From d1c928a1ed1abd72730b923ad6512b30429e3561 Mon Sep 17 00:00:00 2001 From: Nigel Tao Date: Fri, 28 Feb 2020 12:43:53 +1100 Subject: [PATCH] Have tokens include explicit link bits --- example/jsonptr/jsonptr.cc | 37 +- internal/cgen/base/token-public.h | 159 ++++--- internal/cgen/builtin.go | 24 +- internal/cgen/data.go | 20 +- lang/builtin/builtin.go | 8 +- release/c/wuffs-unsupported-snapshot.c | 396 ++++++++++-------- script/print-json-token-debug-format.c | 58 ++- std/json/decode_json.wuffs | 193 +++++---- test/c/testlib/testlib.c | 26 +- .../data/australian-abc-local-stations.tokens | Bin 28384 -> 67792 bytes test/data/json-things.unformatted.tokens | Bin 384 -> 480 bytes 11 files changed, 533 insertions(+), 388 deletions(-) diff --git a/example/jsonptr/jsonptr.cc b/example/jsonptr/jsonptr.cc index c3531cab..c4b28ce6 100644 --- a/example/jsonptr/jsonptr.cc +++ b/example/jsonptr/jsonptr.cc @@ -99,8 +99,6 @@ wuffs_base__token_buffer tok; // token. An invariant is that (curr_token_end_src_index <= src.meta.ri). size_t curr_token_end_src_index; -bool prev_token_incomplete; - uint64_t depth; enum class context { @@ -132,8 +130,6 @@ initialize_globals(int argc, char** argv) { curr_token_end_src_index = 0; - prev_token_incomplete = false; - depth = 0; ctx = context::none; @@ -319,19 +315,20 @@ handle_token(wuffs_base__token t) { goto after_value; } - // Write preceding whitespace and punctuation, if it wasn't ']' or '}'. - if (!prev_token_incomplete) { - if (ctx == context::in_dict_after_key) { - TRY(write_dst(": ", 2)); - } else if (ctx != context::none) { - if ((ctx != context::in_list_after_bracket) && - (ctx != context::in_dict_after_brace)) { - TRY(write_dst(",", 1)); - } - TRY(write_dst("\n", 1)); - for (size_t i = 0; i < depth; i++) { - TRY(write_dst(INDENT_STRING, indent)); - } + // Write preceding whitespace and punctuation, if it wasn't ']', '}' or a + // continuation of a multi-token chain. + if (t.link_prev()) { + // No-op. + } else if (ctx == context::in_dict_after_key) { + TRY(write_dst(": ", 2)); + } else if (ctx != context::none) { + if ((ctx != context::in_list_after_bracket) && + (ctx != context::in_dict_after_brace)) { + TRY(write_dst(",", 1)); + } + TRY(write_dst("\n", 1)); + for (size_t i = 0; i < depth; i++) { + TRY(write_dst(INDENT_STRING, indent)); } } @@ -339,7 +336,7 @@ handle_token(wuffs_base__token t) { // value (number, string or literal). switch (vbc) { case WUFFS_BASE__TOKEN__VBC__STRING: - if (!prev_token_incomplete) { + if (!t.link_prev()) { TRY(write_dst("\"", 1)); } @@ -352,9 +349,7 @@ handle_token(wuffs_base__token t) { return "main: internal error: unexpected string-token conversion"; } - prev_token_incomplete = - vbd & WUFFS_BASE__TOKEN__VBD__STRING__INCOMPLETE; - if (prev_token_incomplete) { + if (t.link_next()) { return nullptr; } TRY(write_dst("\"", 1)); diff --git a/internal/cgen/base/token-public.h b/internal/cgen/base/token-public.h index 389cab65..b34f073a 100644 --- a/internal/cgen/base/token-public.h +++ b/internal/cgen/base/token-public.h @@ -17,22 +17,52 @@ // ---------------- Tokens typedef struct { - // The repr is divided as: - // - Bits 63 .. 40 (24 bits) is the major value. - // - Bits 39 .. 16 (24 bits) is the minor value. + // The repr's 64 bits are divided as: + // + // +-----+-------------+-------+-------------+-----+-----+-----------+ + // | 1 | 21 | 3 | 21 | 1 | 1 | 16 | + // +-----+-------------+-------+-------------+-----+-----+-----------+ + // [..................value..................] LP LN length + // [..0..|.value_major.|.....value_minor.....] + // [..0..|.........VBC.........|.....VBD.....] + // + // The broad divisions are: + // - Bits 63 .. 18 (46 bits) is the value. + // - Bits 17 .. 16 ( 2 bits) is LP and LN (link_prev and link_next). // - Bits 15 .. 0 (16 bits) is the length. // - // The major value is a [Base38](doc/note/base38-and-fourcc.md) value. If - // zero (special cased for Wuffs' built-in "base" package) then the minor - // value is further sub-divided: - // - Bits 39 .. 37 ( 3 bits) is the value_base_category. - // - Bits 36 .. 16 (21 bits) is the value_base_detail. + // The value bits can be sub-divided in multiple ways: + // - Bits 63 .. 63 ( 1 bits) is reserved (a zero bit). + // - Bits 62 .. 42 (21 bits) is the value_major. + // - Bits 41 .. 18 (24 bits) is the value_minor. + // - Bits 62 .. 39 (24 bits) is the VBC (value_base_category). + // - Bits 38 .. 18 (21 bits) is the VBD (value_base_detail). // - // In particular, at 21 bits, the value_base_detail can hold every valid - // Unicode code point. + // The value_major is a 21-bit [Base38](doc/note/base38-and-fourcc.md) value. + // If all of its bits are zero (special cased for Wuffs' built-in "base" + // package) then the value_minor is further sub-divided: + // - Bits 41 .. 39 ( 3 bits) is the VBC (value_base_category). + // - Bits 38 .. 18 (21 bits) is the VBD (value_base_detail). // - // If the major value is non-zero then the minor value has whatever arbitrary - // meaning the tokenizer's package assigns to it. + // The high 46 bits (bits 63 .. 18) only have VBC and VBD semantics when the + // high 22 bits (the value_major) are all zero. An equivalent test is that + // the high 25 bits (the notional VBC) has a numerical value less than 8. + // + // At 21 bits, the VBD can hold every valid Unicode code point. + // + // If value_major is non-zero then value_minor has whatever arbitrary meaning + // the tokenizer's package assigns to it. + // + // Multiple consecutive tokens can form a larger conceptual unit. For + // example, an "abc\tz" string is a single higher level concept but at the + // lower level, it could consist of multiple tokens: the quotes '"', the + // ASCII texts "abc" and "z" and the backslash-escaped tab '\t'. The LP and + // LN (link_prev and link_next) bits denote tokens that are part of a + // multi-token chain: + // - LP means that this token is not the first (there is a previous token). + // - LN means that this token is not the last (there is a next token). + // + // In particular, a stand-alone token will have both link bits set to zero. uint64_t repr; #ifdef __cplusplus @@ -41,6 +71,8 @@ typedef struct { inline uint64_t value_minor() const; inline uint64_t value_base_category() const; inline uint64_t value_base_detail() const; + inline bool link_prev() const; + inline bool link_next() const; inline uint64_t length() const; #endif // __cplusplus @@ -55,20 +87,25 @@ wuffs_base__make_token(uint64_t repr) { // -------- -#define WUFFS_BASE__TOKEN__VALUE__MASK 0xFFFFFFFFFFFF -#define WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK 0xFFFFFF +#define WUFFS_BASE__TOKEN__VALUE__MASK 0x3FFFFFFFFFFF +#define WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK 0x3FFFFF #define WUFFS_BASE__TOKEN__VALUE_MINOR__MASK 0xFFFFFF -#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK 0x7FFFFFF +#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK 0x1FFFFFF #define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__MASK 0x1FFFFF +#define WUFFS_BASE__TOKEN__LINK__MASK 0x3 #define WUFFS_BASE__TOKEN__LENGTH__MASK 0xFFFF -#define WUFFS_BASE__TOKEN__VALUE__SHIFT 16 -#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 40 -#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 16 -#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 37 -#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 16 +#define WUFFS_BASE__TOKEN__VALUE__SHIFT 18 +#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 42 +#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 18 +#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 39 +#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 18 +#define WUFFS_BASE__TOKEN__LINK__SHIFT 16 #define WUFFS_BASE__TOKEN__LENGTH__SHIFT 0 +#define WUFFS_BASE__TOKEN__LINK_PREV 0x20000 +#define WUFFS_BASE__TOKEN__LINK_NEXT 0x10000 + // -------- #define WUFFS_BASE__TOKEN__VBC__FILLER 0 @@ -77,41 +114,19 @@ wuffs_base__make_token(uint64_t repr) { #define WUFFS_BASE__TOKEN__VBC__NUMBER 3 #define WUFFS_BASE__TOKEN__VBC__STRUCTURE 4 -// -------- + // -------- -// INCOMPLETE means that this token combines with the following token. -// -// For example, tokenizing a comment that is longer than the source buffer can -// result in multiple VBC__FILLER tokens. All but the last one is INCOMPLETE. -// -// By convention, whitespace is not marked incomplete. Two whitespace tokens of -// length 30 are equivalent to one whitespace token of length 60. -// -// Bits 0x2, 0x4 and 0x8 are reserved for flags that are common between -// VBD_FILLER and VBD_STRING. -#define WUFFS_BASE__TOKEN__VBD__FILLER__INCOMPLETE 0x00001 - -#define WUFFS_BASE__TOKEN__VBD__FILLER__END_OF_CONSECUTIVE_COMMENTS 0x00010 -#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_LINE 0x00020 -#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_BLOCK 0x00040 +#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_LINE 0x00001 +#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_BLOCK 0x00002 // -------- -// INCOMPLETE means that this token combines with the following token. -// -// For example, tokenizing a string that is longer than the source buffer can -// result in multiple VBC__STRING tokens. All but the last one is INCOMPLETE. -// -// Bits 0x2, 0x4 and 0x8 are reserved for flags that are common between -// VBD_FILLER and VBD_STRING. -#define WUFFS_BASE__TOKEN__VBD__STRING__INCOMPLETE 0x00001 - -// "DEFINITELY_FOO" means that the source bytes (and also the destination -// bytes, assuming 1_DST_1_SRC_COPY) are in the FOO format. Definitely means -// that the lack of the bit is conservative: it is valid for all-ASCII strings -// to have neither DEFINITELY_ASCII or DEFINITELY_UTF_8 bits set. -#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_ASCII 0x00010 -#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_UTF_8 0x00020 +// "DEFINITELY_FOO" means that the destination bytes (and also the source +// bytes, for 1_DST_1_SRC_COPY) are in the FOO format. Definitely means that +// the lack of the bit is conservative: it is valid for all-ASCII strings to +// have neither DEFINITELY_UTF_8 or DEFINITELY_ASCII bits set. +#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_UTF_8 0x00001 +#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_ASCII 0x00002 // "CONVERT_D_DST_S_SRC" means that multiples of S source bytes (possibly // padded) produces multiples of D destination bytes. For example, @@ -121,21 +136,13 @@ wuffs_base__make_token(uint64_t repr) { // When src is the empty string, multiple conversion algorithms are applicable // (so these bits are not necessarily mutually exclusive), all producing the // same empty dst string. -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP 0x00100 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY 0x00200 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_2_SRC_HEXADECIMAL 0x00400 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X 0x00800 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_STD 0x01000 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_URL 0x02000 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_4_DST_5_SRC_ASCII_85 0x04000 - - // -------- - - // UNICODE_CODE_POINT tokens have no room in their VBD for an INCOMPLETE bit. - // All 21 bits are used to hold the Unicode code point. Such tokens preserve - // the INCOMPLETEness of the previous token (if a FILLER or STRING). - -#define WUFFS_BASE__TOKEN__VBD__UNICODE_CODE_POINT__MAX_INCL 0x10FFFF +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP 0x00010 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY 0x00020 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_2_SRC_HEXADECIMAL 0x00040 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X 0x00080 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_STD 0x00100 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_URL 0x00200 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_4_DST_5_SRC_ASCII_85 0x00400 // -------- @@ -201,6 +208,16 @@ wuffs_base__token__value_base_detail(const wuffs_base__token* t) { WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__MASK; } +static inline bool // +wuffs_base__token__link_prev(const wuffs_base__token* t) { + return t->repr & WUFFS_BASE__TOKEN__LINK_PREV; +} + +static inline bool // +wuffs_base__token__link_next(const wuffs_base__token* t) { + return t->repr & WUFFS_BASE__TOKEN__LINK_NEXT; +} + static inline uint64_t // wuffs_base__token__length(const wuffs_base__token* t) { return (t->repr >> WUFFS_BASE__TOKEN__LENGTH__SHIFT) & @@ -234,6 +251,16 @@ wuffs_base__token::value_base_detail() const { return wuffs_base__token__value_base_detail(this); } +inline bool // +wuffs_base__token::link_prev() const { + return wuffs_base__token__link_prev(this); +} + +inline bool // +wuffs_base__token::link_next() const { + return wuffs_base__token__link_next(this); +} + inline uint64_t // wuffs_base__token::length() const { return wuffs_base__token__length(this); diff --git a/internal/cgen/builtin.go b/internal/cgen/builtin.go index 55562a4c..a6676fc4 100644 --- a/internal/cgen/builtin.go +++ b/internal/cgen/builtin.go @@ -325,11 +325,27 @@ func (g *gen) writeBuiltinIOWriter(b *buffer, recv *a.Expr, method t.ID, args [] func (g *gen) writeBuiltinTokenWriter(b *buffer, recv *a.Expr, method t.ID, args []*a.Node, depth uint32) error { if method == t.IDWriteFastToken { b.printf("*iop_a_dst++ = wuffs_base__make_token((((uint64_t)(") - if err := g.writeExpr(b, args[0].AsArg().Value(), depth); err != nil { + + if cv := args[0].AsArg().Value().ConstValue(); (cv == nil) || (cv.Sign() != 0) { + if err := g.writeExpr(b, args[0].AsArg().Value(), depth); err != nil { + return err + } + b.writes(")) << WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT) | (((uint64_t)(") + } + + if err := g.writeExpr(b, args[1].AsArg().Value(), depth); err != nil { return err } - b.writes(")) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | (((uint64_t)(") - if err := g.writeExpr(b, args[1].AsArg().Value(), depth); err != nil { + b.writes(")) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(") + + if cv := args[2].AsArg().Value().ConstValue(); (cv == nil) || (cv.Sign() != 0) { + if err := g.writeExpr(b, args[2].AsArg().Value(), depth); err != nil { + return err + } + b.writes(")) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(") + } + + if err := g.writeExpr(b, args[3].AsArg().Value(), depth); err != nil { return err } b.writes(")) << WUFFS_BASE__TOKEN__LENGTH__SHIFT))") @@ -699,7 +715,7 @@ func (g *gen) writeBuiltinQuestionCall(b *buffer, n *a.Expr, depth uint32) error if err := g.writeExpr(b, n.Args()[0].AsArg().Value(), depth); err != nil { return err } - b.writes(")) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | (((uint64_t)(") + b.writes(")) << WUFFS_BASE__TOKEN__VLL__SHIFT) | (((uint64_t)(") if err := g.writeExpr(b, n.Args()[1].AsArg().Value(), depth); err != nil { return err } diff --git a/internal/cgen/data.go b/internal/cgen/data.go index ebb7fa9a..fee98742 100644 --- a/internal/cgen/data.go +++ b/internal/cgen/data.go @@ -308,26 +308,26 @@ const baseTokenPrivateH = "" + "" const baseTokenPublicH = "" + - "// ---------------- Tokens\n\ntypedef struct {\n // The repr is divided as:\n // - Bits 63 .. 40 (24 bits) is the major value.\n // - Bits 39 .. 16 (24 bits) is the minor value.\n // - Bits 15 .. 0 (16 bits) is the length.\n //\n // The major value is a [Base38](doc/note/base38-and-fourcc.md) value. If\n // zero (special cased for Wuffs' built-in \"base\" package) then the minor\n // value is further sub-divided:\n // - Bits 39 .. 37 ( 3 bits) is the value_base_category.\n // - Bits 36 .. 16 (21 bits) is the value_base_detail.\n //\n // In particular, at 21 bits, the value_base_detail can hold every valid\n // Unicode code point.\n //\n // If the major value is non-zero then the minor value has whatever arbitrary\n // meaning the tokenizer's package assigns to it.\n uint64_t repr;\n\n#ifdef __cplusplus\n inline uint64_t value() const;\n inline uint64_t value_major() const;\n inline uint64_t value_minor() const;\n inline uint64_t value_base_category() const;\n inline uint64_t value_base_detail() const;\n inlin" + - "e uint64_t length() const;\n#endif // __cplusplus\n\n} wuffs_base__token;\n\nstatic inline wuffs_base__token //\nwuffs_base__make_token(uint64_t repr) {\n wuffs_base__token ret;\n ret.repr = repr;\n return ret;\n}\n\n " + + "// ---------------- Tokens\n\ntypedef struct {\n // The repr's 64 bits are divided as:\n //\n // +-----+-------------+-------+-------------+-----+-----+-----------+\n // | 1 | 21 | 3 | 21 | 1 | 1 | 16 |\n // +-----+-------------+-------+-------------+-----+-----+-----------+\n // [..................value..................] LP LN length\n // [..0..|.value_major.|.....value_minor.....]\n // [..0..|.........VBC.........|.....VBD.....]\n //\n // The broad divisions are:\n // - Bits 63 .. 18 (46 bits) is the value.\n // - Bits 17 .. 16 ( 2 bits) is LP and LN (link_prev and link_next).\n // - Bits 15 .. 0 (16 bits) is the length.\n //\n // The value bits can be sub-divided in multiple ways:\n // - Bits 63 .. 63 ( 1 bits) is reserved (a zero bit).\n // - Bits 62 .. 42 (21 bits) is the value_major.\n // - Bits 41 .. 18 (24 bits) is the value_minor.\n // - Bits 62 .. 39 (24 bits) is the VBC (value_base_category).\n // - Bits 38 .. 18 (21 bits) is the VBD (value_base_de" + + "tail).\n //\n // The value_major is a 21-bit [Base38](doc/note/base38-and-fourcc.md) value.\n // If all of its bits are zero (special cased for Wuffs' built-in \"base\"\n // package) then the value_minor is further sub-divided:\n // - Bits 41 .. 39 ( 3 bits) is the VBC (value_base_category).\n // - Bits 38 .. 18 (21 bits) is the VBD (value_base_detail).\n //\n // The high 46 bits (bits 63 .. 18) only have VBC and VBD semantics when the\n // high 22 bits (the value_major) are all zero. An equivalent test is that\n // the high 25 bits (the notional VBC) has a numerical value less than 8.\n //\n // At 21 bits, the VBD can hold every valid Unicode code point.\n //\n // If value_major is non-zero then value_minor has whatever arbitrary meaning\n // the tokenizer's package assigns to it.\n //\n // Multiple consecutive tokens can form a larger conceptual unit. For\n // example, an \"abc\\tz\" string is a single higher level concept but at the\n // lower level, it could consist of multiple tokens: the quotes '\"', the\n " + + "// ASCII texts \"abc\" and \"z\" and the backslash-escaped tab '\\t'. The LP and\n // LN (link_prev and link_next) bits denote tokens that are part of a\n // multi-token chain:\n // - LP means that this token is not the first (there is a previous token).\n // - LN means that this token is not the last (there is a next token).\n //\n // In particular, a stand-alone token will have both link bits set to zero.\n uint64_t repr;\n\n#ifdef __cplusplus\n inline uint64_t value() const;\n inline uint64_t value_major() const;\n inline uint64_t value_minor() const;\n inline uint64_t value_base_category() const;\n inline uint64_t value_base_detail() const;\n inline bool link_prev() const;\n inline bool link_next() const;\n inline uint64_t length() const;\n#endif // __cplusplus\n\n} wuffs_base__token;\n\nstatic inline wuffs_base__token //\nwuffs_base__make_token(uint64_t repr) {\n wuffs_base__token ret;\n ret.repr = repr;\n return ret;\n}\n\n " + "" + - "// --------\n\n#define WUFFS_BASE__TOKEN__VALUE__MASK 0xFFFFFFFFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK 0xFFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_MINOR__MASK 0xFFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK 0x7FFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__MASK 0x1FFFFF\n#define WUFFS_BASE__TOKEN__LENGTH__MASK 0xFFFF\n\n#define WUFFS_BASE__TOKEN__VALUE__SHIFT 16\n#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 40\n#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 16\n#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 37\n#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 16\n#define WUFFS_BASE__TOKEN__LENGTH__SHIFT 0\n\n " + + "// --------\n\n#define WUFFS_BASE__TOKEN__VALUE__MASK 0x3FFFFFFFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK 0x3FFFFF\n#define WUFFS_BASE__TOKEN__VALUE_MINOR__MASK 0xFFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK 0x1FFFFFF\n#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__MASK 0x1FFFFF\n#define WUFFS_BASE__TOKEN__LINK__MASK 0x3\n#define WUFFS_BASE__TOKEN__LENGTH__MASK 0xFFFF\n\n#define WUFFS_BASE__TOKEN__VALUE__SHIFT 18\n#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 42\n#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 18\n#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 39\n#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 18\n#define WUFFS_BASE__TOKEN__LINK__SHIFT 16\n#define WUFFS_BASE__TOKEN__LENGTH__SHIFT 0\n\n#define WUFFS_BASE__TOKEN__LINK_PREV 0x20000\n#define WUFFS_BASE__TOKEN__LINK_NEXT 0x10000\n\n " + "" + - "// --------\n\n#define WUFFS_BASE__TOKEN__VBC__FILLER 0\n#define WUFFS_BASE__TOKEN__VBC__STRING 1\n#define WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT 2\n#define WUFFS_BASE__TOKEN__VBC__NUMBER 3\n#define WUFFS_BASE__TOKEN__VBC__STRUCTURE 4\n\n" + + "// --------\n\n#define WUFFS_BASE__TOKEN__VBC__FILLER 0\n#define WUFFS_BASE__TOKEN__VBC__STRING 1\n#define WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT 2\n#define WUFFS_BASE__TOKEN__VBC__NUMBER 3\n#define WUFFS_BASE__TOKEN__VBC__STRUCTURE 4\n\n " + "" + - "// --------\n\n// INCOMPLETE means that this token combines with the following token.\n//\n// For example, tokenizing a comment that is longer than the source buffer can\n// result in multiple VBC__FILLER tokens. All but the last one is INCOMPLETE.\n//\n// By convention, whitespace is not marked incomplete. Two whitespace tokens of\n// length 30 are equivalent to one whitespace token of length 60.\n//\n// Bits 0x2, 0x4 and 0x8 are reserved for flags that are common between\n// VBD_FILLER and VBD_STRING.\n#define WUFFS_BASE__TOKEN__VBD__FILLER__INCOMPLETE 0x00001\n\n#define WUFFS_BASE__TOKEN__VBD__FILLER__END_OF_CONSECUTIVE_COMMENTS 0x00010\n#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_LINE 0x00020\n#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_BLOCK 0x00040\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" + "" + - "// --------\n\n// INCOMPLETE means that this token combines with the following token.\n//\n// For example, tokenizing a string that is longer than the source buffer can\n// result in multiple VBC__STRING tokens. All but the last one is INCOMPLETE.\n//\n// Bits 0x2, 0x4 and 0x8 are reserved for flags that are common between\n// VBD_FILLER and VBD_STRING.\n#define WUFFS_BASE__TOKEN__VBD__STRING__INCOMPLETE 0x00001\n\n// \"DEFINITELY_FOO\" means that the source bytes (and also the destination\n// bytes, assuming 1_DST_1_SRC_COPY) are in the FOO format. Definitely means\n// that the lack of the bit is conservative: it is valid for all-ASCII strings\n// to have neither DEFINITELY_ASCII or DEFINITELY_UTF_8 bits set.\n#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_ASCII 0x00010\n#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_UTF_8 0x00020\n\n// \"CONVERT_D_DST_S_SRC\" means that multiples of S source bytes (possibly\n// padded) produces multiples of D destination bytes. For example,\n// CONVERT_1_DST_4_SRC_BACKSLASH_X means a sourc" + - "e like \"\\\\x23\\\\x67\\\\xAB\", where\n// 12 src bytes encode 3 dst bytes.\n//\n// When src is the empty string, multiple conversion algorithms are applicable\n// (so these bits are not necessarily mutually exclusive), all producing the\n// same empty dst string.\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP 0x00100\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY 0x00200\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_2_SRC_HEXADECIMAL 0x00400\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X 0x00800\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_STD 0x01000\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_URL 0x02000\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_4_DST_5_SRC_ASCII_85 0x04000\n\n " + - "" + - "// --------\n\n // UNICODE_CODE_POINT tokens have no room in their VBD for an INCOMPLETE bit.\n // All 21 bits are used to hold the Unicode code point. Such tokens preserve\n // the INCOMPLETEness of the previous token (if a FILLER or STRING).\n\n#define WUFFS_BASE__TOKEN__VBD__UNICODE_CODE_POINT__MAX_INCL 0x10FFFF\n\n " + + "// --------\n\n// \"DEFINITELY_FOO\" means that the destination bytes (and also the source\n// bytes, for 1_DST_1_SRC_COPY) are in the FOO format. Definitely means that\n// the lack of the bit is conservative: it is valid for all-ASCII strings to\n// have neither DEFINITELY_UTF_8 or DEFINITELY_ASCII bits set.\n#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_UTF_8 0x00001\n#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_ASCII 0x00002\n\n// \"CONVERT_D_DST_S_SRC\" means that multiples of S source bytes (possibly\n// padded) produces multiples of D destination bytes. For example,\n// CONVERT_1_DST_4_SRC_BACKSLASH_X means a source like \"\\\\x23\\\\x67\\\\xAB\", where\n// 12 src bytes encode 3 dst bytes.\n//\n// When src is the empty string, multiple conversion algorithms are applicable\n// (so these bits are not necessarily mutually exclusive), all producing the\n// same empty dst string.\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP 0x00010\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY 0x00020\n#" + + "define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_2_SRC_HEXADECIMAL 0x00040\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X 0x00080\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_STD 0x00100\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_URL 0x00200\n#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_4_DST_5_SRC_ASCII_85 0x00400\n\n " + "" + "// --------\n\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__LITERAL 0x00001\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__LITERAL__UNDEFINED 0x00101\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__LITERAL__NULL 0x00201\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__LITERAL__FALSE 0x00401\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__LITERAL__TRUE 0x00801\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__FLOATING_POINT.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__INTEGER_SIGNED.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__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__FLOATING_POINT 0x00002\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__INTEGER_SIGNED 0x00004\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__INTEGER_UNSIGNED 0x00008\n\n " + "" + "// --------\n\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__PUSH 0x00001\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__POP 0x00002\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__FROM_NONE 0x00010\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__FROM_LIST 0x00020\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__FROM_DICT 0x00040\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_NONE 0x01000\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_LIST 0x02000\n#define WUFFS_BASE__TOKEN__VBD__STRUCTURE__TO_DICT 0x04000\n\n" + "" + "// --------\n\nstatic inline uint64_t //\nwuffs_base__token__value(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__VALUE__SHIFT) &\n WUFFS_BASE__TOKEN__VALUE__MASK;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__value_major(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT) &\n WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK;\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) &\n WUFFS_BASE__TOKEN__VALUE_MINOR__MASK;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__value_base_category(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT) &\n WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK;\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) &\n WUFFS_BASE__TOKEN__VALUE_BASE_DETA" + - "IL__MASK;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__length(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__LENGTH__SHIFT) &\n WUFFS_BASE__TOKEN__LENGTH__MASK;\n}\n\n#ifdef __cplusplus\n\ninline uint64_t //\nwuffs_base__token::value() const {\n return wuffs_base__token__value(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_major() const {\n return wuffs_base__token__value_major(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_minor() const {\n return wuffs_base__token__value_minor(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_base_category() const {\n return wuffs_base__token__value_base_category(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_base_detail() const {\n return wuffs_base__token__value_base_detail(this);\n}\n\ninline uint64_t //\nwuffs_base__token::length() const {\n return wuffs_base__token__length(this);\n}\n\n#endif // __cplusplus\n\n" + + "IL__MASK;\n}\n\nstatic inline bool //\nwuffs_base__token__link_prev(const wuffs_base__token* t) {\n return t->repr & WUFFS_BASE__TOKEN__LINK_PREV;\n}\n\nstatic inline bool //\nwuffs_base__token__link_next(const wuffs_base__token* t) {\n return t->repr & WUFFS_BASE__TOKEN__LINK_NEXT;\n}\n\nstatic inline uint64_t //\nwuffs_base__token__length(const wuffs_base__token* t) {\n return (t->repr >> WUFFS_BASE__TOKEN__LENGTH__SHIFT) &\n WUFFS_BASE__TOKEN__LENGTH__MASK;\n}\n\n#ifdef __cplusplus\n\ninline uint64_t //\nwuffs_base__token::value() const {\n return wuffs_base__token__value(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_major() const {\n return wuffs_base__token__value_major(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_minor() const {\n return wuffs_base__token__value_minor(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_base_category() const {\n return wuffs_base__token__value_base_category(this);\n}\n\ninline uint64_t //\nwuffs_base__token::value_base_detail() const {\n return wuffs_base__to" + + "ken__value_base_detail(this);\n}\n\ninline bool //\nwuffs_base__token::link_prev() const {\n return wuffs_base__token__link_prev(this);\n}\n\ninline bool //\nwuffs_base__token::link_next() const {\n return wuffs_base__token__link_next(this);\n}\n\ninline uint64_t //\nwuffs_base__token::length() const {\n return wuffs_base__token__length(this);\n}\n\n#endif // __cplusplus\n\n" + "" + "// --------\n\ntypedef WUFFS_BASE__SLICE(wuffs_base__token) wuffs_base__slice_token;\n\nstatic inline wuffs_base__slice_token //\nwuffs_base__make_slice_token(wuffs_base__token* ptr, size_t len) {\n wuffs_base__slice_token ret;\n ret.ptr = ptr;\n ret.len = len;\n return ret;\n}\n\n" + "" + diff --git a/lang/builtin/builtin.go b/lang/builtin/builtin.go index 22e2dd68..2ec22216 100644 --- a/lang/builtin/builtin.go +++ b/lang/builtin/builtin.go @@ -326,9 +326,13 @@ var Funcs = []string{ // ---- token_writer - "token_writer.write_token?(value: u64[..= 0xFFFFFFFFFFFF], length: u64[..= 0xFFFF])", + "token_writer.write_token?(" + + "value_major: u32[..= 0x1F_FFFF], value_minor: u32[..= 0xFF_FFFF]," + + "link: u32[..= 0x3], length: u32[..= 0xFFFF])", - "token_writer.write_fast_token!(value: u64[..= 0xFFFFFFFFFFFF], length: u64[..= 0xFFFF])", + "token_writer.write_fast_token!(" + + "value_major: u32[..= 0x1F_FFFF], value_minor: u32[..= 0xFF_FFFF]," + + "link: u32[..= 0x3], length: u32[..= 0xFFFF])", "token_writer.available() u64", diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index 6b3a64a4..75ca1012 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -1733,22 +1733,52 @@ wuffs_base__io_buffer::writer_io_position() const { // ---------------- Tokens typedef struct { - // The repr is divided as: - // - Bits 63 .. 40 (24 bits) is the major value. - // - Bits 39 .. 16 (24 bits) is the minor value. + // The repr's 64 bits are divided as: + // + // +-----+-------------+-------+-------------+-----+-----+-----------+ + // | 1 | 21 | 3 | 21 | 1 | 1 | 16 | + // +-----+-------------+-------+-------------+-----+-----+-----------+ + // [..................value..................] LP LN length + // [..0..|.value_major.|.....value_minor.....] + // [..0..|.........VBC.........|.....VBD.....] + // + // The broad divisions are: + // - Bits 63 .. 18 (46 bits) is the value. + // - Bits 17 .. 16 ( 2 bits) is LP and LN (link_prev and link_next). // - Bits 15 .. 0 (16 bits) is the length. // - // The major value is a [Base38](doc/note/base38-and-fourcc.md) value. If - // zero (special cased for Wuffs' built-in "base" package) then the minor - // value is further sub-divided: - // - Bits 39 .. 37 ( 3 bits) is the value_base_category. - // - Bits 36 .. 16 (21 bits) is the value_base_detail. + // The value bits can be sub-divided in multiple ways: + // - Bits 63 .. 63 ( 1 bits) is reserved (a zero bit). + // - Bits 62 .. 42 (21 bits) is the value_major. + // - Bits 41 .. 18 (24 bits) is the value_minor. + // - Bits 62 .. 39 (24 bits) is the VBC (value_base_category). + // - Bits 38 .. 18 (21 bits) is the VBD (value_base_detail). // - // In particular, at 21 bits, the value_base_detail can hold every valid - // Unicode code point. + // The value_major is a 21-bit [Base38](doc/note/base38-and-fourcc.md) value. + // If all of its bits are zero (special cased for Wuffs' built-in "base" + // package) then the value_minor is further sub-divided: + // - Bits 41 .. 39 ( 3 bits) is the VBC (value_base_category). + // - Bits 38 .. 18 (21 bits) is the VBD (value_base_detail). // - // If the major value is non-zero then the minor value has whatever arbitrary - // meaning the tokenizer's package assigns to it. + // The high 46 bits (bits 63 .. 18) only have VBC and VBD semantics when the + // high 22 bits (the value_major) are all zero. An equivalent test is that + // the high 25 bits (the notional VBC) has a numerical value less than 8. + // + // At 21 bits, the VBD can hold every valid Unicode code point. + // + // If value_major is non-zero then value_minor has whatever arbitrary meaning + // the tokenizer's package assigns to it. + // + // Multiple consecutive tokens can form a larger conceptual unit. For + // example, an "abc\tz" string is a single higher level concept but at the + // lower level, it could consist of multiple tokens: the quotes '"', the + // ASCII texts "abc" and "z" and the backslash-escaped tab '\t'. The LP and + // LN (link_prev and link_next) bits denote tokens that are part of a + // multi-token chain: + // - LP means that this token is not the first (there is a previous token). + // - LN means that this token is not the last (there is a next token). + // + // In particular, a stand-alone token will have both link bits set to zero. uint64_t repr; #ifdef __cplusplus @@ -1757,6 +1787,8 @@ typedef struct { inline uint64_t value_minor() const; inline uint64_t value_base_category() const; inline uint64_t value_base_detail() const; + inline bool link_prev() const; + inline bool link_next() const; inline uint64_t length() const; #endif // __cplusplus @@ -1771,20 +1803,25 @@ wuffs_base__make_token(uint64_t repr) { // -------- -#define WUFFS_BASE__TOKEN__VALUE__MASK 0xFFFFFFFFFFFF -#define WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK 0xFFFFFF +#define WUFFS_BASE__TOKEN__VALUE__MASK 0x3FFFFFFFFFFF +#define WUFFS_BASE__TOKEN__VALUE_MAJOR__MASK 0x3FFFFF #define WUFFS_BASE__TOKEN__VALUE_MINOR__MASK 0xFFFFFF -#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK 0x7FFFFFF +#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__MASK 0x1FFFFFF #define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__MASK 0x1FFFFF +#define WUFFS_BASE__TOKEN__LINK__MASK 0x3 #define WUFFS_BASE__TOKEN__LENGTH__MASK 0xFFFF -#define WUFFS_BASE__TOKEN__VALUE__SHIFT 16 -#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 40 -#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 16 -#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 37 -#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 16 +#define WUFFS_BASE__TOKEN__VALUE__SHIFT 18 +#define WUFFS_BASE__TOKEN__VALUE_MAJOR__SHIFT 42 +#define WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT 18 +#define WUFFS_BASE__TOKEN__VALUE_BASE_CATEGORY__SHIFT 39 +#define WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__SHIFT 18 +#define WUFFS_BASE__TOKEN__LINK__SHIFT 16 #define WUFFS_BASE__TOKEN__LENGTH__SHIFT 0 +#define WUFFS_BASE__TOKEN__LINK_PREV 0x20000 +#define WUFFS_BASE__TOKEN__LINK_NEXT 0x10000 + // -------- #define WUFFS_BASE__TOKEN__VBC__FILLER 0 @@ -1793,41 +1830,19 @@ wuffs_base__make_token(uint64_t repr) { #define WUFFS_BASE__TOKEN__VBC__NUMBER 3 #define WUFFS_BASE__TOKEN__VBC__STRUCTURE 4 -// -------- + // -------- -// INCOMPLETE means that this token combines with the following token. -// -// For example, tokenizing a comment that is longer than the source buffer can -// result in multiple VBC__FILLER tokens. All but the last one is INCOMPLETE. -// -// By convention, whitespace is not marked incomplete. Two whitespace tokens of -// length 30 are equivalent to one whitespace token of length 60. -// -// Bits 0x2, 0x4 and 0x8 are reserved for flags that are common between -// VBD_FILLER and VBD_STRING. -#define WUFFS_BASE__TOKEN__VBD__FILLER__INCOMPLETE 0x00001 - -#define WUFFS_BASE__TOKEN__VBD__FILLER__END_OF_CONSECUTIVE_COMMENTS 0x00010 -#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_LINE 0x00020 -#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_BLOCK 0x00040 +#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_LINE 0x00001 +#define WUFFS_BASE__TOKEN__VBD__FILLER__COMMENT_BLOCK 0x00002 // -------- -// INCOMPLETE means that this token combines with the following token. -// -// For example, tokenizing a string that is longer than the source buffer can -// result in multiple VBC__STRING tokens. All but the last one is INCOMPLETE. -// -// Bits 0x2, 0x4 and 0x8 are reserved for flags that are common between -// VBD_FILLER and VBD_STRING. -#define WUFFS_BASE__TOKEN__VBD__STRING__INCOMPLETE 0x00001 - -// "DEFINITELY_FOO" means that the source bytes (and also the destination -// bytes, assuming 1_DST_1_SRC_COPY) are in the FOO format. Definitely means -// that the lack of the bit is conservative: it is valid for all-ASCII strings -// to have neither DEFINITELY_ASCII or DEFINITELY_UTF_8 bits set. -#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_ASCII 0x00010 -#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_UTF_8 0x00020 +// "DEFINITELY_FOO" means that the destination bytes (and also the source +// bytes, for 1_DST_1_SRC_COPY) are in the FOO format. Definitely means that +// the lack of the bit is conservative: it is valid for all-ASCII strings to +// have neither DEFINITELY_UTF_8 or DEFINITELY_ASCII bits set. +#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_UTF_8 0x00001 +#define WUFFS_BASE__TOKEN__VBD__STRING__DEFINITELY_ASCII 0x00002 // "CONVERT_D_DST_S_SRC" means that multiples of S source bytes (possibly // padded) produces multiples of D destination bytes. For example, @@ -1837,21 +1852,13 @@ wuffs_base__make_token(uint64_t repr) { // When src is the empty string, multiple conversion algorithms are applicable // (so these bits are not necessarily mutually exclusive), all producing the // same empty dst string. -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP 0x00100 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY 0x00200 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_2_SRC_HEXADECIMAL 0x00400 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X 0x00800 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_STD 0x01000 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_URL 0x02000 -#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_4_DST_5_SRC_ASCII_85 0x04000 - - // -------- - - // UNICODE_CODE_POINT tokens have no room in their VBD for an INCOMPLETE bit. - // All 21 bits are used to hold the Unicode code point. Such tokens preserve - // the INCOMPLETEness of the previous token (if a FILLER or STRING). - -#define WUFFS_BASE__TOKEN__VBD__UNICODE_CODE_POINT__MAX_INCL 0x10FFFF +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_0_DST_1_SRC_DROP 0x00010 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_1_SRC_COPY 0x00020 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_2_SRC_HEXADECIMAL 0x00040 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_1_DST_4_SRC_BACKSLASH_X 0x00080 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_STD 0x00100 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_3_DST_4_SRC_BASE_64_URL 0x00200 +#define WUFFS_BASE__TOKEN__VBD__STRING__CONVERT_4_DST_5_SRC_ASCII_85 0x00400 // -------- @@ -1917,6 +1924,16 @@ wuffs_base__token__value_base_detail(const wuffs_base__token* t) { WUFFS_BASE__TOKEN__VALUE_BASE_DETAIL__MASK; } +static inline bool // +wuffs_base__token__link_prev(const wuffs_base__token* t) { + return t->repr & WUFFS_BASE__TOKEN__LINK_PREV; +} + +static inline bool // +wuffs_base__token__link_next(const wuffs_base__token* t) { + return t->repr & WUFFS_BASE__TOKEN__LINK_NEXT; +} + static inline uint64_t // wuffs_base__token__length(const wuffs_base__token* t) { return (t->repr >> WUFFS_BASE__TOKEN__LENGTH__SHIFT) & @@ -1950,6 +1967,16 @@ wuffs_base__token::value_base_detail() const { return wuffs_base__token__value_base_detail(this); } +inline bool // +wuffs_base__token::link_prev() const { + return wuffs_base__token__link_prev(this); +} + +inline bool // +wuffs_base__token::link_next() const { + return wuffs_base__token__link_next(this); +} + inline uint64_t // wuffs_base__token::length() const { return wuffs_base__token__length(this); @@ -19257,7 +19284,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, self->private_impl.active_coroutine = 0; wuffs_base__status status = wuffs_base__make_status(NULL); - uint64_t v_token_value = 0; + uint32_t v_vminor = 0; uint32_t v_number_length = 0; uint32_t v_number_status = 0; uint32_t v_string_length = 0; @@ -19329,8 +19356,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, if (((uint64_t)(io2_a_src - iop_a_src)) <= 0) { if (v_whitespace_length > 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_whitespace_length)))) + (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(v_whitespace_length)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_whitespace_length = 0; } @@ -19352,7 +19379,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 1, wuffs_base__make_empty_struct()); if (v_whitespace_length >= 65534) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(65535)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_whitespace_length = 0; goto label__outer__continue; @@ -19362,8 +19389,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, label__ws__break:; if (v_whitespace_length > 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_whitespace_length)))) + (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(v_whitespace_length)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_whitespace_length = 0; if (((uint64_t)(io2_a_dst - iop_a_dst)) <= 0) { @@ -19376,53 +19403,28 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } if (v_class == 1) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(2097171)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); (iop_a_src += 1, wuffs_base__make_empty_struct()); - label__0__continue:; - while (true) { - if (((uint64_t)(io2_a_src - iop_a_src)) <= 0) { - if (a_src && a_src->meta.closed) { - status = wuffs_base__make_status(wuffs_json__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__0__continue; - } - if (wuffs_base__load_u8be__no_bounds_check(iop_a_src) != 92) { - goto label__0__break; - } - 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(4); - goto label__0__continue; - } - *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2129713)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); - goto label__0__break; - } - label__0__break:; label__string_loop__continue:; 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(5); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(3); goto label__string_loop__continue; } v_string_length = 0; - label__1__continue:; + label__0__continue:; while (true) { if (((uint64_t)(io2_a_src - iop_a_src)) <= 0) { if (v_string_length > 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_string_length)))) + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)(v_string_length)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; } @@ -19433,7 +19435,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status(wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(6); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(4); v_string_length = 0; goto label__string_loop__continue; } @@ -19443,28 +19445,34 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 1, wuffs_base__make_empty_struct()); if (v_string_length >= 65531) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(65532)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; goto label__string_loop__continue; } v_string_length += 1; - goto label__1__continue; + goto label__0__continue; } else if (v_char == 1) { - *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097696)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_string_length)))) - << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); - v_string_length = 0; + if (v_string_length != 0) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)(v_string_length)) + << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + v_string_length = 0; + } goto label__string_loop__break; } else if (v_char == 2) { if (v_string_length > 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_string_length)))) + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)(v_string_length)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; if (((uint64_t)(io2_a_dst - iop_a_dst)) <= 0) { @@ -19479,7 +19487,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status( wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(7); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(5); v_string_length = 0; v_char = 0; goto label__string_loop__continue; @@ -19491,8 +19499,9 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 2, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( (((uint64_t)( - (4194304 | ((uint64_t)((v_backslash & 127)))))) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (4194304 | ((uint32_t)((v_backslash & 127)))))) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(2)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); goto label__string_loop__continue; } else if (v_c == 117) { @@ -19504,7 +19513,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status( wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(8); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(6); v_string_length = 0; v_char = 0; goto label__string_loop__continue; @@ -19535,8 +19544,9 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else if ((v_uni4_value < 55296) || (57343 < v_uni4_value)) { (iop_a_src += 6, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)((4194304 | ((uint64_t)(v_uni4_value))))) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)((4194304 | v_uni4_value))) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(6)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); goto label__string_loop__continue; } else if (v_uni4_value >= 56320) { @@ -19549,7 +19559,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status( wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(9); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(7); v_string_length = 0; v_uni4_value = 0; v_char = 0; @@ -19592,10 +19602,10 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, v_uni4_value -= 56320; (iop_a_src += 8, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)( - (4194304 | ((uint64_t)((v_uni4_high_surrogate + - v_uni4_value)))))) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)((4194304 | v_uni4_high_surrogate | + v_uni4_value))) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(12)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); goto label__string_loop__continue; @@ -19626,7 +19636,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status( wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(10); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(8); v_string_length = 0; v_char = 0; goto label__string_loop__continue; @@ -19639,15 +19649,16 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 2, wuffs_base__make_empty_struct()); if (v_string_length >= 65528) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)((v_string_length + 2))))) + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)((v_string_length + 2))) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; goto label__string_loop__continue; } v_string_length += 2; - goto label__1__continue; + goto label__0__continue; } } else if (v_char == 4) { if (((uint64_t)(io2_a_src - iop_a_src)) < 3) { @@ -19658,7 +19669,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status( wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(11); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(9); v_string_length = 0; v_char = 0; goto label__string_loop__continue; @@ -19675,15 +19686,16 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 3, wuffs_base__make_empty_struct()); if (v_string_length >= 65528) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)((v_string_length + 3))))) + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)((v_string_length + 3))) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; goto label__string_loop__continue; } v_string_length += 3; - goto label__1__continue; + goto label__0__continue; } } } else if (v_char == 5) { @@ -19695,7 +19707,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status( wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(12); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(10); v_string_length = 0; v_char = 0; goto label__string_loop__continue; @@ -19712,22 +19724,25 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 4, wuffs_base__make_empty_struct()); if (v_string_length >= 65528) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)((v_string_length + 4))))) + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)((v_string_length + 4))) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; goto label__string_loop__continue; } v_string_length += 4; - goto label__1__continue; + goto label__0__continue; } } } if (v_string_length > 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(2097697)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_string_length)))) + (((uint64_t)(2097185)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | + (((uint64_t)(v_string_length)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_string_length = 0; } @@ -19741,7 +19756,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } } label__string_loop__break:; - label__2__continue:; + label__1__continue:; while (true) { if (((uint64_t)(io2_a_src - iop_a_src)) <= 0) { if (a_src && a_src->meta.closed) { @@ -19750,22 +19765,24 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } status = wuffs_base__make_status(wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(13); - goto label__2__continue; + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(11); + goto label__1__continue; } 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(14); - goto label__2__continue; + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(12); + goto label__1__continue; } (iop_a_src += 1, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(2097171)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(2)) << WUFFS_BASE__TOKEN__LINK__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); - goto label__2__break; + goto label__1__break; } - label__2__break:; + label__1__break:; if (0 == (v_expect & 16)) { v_expect = 8; goto label__outer__continue; @@ -19774,7 +19791,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else if (v_class == 2) { (iop_a_src += 1, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); if (0 == (v_expect & 256)) { v_expect = 2; @@ -19785,7 +19802,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else if (v_class == 3) { (iop_a_src += 1, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(0)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 3762; goto label__outer__continue; @@ -19799,18 +19816,18 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, iop_a_src = a_src->data.ptr + a_src->meta.ri; } v_number_status = (v_number_length >> 8); - v_token_value = 6291460; + v_vminor = 6291460; if ((v_number_length & 128) != 0) { - v_token_value = 6291458; + v_vminor = 6291458; } v_number_length = (v_number_length & 127); if (v_number_status == 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(v_token_value)) - << WUFFS_BASE__TOKEN__VALUE__SHIFT) | - (((uint64_t)(((uint64_t)(v_number_length)))) + (((uint64_t)(v_vminor)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(v_number_length)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); - goto label__3__break; + goto label__2__break; } while (v_number_length > 0) { v_number_length -= 1; @@ -19832,23 +19849,23 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else { status = wuffs_base__make_status(wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(15); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(13); while (((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(16); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(14); } } } - label__3__break:; + label__2__break:; goto label__goto_parsed_a_leaf_value__break; } else if (v_class == 5) { - v_token_value = 8405009; + v_vminor = 8405009; if (v_depth == 0) { } else if (0 != (v_expect_after_value & 64)) { - v_token_value = 8405057; + v_vminor = 8405057; } else { - v_token_value = 8405025; + v_vminor = 8405025; } if (v_depth >= 1024) { status = wuffs_base__make_status( @@ -19862,7 +19879,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, v_depth += 1; (iop_a_src += 1, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(v_token_value)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(v_vminor)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 66; v_expect_after_value = 68; @@ -19871,7 +19889,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 1, wuffs_base__make_empty_struct()); if (v_depth <= 1) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(8392770)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(8392770)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); goto label__outer__break; } @@ -19881,25 +19900,27 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, if (0 == (self->private_data.f_stack[v_stack_byte] & (((uint32_t)(1)) << v_stack_bit))) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(8396866)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(8396866)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 260; v_expect_after_value = 260; } else { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(8405058)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(8405058)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 68; v_expect_after_value = 68; } goto label__outer__continue; } else if (v_class == 7) { - v_token_value = 8396817; + v_vminor = 8396817; if (v_depth == 0) { } else if (0 != (v_expect_after_value & 64)) { - v_token_value = 8396865; + v_vminor = 8396865; } else { - v_token_value = 8396833; + v_vminor = 8396833; } if (v_depth >= 1024) { status = wuffs_base__make_status( @@ -19913,7 +19934,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, v_depth += 1; (iop_a_src += 1, wuffs_base__make_empty_struct()); *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(v_token_value)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(v_vminor)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 4018; v_expect_after_value = 260; @@ -19922,7 +19944,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, (iop_a_src += 1, wuffs_base__make_empty_struct()); if (v_depth <= 1) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(8392738)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(8392738)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); goto label__outer__break; } @@ -19932,13 +19955,15 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, if (0 == (self->private_data.f_stack[v_stack_byte] & (((uint32_t)(1)) << v_stack_bit))) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(8396834)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(8396834)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 260; v_expect_after_value = 260; } else { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(8405026)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(8405026)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(1)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); v_expect = 68; v_expect_after_value = 68; @@ -19949,7 +19974,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, 111546413966853); if (v_match == 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(6292481)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(6292481)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(5)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); if (((uint64_t)(io2_a_src - iop_a_src)) < 5) { status = wuffs_base__make_status( @@ -19961,7 +19987,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else if (v_match == 1) { status = wuffs_base__make_status(wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(17); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(15); goto label__outer__continue; } } else if (v_class == 10) { @@ -19969,7 +19995,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, 435762131972); if (v_match == 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(6293505)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(6293505)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(4)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); if (((uint64_t)(io2_a_src - iop_a_src)) < 4) { status = wuffs_base__make_status( @@ -19981,7 +20008,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else if (v_match == 1) { status = wuffs_base__make_status(wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(18); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(16); goto label__outer__continue; } } else if (v_class == 11) { @@ -19989,7 +20016,8 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, 465676103172); if (v_match == 0) { *iop_a_dst++ = wuffs_base__make_token( - (((uint64_t)(6291969)) << WUFFS_BASE__TOKEN__VALUE__SHIFT) | + (((uint64_t)(6291969)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | (((uint64_t)(4)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); if (((uint64_t)(io2_a_src - iop_a_src)) < 4) { status = wuffs_base__make_status( @@ -20001,7 +20029,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } else if (v_match == 1) { status = wuffs_base__make_status(wuffs_base__suspension__short_read); - WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(19); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(17); goto label__outer__continue; } } diff --git a/script/print-json-token-debug-format.c b/script/print-json-token-debug-format.c index af38a958..5ddc73fa 100644 --- a/script/print-json-token-debug-format.c +++ b/script/print-json-token-debug-format.c @@ -17,23 +17,38 @@ // print-json-token-debug-format.c parses JSON from stdin and prints the // resulting token stream, eliding any non-essential (e.g. whitespace) tokens. // -// The output format is only for debugging, and certainly not for long term -// storage. It isn't guaranteed to be stable between versions of this program -// and of the Wuffs standard library. +// The output format is only for debugging or regression testing, and certainly +// not for long term storage. It isn't guaranteed to be stable between versions +// of this program and of the Wuffs standard library. +// +// It prints 16 bytes (128 bits) per token, containing big-endian numbers: +// +// +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ +// | POS | LEN | LP| LN| MAJOR | MINOR | +// | | | | | |VBC| VBD | +// +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ +// +// - POS (4 bytes) is the position: the sum of all previous tokens' lengths, +// including elided tokens. +// - LEN (2 bytes) is the length. +// - LP (1 bytes) is the link_prev bit. +// - LN (1 bytes) is the link_next bit +// - MAJOR (4 bytes) is the value_major. +// +// The final 4 bytes are either value_minor (when the value_major is non-zero) +// or 1 + 3 bytes for value_base_category and value_base_detail (otherwise). // -// It prints 16 bytes (4 big-endian uint32's) per token: the token's position -// (total length of all previous tokens), length, major value and minor value. // Together with the hexadecimal WUFFS_BASE__TOKEN__ETC constants defined in // token-public.h, this format is somewhat human-readable when piped through a -// hex-dump program (e.g. /usr/bin/hd). +// hex-dump program (such as /usr/bin/hd), printing one token per line. // // If the input or output is larger than the program's buffers (64 MiB and // 131072 tokens by default), there may be multiple valid tokenizations of any // given input. For example, if a source string "abcde" straddles an I/O -// boundary, it may be tokenized as a 5-length complete string or as a 3-length -// incomplete string followed by a 2-length complete string. +// boundary, it may be tokenized as single (no-link) 5-length string or as a +// 3-length link_next'ed string followed by a 2-length link_prev'ed string. // -// A Wuffs token stream, in general, can support inputs more than (1 << 32) +// A Wuffs token stream, in general, can support inputs more than 0xFFFF_FFFF // bytes long, but this program can not, as it tracks the tokens' cumulative // position as a uint32. @@ -134,17 +149,28 @@ main1(int argc, char** argv) { while (tok.meta.ri < tok.meta.wi) { wuffs_base__token* t = &tok.data.ptr[tok.meta.ri++]; - uint64_t len = wuffs_base__token__length(t); + uint16_t len = wuffs_base__token__length(t); if (wuffs_base__token__value(t) != 0) { - uint64_t maj = wuffs_base__token__value_major(t); - uint64_t min = wuffs_base__token__value_minor(t); + uint8_t lp = wuffs_base__token__link_prev(t) ? 1 : 0; + uint8_t ln = wuffs_base__token__link_next(t) ? 1 : 0; + uint32_t vmajor = wuffs_base__token__value_major(t); uint8_t buf[16]; - wuffs_base__store_u32be__no_bounds_check(&buf[0 * 4], (uint32_t)(pos)); - wuffs_base__store_u32be__no_bounds_check(&buf[1 * 4], (uint32_t)(len)); - wuffs_base__store_u32be__no_bounds_check(&buf[2 * 4], (uint32_t)(maj)); - wuffs_base__store_u32be__no_bounds_check(&buf[3 * 4], (uint32_t)(min)); + wuffs_base__store_u32be__no_bounds_check(&buf[0x0], (uint32_t)(pos)); + wuffs_base__store_u16be__no_bounds_check(&buf[0x4], len); + wuffs_base__store_u8__no_bounds_check(&buf[0x0006], lp); + wuffs_base__store_u8__no_bounds_check(&buf[0x0007], ln); + wuffs_base__store_u32be__no_bounds_check(&buf[0x8], vmajor); + if (vmajor) { + uint32_t vminor = wuffs_base__token__value_minor(t); + wuffs_base__store_u32be__no_bounds_check(&buf[0xC], vminor); + } else { + uint8_t vbc = wuffs_base__token__value_base_category(t); + uint32_t vbd = wuffs_base__token__value_base_detail(t); + wuffs_base__store_u8__no_bounds_check(&buf[0x000C], vbc); + wuffs_base__store_u24be__no_bounds_check(&buf[0xD], vbd); + } const int stdout_fd = 1; write(stdout_fd, &buf[0], 16); } diff --git a/std/json/decode_json.wuffs b/std/json/decode_json.wuffs index 959589e8..beb23a42 100644 --- a/std/json/decode_json.wuffs +++ b/std/json/decode_json.wuffs @@ -45,7 +45,7 @@ pub struct decoder? implements base.token_decoder( ) pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { - var token_value : base.u64[..= 0xFF_FFFF] + var vminor : base.u32[..= 0xFF_FFFF] var number_length : base.u32[..= 0x3FF] var number_status : base.u32[..= 0x3] var string_length : base.u32[..= 0xFFFB] @@ -95,8 +95,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { if args.src.available() <= 0 { if whitespace_length > 0 { args.dst.write_fast_token!( - value: 0, - length: whitespace_length as base.u64) + value_major: 0, + value_minor: 0, + link: 0x0, + length: whitespace_length) whitespace_length = 0 } if args.src.is_closed() { @@ -116,7 +118,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { if whitespace_length >= 0xFFFE { args.dst.write_fast_token!( - value: 0, + value_major: 0, + value_minor: 0, + link: 0x0, length: 0xFFFF) whitespace_length = 0 continue.outer @@ -127,8 +131,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { // Emit whitespace. if whitespace_length > 0 { args.dst.write_fast_token!( - value: 0, - length: whitespace_length as base.u64) + value_major: 0, + value_minor: 0, + link: 0x0, + length: whitespace_length) whitespace_length = 0 if args.dst.available() <= 0 { continue.outer @@ -149,37 +155,14 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { if class == 0x01 { // 0x01 is CLASS_STRING. // -------- BEGIN parse strings. - // The leading '"' is filler. + // Emit the leading '"'. args.dst.write_fast_token!( - value: 0, + value_major: 0, + value_minor: 0x20_0013, + link: 0x1, length: 1) args.src.skip32_fast!(actual: 1, worst_case: 1) - // If the string contents starts with an escape code, issue an - // incomplete zero-length string token (with - // WUFFS_BASE__TOKEN__VBC__STRING) before issuing a token with - // WUFFS_BASE__TOKEN__VBC__UNICODE_CODE_POINT. - while true { - if args.src.available() <= 0 { - if args.src.is_closed() { - return "#bad input" - } - yield? base."$short read" - continue - } - if args.src.peek_u8() <> 0x5C { // 0x5C is '\\'. - break - } - if args.dst.available() <= 0 { - yield? base."$short write" - continue - } - args.dst.write_fast_token!( - value: 0x20_7F31, - length: 0) - break - } - while.string_loop true { if args.dst.available() <= 0 { yield? base."$short write" @@ -193,8 +176,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { if args.src.available() <= 0 { if string_length > 0 { args.dst.write_fast_token!( - value: 0x20_0221, - length: string_length as base.u64) + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length) string_length = 0 } if args.src.is_closed() { @@ -211,7 +196,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) if string_length >= 0xFFFB { args.dst.write_fast_token!( - value: 0x20_0221, + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, length: 0xFFFC) string_length = 0 continue.string_loop @@ -220,17 +207,23 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { continue } else if char == 0x01 { // '"' - args.dst.write_fast_token!( - value: 0x20_0220, - length: string_length as base.u64) - string_length = 0 + if string_length <> 0 { + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length) + string_length = 0 + } break.string_loop } else if char == 0x02 { // '\\'. if string_length > 0 { args.dst.write_fast_token!( - value: 0x20_0221, - length: string_length as base.u64) + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length) string_length = 0 if args.dst.available() <= 0 { continue.string_loop @@ -252,7 +245,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { if backslash > 0 { args.src.skip32_fast!(actual: 2, worst_case: 2) args.dst.write_fast_token!( - value: 0x40_0000 | ((backslash & 0x7F) as base.u64), + value_major: 0, + value_minor: 0x40_0000 | ((backslash & 0x7F) as base.u32), + link: 0x3, length: 2) continue.string_loop @@ -294,7 +289,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { // Not a Unicode surrogate. We're good. args.src.skip32_fast!(actual: 6, worst_case: 6) args.dst.write_fast_token!( - value: 0x40_0000 | (uni4_value as base.u64), + value_major: 0, + value_minor: 0x40_0000 | uni4_value, + link: 0x3, length: 6) continue.string_loop @@ -358,7 +355,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { uni4_value -= 0xDC00 args.src.skip32_fast!(actual: 8, worst_case: 8) args.dst.write_fast_token!( - value: 0x40_0000 | ((uni4_high_surrogate + uni4_value) as base.u64), + value_major: 0, + value_minor: 0x40_0000 | uni4_high_surrogate | uni4_value, + link: 0x3, length: 12) continue.string_loop @@ -399,8 +398,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 2, worst_case: 2) if string_length >= 0xFFF8 { args.dst.write_fast_token!( - value: 0x20_0221, - length: (string_length + 2) as base.u64) + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length + 2) string_length = 0 continue.string_loop } @@ -429,8 +430,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 3, worst_case: 3) if string_length >= 0xFFF8 { args.dst.write_fast_token!( - value: 0x20_0221, - length: (string_length + 3) as base.u64) + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length + 3) string_length = 0 continue.string_loop } @@ -459,8 +462,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 4, worst_case: 4) if string_length >= 0xFFF8 { args.dst.write_fast_token!( - value: 0x20_0221, - length: (string_length + 4) as base.u64) + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length + 4) string_length = 0 continue.string_loop } @@ -472,8 +477,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { if string_length > 0 { args.dst.write_fast_token!( - value: 0x20_0221, - length: string_length as base.u64) + value_major: 0, + value_minor: 0x20_0021, + link: 0x3, + length: string_length) string_length = 0 } if char == 0x80 { @@ -483,7 +490,7 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { } } endwhile.string_loop - // The trailing '"' is filler. + // Emit the trailing '"'. while true { if args.src.available() <= 0 { if args.src.is_closed() { @@ -498,7 +505,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { } args.src.skip32_fast!(actual: 1, worst_case: 1) args.dst.write_fast_token!( - value: 0, + value_major: 0, + value_minor: 0x20_0013, + link: 0x2, length: 1) break } @@ -517,7 +526,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) // The ',' is filler. args.dst.write_fast_token!( - value: 0, + value_major: 0, + value_minor: 0, + link: 0x0, length: 1) // What's valid after a comma depends on whether or not we're // in an array or an object. @@ -532,7 +543,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) // The ':' is filler. args.dst.write_fast_token!( - value: 0, + value_major: 0, + value_minor: 0, + link: 0x0, length: 1) expect = 0x0EB2 // 0x0EB2 is EXPECT_VALUE. continue.outer @@ -544,15 +557,17 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { { number_length = this.decode_number!(src: args.src) number_status = number_length >> 8 - token_value = 0x60_0004 + vminor = 0x60_0004 if (number_length & 0x80) <> 0 { - token_value = 0x60_0002 + vminor = 0x60_0002 } number_length = number_length & 0x7F if number_status == 0 { args.dst.write_fast_token!( - value: token_value, - length: number_length as base.u64) + value_major: 0, + value_minor: vminor, + link: 0x0, + length: number_length) break } @@ -582,13 +597,13 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { // -------- END parse numbers. } else if class == 0x05 { // 0x05 is CLASS_OPEN_CURLY_BRACE. - token_value = 0x80_4011 + vminor = 0x80_4011 if depth == 0 { // No-op. } else if 0 <> (expect_after_value & 0x0040) { // 0x0040 is EXPECT_CLOSE_CURLY_BRACE. - token_value = 0x80_4041 + vminor = 0x80_4041 } else { - token_value = 0x80_4021 + vminor = 0x80_4021 } if depth >= 1024 { return "#unsupported recursion depth" @@ -600,7 +615,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) args.dst.write_fast_token!( - value: token_value, + value_major: 0, + value_minor: vminor, + link: 0x0, length: 1) expect = 0x0042 // 0x0042 is (EXPECT_CLOSE_CURLY_BRACE | EXPECT_STRING). expect_after_value = 0x0044 // 0x0044 is (EXPECT_CURLY_CLOSE_BRACE | EXPECT_COMMA). @@ -610,7 +627,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) if depth <= 1 { args.dst.write_fast_token!( - value: 0x80_1042, + value_major: 0, + value_minor: 0x80_1042, + link: 0x0, length: 1) break.outer } @@ -619,14 +638,18 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { stack_bit = (depth - 1) & 31 if 0 == (this.stack[stack_byte] & ((1 as base.u32) << stack_bit)) { args.dst.write_fast_token!( - value: 0x80_2042, + value_major: 0, + value_minor: 0x80_2042, + link: 0x0, length: 1) // 0x0104 is (EXPECT_SQUARE_CLOSE_BRACKET | EXPECT_COMMA). expect = 0x0104 expect_after_value = 0x0104 } else { args.dst.write_fast_token!( - value: 0x80_4042, + value_major: 0, + value_minor: 0x80_4042, + link: 0x0, length: 1) // 0x0044 is (EXPECT_CURLY_CLOSE_BRACE | EXPECT_COMMA). expect = 0x0044 @@ -635,13 +658,13 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { continue.outer } else if class == 0x07 { // 0x07 is CLASS_OPEN_SQUARE_BRACKET. - token_value = 0x80_2011 + vminor = 0x80_2011 if depth == 0 { // No-op. } else if 0 <> (expect_after_value & 0x0040) { // 0x0040 is EXPECT_CLOSE_CURLY_BRACE. - token_value = 0x80_2041 + vminor = 0x80_2041 } else { - token_value = 0x80_2021 + vminor = 0x80_2021 } if depth >= 1024 { return "#unsupported recursion depth" @@ -653,7 +676,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) args.dst.write_fast_token!( - value: token_value, + value_major: 0, + value_minor: vminor, + link: 0x0, length: 1) expect = 0x0FB2 // 0x0FB2 is (EXPECT_CLOSE_SQUARE_BRACKET | EXPECT_VALUE). expect_after_value = 0x0104 // 0x0104 is (EXPECT_CLOSE_SQUARE_BRACKET | EXPECT_COMMA). @@ -663,7 +688,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { args.src.skip32_fast!(actual: 1, worst_case: 1) if depth <= 1 { args.dst.write_fast_token!( - value: 0x80_1022, + value_major: 0, + value_minor: 0x80_1022, + link: 0x0, length: 1) break.outer } @@ -672,14 +699,18 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { stack_bit = (depth - 1) & 31 if 0 == (this.stack[stack_byte] & ((1 as base.u32) << stack_bit)) { args.dst.write_fast_token!( - value: 0x80_2022, + value_major: 0, + value_minor: 0x80_2022, + link: 0x0, length: 1) // 0x0104 is (EXPECT_CLOSE_SQUARE_BRACKET | EXPECT_COMMA). expect = 0x0104 expect_after_value = 0x0104 } else { args.dst.write_fast_token!( - value: 0x80_4022, + value_major: 0, + value_minor: 0x80_4022, + link: 0x0, length: 1) // 0x0044 is (EXPECT_CLOSE_CURLY_BRACE | EXPECT_COMMA). expect = 0x0044 @@ -691,7 +722,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { match = args.src.match7(a: 0x6573_6C61_6605) // 5 bytes "false". if match == 0 { args.dst.write_fast_token!( - value: 0x60_0401, + value_major: 0, + value_minor: 0x60_0401, + link: 0x0, length: 5) if args.src.available() < 5 { return "#internal error: inconsistent I/O" @@ -707,7 +740,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { match = args.src.match7(a: 0x65_7572_7404) // 4 bytes "true". if match == 0 { args.dst.write_fast_token!( - value: 0x60_0801, + value_major: 0, + value_minor: 0x60_0801, + link: 0x0, length: 4) if args.src.available() < 4 { return "#internal error: inconsistent I/O" @@ -723,7 +758,9 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { match = args.src.match7(a: 0x6C_6C75_6E04) // 4 bytes "null". if match == 0 { args.dst.write_fast_token!( - value: 0x60_0201, + value_major: 0, + value_minor: 0x60_0201, + link: 0x0, length: 4) if args.src.available() < 4 { return "#internal error: inconsistent I/O" diff --git a/test/c/testlib/testlib.c b/test/c/testlib/testlib.c index 91327942..fe61b707 100644 --- a/test/c/testlib/testlib.c +++ b/test/c/testlib/testlib.c @@ -1089,11 +1089,12 @@ do_test__wuffs_base__token_decoder(wuffs_base__token_decoder* b, uint64_t pos = 0; while (tok.meta.ri < tok.meta.wi) { wuffs_base__token* t = &tok.data.ptr[tok.meta.ri++]; - uint64_t len = wuffs_base__token__length(t); + uint16_t len = wuffs_base__token__length(t); if (wuffs_base__token__value(t) != 0) { - uint64_t maj = wuffs_base__token__value_major(t); - uint64_t min = wuffs_base__token__value_minor(t); + uint8_t lp = wuffs_base__token__link_prev(t) ? 1 : 0; + uint8_t ln = wuffs_base__token__link_next(t) ? 1 : 0; + uint32_t vmajor = wuffs_base__token__value_major(t); if ((have.data.len - have.meta.wi) < 16) { return "testlib: output is too long"; @@ -1101,10 +1102,21 @@ do_test__wuffs_base__token_decoder(wuffs_base__token_decoder* b, // This 16-bytes-per-token debug format is the same one used by // `script/print-json-token-debug-format.c`. uint8_t* ptr = have.data.ptr + have.meta.wi; - wuffs_base__store_u32be__no_bounds_check(ptr + (0 * 4), (uint32_t)(pos)); - wuffs_base__store_u32be__no_bounds_check(ptr + (1 * 4), (uint32_t)(len)); - wuffs_base__store_u32be__no_bounds_check(ptr + (2 * 4), (uint32_t)(maj)); - wuffs_base__store_u32be__no_bounds_check(ptr + (3 * 4), (uint32_t)(min)); + + wuffs_base__store_u32be__no_bounds_check(ptr + 0x0, (uint32_t)(pos)); + wuffs_base__store_u16be__no_bounds_check(ptr + 0x4, len); + wuffs_base__store_u8__no_bounds_check(ptr + 0x0006, lp); + wuffs_base__store_u8__no_bounds_check(ptr + 0x0007, ln); + wuffs_base__store_u32be__no_bounds_check(ptr + 0x8, vmajor); + if (vmajor) { + uint32_t vminor = wuffs_base__token__value_minor(t); + wuffs_base__store_u32be__no_bounds_check(ptr + 0xC, vminor); + } else { + uint8_t vbc = wuffs_base__token__value_base_category(t); + uint32_t vbd = wuffs_base__token__value_base_detail(t); + wuffs_base__store_u8__no_bounds_check(ptr + 0x000C, vbc); + wuffs_base__store_u24be__no_bounds_check(ptr + 0xD, vbd); + } have.meta.wi += 16; } diff --git a/test/data/australian-abc-local-stations.tokens b/test/data/australian-abc-local-stations.tokens index 97ffa61e245e29794d61b1a3939edb7e54ceb1be..f8b40ee60358e51ef83e6355f609a19e27a32371 100644 GIT binary patch literal 67792 zcmZPo0Y(tPz`(-bAjrVLzzBw527@pI0|OHS3nN$sBLjmX0|Nsa1BA=K$N-k-KvU1j zAP7+aRWF66UItCQECUCMdQ~*_>L80iE@WX)aAaU$&_M7VKW?#Up_qRWP5osg_kqIy z3Yz|_X!hMi(|;RH{T&7&6#L$xssDhc{v(?I|Dmd9gydj&{$*rj&_FhyQ4>wQHmdoI zIt+X$>aEb!+n}kpMNJQk9%$;l(A0Y~@S~U?g{D3RO?@nb3X1whH1*AB>RV9rGh;WJ z`d&2keGKv_=C4IlkCy%zH=xEB;~_NlN6^ea%3zLS{vR~;f05FY69WU|KLlUF37loX zr7yUAR%BpcLMxw{(8^~fwDOq=t$b$UL=7J%0aW)e38A{5Nf@=fV-iJEFOH^O0=2wj zaz<0{il*KTwY+2ULsK7s2oDEG1_q`;H2p!S`Hd+IO@9QM`bgCBjwudJeFB>LMAY=n zl!K-|58*ygdB>EGTD~z=q48S~>OtilQ!5f56yNP=_IIG>cc!Um_D@GsKLag1=Afyc zho*i$YI(=B5>5STB=>>BA5>n0QzWGRXIhJ9-$pe3o6+ps!XS#AADB*}sXvXT{tTM` z(aKXMwDOec4udJO`OJc-<}(YUn$IkPTHY}$p{Z9vQ?H6zeleS%sW(GYZ;qzk4NbiV zntD$L8x;4qqNztqAIu%7`I)&NP5neP^CvMVqnN)HP5pK>^*d0@JLXep>d&C5Kg(c^ zY(5JIs(Kbqr1a#(z`(+V;DgFL79Ip2)V^ckMKzy=4>i592%(uTf~H;+HGEiP(A3MJ zsh3AB?^qPk)GMQ@S3xcBSp3n{qm@@IL8#>&3tIih0;>PO4JH-_P`OsYpNd-Ev1FmC&p}h4i(1~XRG_J^Lc|ZKykn_GE#Fw$(D)M&>Otil3n)Fp{0EBv z$!PXZLCx@kqn3B9lBnjh zN~4<3DubFHSk=+gqxHX7wNT4DR!cPX)@bJ2Fc_h@KMzfP0h;&xg&p|VPE^2wlx*kpaMl|)CP}_H`FVWP$LQ4Nm3=FKV5qwa2$NCn*2bFiM z@6gPDkDA_CKckud6;1s&wD9?hrv5*wdNxp#5*A*H3=C{csOs5R(A2Y{mUnChXzGp7 z)ElFgcWl;Z>TQwy@5sQwW{0NV9@YPBXzfWhwDu&M8)|vS=8a~bFPiy&sOg(65>0(H z!hN9fjxC0P16({n`?GARXnatA9;|?c0aV_xA+--3oER9`(9$zoA!>eSL+h`zq4n3< z8qvZ7t$o0T);?hCKrQdsCZYLn3c`P&`i5;PTKG&uvkxu*u%YE2HnjZ1hL(TW(DD!4 zCN%%=M|0mnH1`}rE$`S~qp1hwCs=wGW?*1@hg#mTq4mGmK>aOb_5V@R13MS0`Rt(f z36gqtUexlA9j$%Aj@G_lmuFBzF7Mb~(ad*8^)I^zYJO%%YhSRVwJ+F%P|G{^LNxP> z(abMFE$`UT#(&t+#(&s5QOi5_18C-h;tQ7kofsI{4rb%XK`rk%*ip^r0JUe4(gOz< zYI(1GI}Wt|3kO>Ng+moJeRH6d z-yERw8|pq#dB8+xe?F-Bodd1B?=jnUyf#91#151Xh&1uiKe~_&HvNU z)XzjyKMQsIjstDHi360Mki+XJYI(Equ`W6P#%M z3C{DVPUX`j-oB{E>?ZwY=j(t1r1g^(Cx)aAaWMLhH|S zq4npv(8^0LP|HbtlwY=j(n{VMpn{VM}L>*t_RzWqNTMgAc-0G!_{tGnoU!s=k)irP@1BE}2KU(+%pxGCKrauf#eK>0V z=Sf3Tk5-=YWTN>WZT_AIG=GompGwsJAP-u8;{oL_Wc7JOr+ zKa8gS2%3M<+K)VF?MI$#XzoYrU-O{#uX+BX=4W2C`AS~2`AS|+1|{VBfL9yU{k*!U z?&sA*E$?_8(9}Djsdr|uLNUJ( z7p;H5i`GBj?LrHmiD>RY8!zXbf?D43PD4{a1I;}%QOi4CwDl0Y$I#RtM=kGo&!d@t z5y}6K3=F)N(DYwM^Z#`;{WsCn-$E_#c<-aBe~7035o&qI`xZ_8dxZNywf-tn;^@tqhL_|VcPA17*l=MzJBj*P`wDOe?RK6mWM|}Qh{*OU3KMu|Oc+~a< zA6k0lLrbrGGf~SgzU649m#w-ZhMZZ!3K(A1wsQ-2ms{W%78l<-08|MH>r zfBAl(=4U>%{uDo2e~O<8wf)GCwjPcjR39USKfeZQdB=~oUWXrTy$-)E>ijG}+WZMW zXgwKHdB+c$Pl575mTs9qJ|HD4{G@H_o12J zk6PaGqm_^RXyqgS6x8yLA8mY@A2eQx9G-_z%R7Fw@kf5p_#-qtKd4E${eWps7ciZvvHf{I3}}!On%wSMYyDpC!mb#KLJ&w z^ykRHAb{4N5>Q9853Ri*fL1;V=rM>O=Vt*YH2Yjo!(YG^&Hq7Y>O;`fhoY`u6R1Q} zUyY`|2DQ8s=t5K9gQmV0%{^%IGXiMyGXir_%R7NJXy&g&Gk-mU6H54?^>+k7{T<}^ z{(_pH1%9KM4_fbrto|SB_?@6Us(L|1RQCufp_X@oXzOnT(bnGxnxl^22^OH4Ux<{R zoER7cix7NJc_&zc;Dg$Cf~9EgDML+ff>mhdgUU-J{|eTkhL2zin))_0^V?C&JHbvg z_1$Rddl+<(<4X{&|0{^r{}tSgTHXmBMl&BYKLaZd92pn{(b|WC$I<+cR(}ejm5+kw zP|G_(wEl%4TK_`u8ftkb_z=x~j}h?$D(?iJpq6if?@-%!fjIlt*;akMlJ7z(AtASp!OhA_zNkZ*{6tRA6k1s z2(7&!q`@G7oIZrm`sYGu{c|B}H2 zc_)n49u!7v4+>8~E$@WU*2fBi*2lv9@5sO)JRi+{3()+(1kFFo(9|zSE$@WU>RVy7 z`c`;7YI!HT3(b6_`6N(zC%hN6d=ox_#s~GsVg6BKU=Rk)_rUm|_(v;mh0*2%h0*2% zh0*2%g`c5?$161VzCjC*x2WZvFk1Q-2Bm+P`#|CU3r+uTH2cu{QzB^XRS_oC{4avG z9#{lzJ+O!fYI!Guw*FHDZT+W+3hH_k5nojQi=d4^iUgpRUm|Gh&qdJIpNk}-nU{ZX8t_X{425)P5p8-^(#=@b0TQ#yG78}cZ-}s zE#E}Y=5IvM=5ItEqPFKm(dPd|LF*S`>BotIK@>C}2<3yyH&M`f94H@DzKMeBSETY* zQ~@=8iK5Nlh@$m3MA7DNMA7DNMA7DNL=92PH&J8M@Dw#g4Np-s)bdRftvnP(D-T6e zQOh?`wD~+y(0m@u|Beg{qG;QQiJ79A zZ;ocZ1!{RG=8mQwwEhSwe8jv^$J4~n@`D&CKOn0wKrQdY%F)aRjgKL#uVRozaX;F6 z9x=4_JYus@%R4c&_J9~#dq8YG>V6inCur^g#V;&9IWaJZJwxBpw-u6zfsFOF|_r_VxaLRr0^7HKrQdYbx_S0 z*F$x`xISumCvJ|W-VzZWpz%6!wEm4a+Wez9+Wez9+We!q6KZ)Uj+Xw!(bAu|7ixJY z9*X9^aD@9n<(+s0YWXIfjK&A`pJ4H+#K0hqw!T;#tvnY;TVE`Ww!T;#ZGEvg+WKN~ zwDrZ}XzPo`(bgA>qpj}{2d(dbxz~|_K^$#;u{hfLVsW(bD{-{(EAhFg<(K#>wD3gR zpC-N*O+8wBS{&4#MoxcwQTu=5XyvImD8C`Ae}-D#iGM+J&o{L6`W;oh1S_g~33gQV z5*(=IorE}=dPy|(QVh<><&gwhe@gvtv5*6&KH zqLz1(x~S64NxeLwz z$I;ZEL^J;s>iE7C+W3+b+W3+bKWh0Ug*IL(h1Q>vQb29rNujj|rO?`gQaY&RofO)B zAt}&)Ate7w*)cdEmv>TV^_vu0{U+6jnxCc6;!g?`f5_%{Gf1PDzYH}zrBjc-c(qn3BlXzRnI(bk7ahohEv(rEp2 zX{7QBRNhIaqLy#c1!#Pv`F7C!gfwV=5Ej2q3=Gm}<&kt9YJQhS>rYFA`qRka--i|+ zXzev=PzEk?5sZT*Hc+WHOYRcQV}EAOPy$~$Sa z@=hA9ypu*N@1#$mu0N4MTTdy2ww_Xk1+~1BL0g|LgSI|hMiezY$S9zuR~aSL@R3nQ zE$?K|*5}BWps6=S?a#`ftq+$$TOTeHjhdfj(9)+2D19Quw@fPP`eGTh`AHeH`AL~} z)bdVdCYpa|qq%<$TKJ>Q|H**nOJM2AiGe`|G(HXGgUUOZ>qz-sk%2+x2AcUdQPZ2u zT{QQgwFhM$poPzKH1%lxWtmr~<(&-L`f(Yw_2V+{QOi47X;kxNWl`NPD~DR%$*Q5L zM_zyH$iN_rw*Oid?fffQwDCw;wDzE^A!>OiYldc@1)BRTQPa09T6rvsR33xMJ6U(s z@=Z1fjgNfZff56QEaG|-$3>~n);7u>OY~DcXDXuuN+$W zE60LbUdW-f2j$S(gK~DL`B@Haf2|x^`#{bEwLdGDf$Dy_EL8W)WuumNa%kh-a%kh- zat)~cSvj=z7;>Qc43?gp7#QS0>ocHyP zgId1Ho1^hTnX_@j*n$!DS2hjxC4Jlc4qd?AA{a(_o2t$iVn*1nK$M=kH< z(asx`2c0*Fl>X$WGuWY+e-$k~fZ7kp>TjTyck*caf8;^q=g8_`qLz2^XzPFF(boUU z|3EG86wvmME1>NkSKvbJ?<=6y-wJ5;w*uPvl?rI5@8lQ!wCnp951*G#HLFJtS=)66cydndG!Uojx zLSZ9ndQ(6-zlK&k5E8c|E6#pwY*b68}Cy9 z_0N&r^9{AUQ$X8Ks0i9m2=l)q1A`*kdIm+b^Zpdk_RlDy?N3nTL@nLsNeNP5nhQ^=R!!MYQ&#;$zhL3MI7ujuNQ9gXDfCP1O9Xgm!*~66pL2Wc9|V z>rs@@#&eX=#&eXyQOi4}OjQ3WWuy96DF<~uiW1uT7$wm916X=;Vqj1Lt(SrFLFJti zX#4}p2et2%(9SDUT8*0Cl+e!8Q9?UUM`;sU_@M27Q9|4QqO=FKyi?kT=HCNo{ym6V z-YKE&w^9P_w?Yp852)px654#R5@@~{=6^>91|_uhw@QCd{jZGHo>WF_PbxE`mUqf% z=bT%4qu|mC^P` zD&Im~zo(40{!s<3y`aK`TArzhpyqECG1T;^BFz}J+qn2kXXyZdFXyZdF z1*qkj3fg*L70`NMr1(^sk6NCopp9p%AeZNk3=Ar0<+%!4d9H$1o~xjh=PH{}%QF?U z@o*K;_yMx}cA=JMDo4@c`#5TOrh>L!SmiQmd8Tp~p&ryrbhm^`}&Rpq6KYvs z6>Yvsl@~StsG{vpQbpUJq^gAKe^q_d@K;4E|5c4p%QIE9{=F(%|6Vl|wLDWzMl(MZ zHNUB*p_W&wXzMvu(bjXSqMc`@ind=}6>Yz|Y8QhKa(SV84b44hu~e-UbVr-nAZ zt%f$ftyYd&-l;7{Gk+;+`cYekTHdLxLsJjxpTqp`$iSe6R(`2%Li7JNH2rAj>#FTU zE$`IO+M{Y{?NPOZsO6m++IXfKXgm|@K2ZBj?L2DvrgjUB{{*$XQvgK5ZRduxWY3iW-gyetq z1l00Q9c_G39c_G3y%2T1i8|VTOm(#VnCi8t<)b><{&RJ-{pafasOzEC(ayV9M?3Fc zeG6)SR!3V8qK-D6uf7j;J+b<2)cm3js*jPw=N@W#r~Uy=J!t(1vii@c>xnhc&O_4x z%_qXrlM@4jh87|}fXX`!s+r6E*5l z^P5I9n)^WaJs|tP4Yj<}KwH0|fwq1_qYt&b(?FXq)Bw#FLfr={?=@XoB(+QhLzTMJ?|%(dIif(dIif?NG}vO|<=InrQpaGy~Ao$D_GF5jDