {\n ret" + + "urn DecodeJson_NoMatch;\n }\n skip_depth--;\n }\n\n if (skip_depth > 0) {\n continue;\n }\n remaining--;\n if (remaining == 0) {\n goto check_that_a_value_follows;\n }\n }\n } while (false); // do_list\n\ncheck_that_a_value_follows:\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 }\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\nD" + + "ecodeJsonResult //\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 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 for (size_t i = 0; i < quirks.len; i++) {\n dec" + + "->set_quirk_enabled(quirks.ptr[i], true);\n }\n\n // Prepare the wuffs_base__tok_buffer.\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 split =\n DecodeJson_SplitJsonPointer(json_pointer, i + 1);\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 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__T" + + "OKEN__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 if (\n vbd &\n WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) {\n ret_error_message =\n DecodeJson_DecodeBackslashX(str, token_ptr, token_len);\n if (!ret_error_message.empty()) {\n goto done;\n }\n } else {\n goto fail;\n }\n if (token.continued()) {\n continue;\n }\n ret_error_message =\n (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CHAIN_MUST_BE_UTF_8)\n ? callbacks.AppendTextString(std::move(str))\n : callbacks.AppendByteString(std::move(str));\n " + + " str.clear();\n goto parsed_a_value;\n }\n\n case WUFFS_BASE__TOKEN__VBC__UNICODE_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__NUM" + + "BER: {\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__T" + + "OKEN__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 = "" + "// ---------------- Auxiliary - JSON\n\nnamespace wuffs_aux {\n\nstruct DecodeJsonResult {\n DecodeJsonResult(std::string&& error_message0, uint64_t cursor_position0);\n\n std::string error_message;\n uint64_t cursor_position;\n};\n\nclass DecodeJsonCallbacks {\n public:\n // AppendXxx are called for leaf nodes: literals, numbers and strings. For\n // strings, the Callbacks implementation is responsible for tracking map keys\n // versus other values.\n\n // The JSON file format as specified deals only with (UTF-8) text strings,\n // but an unofficial extension allows \"ijk\\x89m\" escapes within those\n // strings. DecodeJsonCallbacks' AppendByteString will not be called unless\n // WUFFS_JSON__QUIRK_ALLOW_BACKSLASH_X_AS_BYTES is passed to DecodeJson. If\n // it is passed, AppendTextString will not be called and all byte strings are\n // potentially invalid UTF-8. It is up to the AppendByteString implementation\n // whether to test the std::string for UTF-8 validity.\n //\n // The default AppendByteString implementation r" + "eturns an error message.\n\n virtual std::string AppendNull() = 0;\n virtual std::string AppendBool(bool val) = 0;\n virtual std::string AppendF64(double val) = 0;\n virtual std::string AppendI64(int64_t val) = 0;\n virtual std::string AppendByteString(std::string&& val);\n virtual std::string AppendTextString(std::string&& val) = 0;\n\n // Push and Pop are called for container nodes: JSON arrays (lists) and JSON\n // objects (dictionaries).\n //\n // The flags bits combine exactly one of:\n // - WUFFS_BASE__TOKEN__VBD__STRUCTURE__FROM_NONE\n // - WUFFS_BASE__TOKEN__VBD__STRUCTURE__FROM_LIST\n // - WUFFS_BASE__TOKEN__VBD__STRUCTURE__FROM_DICT\n // and exactly one of:\n // - WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_NONE\n // - WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_LIST\n // - WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_DICT\n\n virtual std::string Push(uint32_t flags) = 0;\n virtual std::string Pop(uint32_t flags) = 0;\n\n // Done is always the last Callback method called by DecodeJson, whether or\n // not parsing th" + - "e input as JSON encountered an error. Even when successful,\n // trailing data may remain in input and buffer. See \"Unintuitive JSON\n // Parsing\" (https://nullprogram.com/blog/2019/12/28/) which discusses JSON\n // parsing and when it stops.\n //\n // Do not keep a reference to buffer or buffer.data.ptr after Done returns,\n // as DecodeJson may then de-allocate the backing array.\n //\n // The default Done implementation is a no-op.\n virtual void Done(DecodeJsonResult& result,\n sync_io::Input& input,\n IOBuffer& buffer);\n};\n\n// DecodeJson calls callbacks based on the JSON-formatted data in input.\n//\n// On success, the returned error_message is empty and cursor_position counts\n// the number of bytes consumed. On failure, error_message is non-empty and\n// cursor_position is the location of the error. That error may be a content\n// error (invalid JSON) or an input error (e.g. network failure).\nDecodeJsonResult DecodeJson(\n DecodeJsonCallbacks&& callbacks,\n sync_io::I" + - "nput&& input,\n wuffs_base__slice_u32 quirks = wuffs_base__empty_slice_u32());\n\n} // namespace wuffs_aux\n" + + "e input as JSON encountered an error. Even when successful,\n // trailing data may remain in input and buffer. See \"Unintuitive JSON\n // Parsing\" (https://nullprogram.com/blog/2019/12/28/) which discusses JSON\n // parsing and when it stops.\n //\n // Do not keep a reference to buffer or buffer.data.ptr after Done returns,\n // as DecodeJson may then de-allocate the backing array.\n //\n // The default Done implementation is a no-op.\n virtual void Done(DecodeJsonResult& result,\n sync_io::Input& input,\n IOBuffer& buffer);\n};\n\nextern const char DecodeJson_BadJsonPointer[];\nextern const char DecodeJson_NoMatch[];\n\n// DecodeJson calls callbacks based on the JSON-formatted data in input.\n//\n// On success, the returned error_message is empty and cursor_position counts\n// the number of bytes consumed. On failure, error_message is non-empty and\n// cursor_position is the location of the error. That error may be a content\n// error (invalid JSON) or an input error (e.g. network fa" + + "ilure).\n//\n// json_pointer is a query in the JSON Pointer (RFC 6901) syntax. The callbacks\n// run for the input's sub-node that matches the query. DecodeJson_NoMatch is\n// returned if no matching sub-node was found. The empty query matches the\n// input's root node, consistent with JSON Pointer semantics.\n//\n// The JSON Pointer implementation is greedy: duplicate keys are not rejected\n// but only the first match for each '/'-separated fragment is followed.\nDecodeJsonResult DecodeJson(\n DecodeJsonCallbacks&& callbacks,\n sync_io::Input&& input,\n wuffs_base__slice_u32 quirks = wuffs_base__empty_slice_u32(),\n std::string json_pointer = std::string());\n\n} // namespace wuffs_aux\n" + "" var AuxNonBaseCcFiles = []string{ diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index 80f5875e..6e58bbcb 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -8604,16 +8604,28 @@ class DecodeJsonCallbacks { IOBuffer& buffer); }; +extern const char DecodeJson_BadJsonPointer[]; +extern const char DecodeJson_NoMatch[]; + // DecodeJson calls callbacks based on the JSON-formatted data in input. // // On success, the returned error_message is empty and cursor_position counts // the number of bytes consumed. On failure, error_message is non-empty and // cursor_position is the location of the error. That error may be a content // error (invalid JSON) or an input error (e.g. network failure). +// +// json_pointer is a query in the JSON Pointer (RFC 6901) syntax. The callbacks +// run for the input's sub-node that matches the query. DecodeJson_NoMatch is +// returned if no matching sub-node was found. The empty query matches the +// input's root node, consistent with JSON Pointer semantics. +// +// The JSON Pointer implementation is greedy: duplicate keys are not rejected +// but only the first match for each '/'-separated fragment is followed. DecodeJsonResult DecodeJson( DecodeJsonCallbacks&& callbacks, sync_io::Input&& input, - wuffs_base__slice_u32 quirks = wuffs_base__empty_slice_u32()); + wuffs_base__slice_u32 quirks = wuffs_base__empty_slice_u32(), + std::string json_pointer = std::string()); } // namespace wuffs_aux @@ -29355,7 +29367,7 @@ DecodeJsonResult::DecodeJsonResult(std::string&& error_message0, std::string // DecodeJsonCallbacks::AppendByteString(std::string&& val) { - return "wuffs_aux::JsonDecoder: unexpected JSON byte string"; + return "wuffs_aux::DecodeJson: unexpected JSON byte string"; } void // @@ -29363,6 +29375,10 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, sync_io::Input& input, IOBuffer& buffer) {} +const char DecodeJson_BadJsonPointer[] = + "wuffs_aux::DecodeJson: bad JSON Pointer"; +const char DecodeJson_NoMatch[] = "wuffs_aux::DecodeJson: no match"; + #define WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN \ while (tok_buf.meta.ri >= tok_buf.meta.wi) { \ if (tok_status.repr == nullptr) { \ @@ -29374,17 +29390,17 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, goto done; \ } else if (cursor_index != io_buf->meta.ri) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: bad cursor_index"; \ + "wuffs_aux::DecodeJson: internal error: bad cursor_index"; \ goto done; \ } else if (io_buf->meta.closed) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: io_buf is closed"; \ + "wuffs_aux::DecodeJson: internal error: io_buf is closed"; \ goto done; \ } \ io_buf->compact(); \ if (io_buf->meta.wi >= io_buf->data.len) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: io_buf is full"; \ + "wuffs_aux::DecodeJson: internal error: io_buf is full"; \ goto done; \ } \ cursor_index = io_buf->meta.ri; \ @@ -29395,7 +29411,7 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, } \ if (WUFFS_JSON__DECODER_WORKBUF_LEN_MAX_INCL_WORST_CASE != 0) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: bad WORKBUF_LEN"; \ + "wuffs_aux::DecodeJson: internal error: bad WORKBUF_LEN"; \ goto done; \ } \ wuffs_base__slice_u8 work_buf = wuffs_base__empty_slice_u8(); \ @@ -29406,16 +29422,268 @@ DecodeJsonCallbacks::Done(DecodeJsonResult& result, if ((io_buf->meta.ri < cursor_index) || \ ((io_buf->meta.ri - cursor_index) < token_len)) { \ ret_error_message = \ - "wuffs_aux::JsonDecoder: internal error: bad token indexes"; \ + "wuffs_aux::DecodeJson: internal error: bad token indexes"; \ goto done; \ } \ uint8_t* token_ptr = io_buf->data.ptr + cursor_index; \ cursor_index += token_len +namespace { + +// DecodeJson_SplitJsonPointer returns ("bar", 8) for ("/foo/bar/baz/qux", 5). +// It returns a 0 size_t when s has invalid JSON Pointer syntax. +std::pair // +DecodeJson_SplitJsonPointer(std::string& s, size_t i) { + std::string fragment; + while (i < s.size()) { + char c = s[i]; + if (c == '/') { + break; + } else if (c != '~') { + fragment.push_back(c); + i++; + continue; + } + i++; + if (i >= s.size()) { + return std::make_pair(std::string(), 0); + } + c = s[i]; + if (c == '0') { + fragment.push_back('~'); + i++; + continue; + } else if (c == '1') { + fragment.push_back('/'); + i++; + continue; + } + return std::make_pair(std::string(), 0); + } + return std::make_pair(std::move(fragment), i); +} + +std::string // +DecodeJson_DecodeBackslashX(std::string& str, + uint8_t* token_ptr, + size_t token_len) { + wuffs_base__slice_u8 encoded = + wuffs_base__make_slice_u8(token_ptr, token_len); + while (encoded.len > 0) { + uint8_t decoded[64]; + constexpr bool src_closed = true; + wuffs_base__transform__output o = wuffs_base__base_16__decode4( + wuffs_base__make_slice_u8(&decoded[0], sizeof decoded), encoded, + src_closed, WUFFS_BASE__BASE_16__DEFAULT_OPTIONS); + if (o.status.is_error()) { + return o.status.message(); + } else if ((o.num_dst > (sizeof decoded)) || (o.num_src > encoded.len)) { + return "wuffs_aux::DecodeJson: internal error: inconsistent base16 " + "decoding"; + } + str.append( // Convert from (uint8_t*). + static_cast(static_cast(&decoded[0])), o.num_dst); + encoded.ptr += o.num_src; + encoded.len -= o.num_src; + } + return ""; +} + +std::string // +DecodeJson_WalkJsonPointerFragment(wuffs_base__token_buffer& tok_buf, + wuffs_base__status& tok_status, + wuffs_json__decoder::unique_ptr& dec, + wuffs_base__io_buffer* io_buf, + std::string& io_error_message, + size_t& cursor_index, + sync_io::Input& input, + std::string& json_pointer_fragment) { + std::string ret_error_message; + while (true) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (vbc == WUFFS_BASE__TOKEN__VBC__FILLER) { + continue; + } else if ((vbc != WUFFS_BASE__TOKEN__VBC__STRUCTURE) || + !(vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH)) { + return DecodeJson_NoMatch; + } else if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_LIST) { + goto do_list; + } + goto do_dict; + } + +do_dict: + // Alternate between these two things: + // 1. Decode the next dict key (a string). If it matches the fragment, we're + // done (success). If we've reached the dict's end (VBD__STRUCTURE__POP) + // so that there was no next dict key, we're done (failure). + // 2. Otherwise, skip the next dict value. + while (true) { + for (std::string str; true;) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + switch (vbc) { + case WUFFS_BASE__TOKEN__VBC__FILLER: + continue; + + case WUFFS_BASE__TOKEN__VBC__STRUCTURE: + if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) { + goto fail; + } + return DecodeJson_NoMatch; + + case WUFFS_BASE__TOKEN__VBC__STRING: { + if (vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP) { + // No-op. + } else if (vbd & + WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY) { + const char* ptr = // Convert from (uint8_t*). + static_cast(static_cast(token_ptr)); + str.append(ptr, token_len); + } else if ( + vbd & + WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) { + ret_error_message = + DecodeJson_DecodeBackslashX(str, token_ptr, token_len); + if (!ret_error_message.empty()) { + goto done; + } + } else { + goto fail; + } + break; + } + + case WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT: { + uint8_t u[WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL]; + size_t n = wuffs_base__utf_8__encode( + wuffs_base__make_slice_u8( + &u[0], WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL), + static_cast(vbd)); + const char* ptr = // Convert from (uint8_t*). + static_cast(static_cast(&u[0])); + str.append(ptr, n); + break; + } + + default: + goto fail; + } + + if (token.continued()) { + continue; + } + if (str == json_pointer_fragment) { + return ""; + } + goto skip_the_next_dict_value; + } + + skip_the_next_dict_value: + for (uint32_t skip_depth = 0; true;) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (token.continued() || (vbc == WUFFS_BASE__TOKEN__VBC__FILLER)) { + continue; + } else if (vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) { + if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) { + skip_depth++; + continue; + } + skip_depth--; + } + + if (skip_depth == 0) { + break; + } + } // skip_the_next_dict_value + } // do_dict + +do_list: + do { + wuffs_base__result_u64 result_u64 = wuffs_base__parse_number_u64( + wuffs_base__make_slice_u8( + static_cast(static_cast( + const_cast(json_pointer_fragment.data()))), + json_pointer_fragment.size()), + WUFFS_BASE__PARSE_NUMBER_XXX__DEFAULT_OPTIONS); + if (!result_u64.status.is_ok()) { + return DecodeJson_NoMatch; + } + uint64_t remaining = result_u64.value; + if (remaining == 0) { + goto check_that_a_value_follows; + } + for (uint32_t skip_depth = 0; true;) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (token.continued() || (vbc == WUFFS_BASE__TOKEN__VBC__FILLER)) { + continue; + } else if (vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) { + if (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH) { + skip_depth++; + continue; + } + if (skip_depth == 0) { + return DecodeJson_NoMatch; + } + skip_depth--; + } + + if (skip_depth > 0) { + continue; + } + remaining--; + if (remaining == 0) { + goto check_that_a_value_follows; + } + } + } while (false); // do_list + +check_that_a_value_follows: + while (true) { + WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN; + + int64_t vbc = token.value_base_category(); + uint64_t vbd = token.value_base_detail(); + if (vbc == WUFFS_BASE__TOKEN__VBC__FILLER) { + continue; + } + + // Undo the last part of WUFFS_AUX__DECODE_JSON__GET_THE_NEXT_TOKEN, so that + // we're only peeking at the next token. + tok_buf.meta.ri--; + cursor_index -= token_len; + + if ((vbc == WUFFS_BASE__TOKEN__VBC__STRUCTURE) && + (vbd & WUFFS_BASE__TOKEN__VBD__STRUCTURE__POP)) { + return DecodeJson_NoMatch; + } + return ""; + } // check_that_a_value_follows + +fail: + return "wuffs_aux::DecodeJson: internal error: unexpected token"; +done: + return ret_error_message; +} + +} // namespace + DecodeJsonResult // DecodeJson(DecodeJsonCallbacks&& callbacks, sync_io::Input&& input, - wuffs_base__slice_u32 quirks) { + wuffs_base__slice_u32 quirks, + std::string json_pointer) { // Prepare the wuffs_base__io_buffer and the resultant error_message. wuffs_base__io_buffer* io_buf = input.BringsItsOwnIOBuffer(); wuffs_base__io_buffer fallback_io_buf = wuffs_base__empty_io_buffer(); @@ -29433,7 +29701,7 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, // Prepare the low-level JSON decoder. wuffs_json__decoder::unique_ptr dec = wuffs_json__decoder::alloc(); if (!dec) { - ret_error_message = "wuffs_aux::JsonDecoder: out of memory"; + ret_error_message = "wuffs_aux::DecodeJson: out of memory"; goto done; } for (size_t i = 0; i < quirks.len; i++) { @@ -29451,6 +29719,27 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, uint32_t depth = 0; std::string str; + // Walk the (optional) JSON Pointer. + for (size_t i = 0; i < json_pointer.size();) { + if (json_pointer[i] != '/') { + ret_error_message = DecodeJson_BadJsonPointer; + goto done; + } + std::pair split = + DecodeJson_SplitJsonPointer(json_pointer, i + 1); + i = std::move(split.second); + if (i == 0) { + ret_error_message = DecodeJson_BadJsonPointer; + goto done; + } + ret_error_message = DecodeJson_WalkJsonPointerFragment( + tok_buf, tok_status, dec, io_buf, io_error_message, cursor_index, + input, split.first); + if (!ret_error_message.empty()) { + goto done; + } + } + // Loop, doing these two things: // 1. Get the next token. // 2. Process that token. @@ -29488,29 +29777,10 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, } else if ( vbd & WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X) { - wuffs_base__slice_u8 encoded = - wuffs_base__make_slice_u8(token_ptr, token_len); - while (encoded.len > 0) { - uint8_t decoded[64]; - constexpr bool src_closed = true; - wuffs_base__transform__output o = wuffs_base__base_16__decode4( - wuffs_base__make_slice_u8(&decoded[0], sizeof decoded), - encoded, src_closed, WUFFS_BASE__BASE_16__DEFAULT_OPTIONS); - if (o.status.is_error()) { - ret_error_message = o.status.message(); - goto done; - } else if ((o.num_dst > (sizeof decoded)) || - (o.num_src > encoded.len)) { - ret_error_message = - "wuffs_aux::JsonDecoder: internal error: inconsistent " - "base16 decoding"; - goto done; - } - str.append( // Convert from (uint8_t*). - static_cast(static_cast(&decoded[0])), - o.num_dst); - encoded.ptr += o.num_src; - encoded.len -= o.num_src; + ret_error_message = + DecodeJson_DecodeBackslashX(str, token_ptr, token_len); + if (!ret_error_message.empty()) { + goto done; } } else { goto fail; @@ -29597,7 +29867,7 @@ DecodeJson(DecodeJsonCallbacks&& callbacks, fail: ret_error_message = - "wuffs_aux::JsonDecoder: internal error: unexpected token"; + "wuffs_aux::DecodeJson: internal error: unexpected token"; goto done; parsed_a_value: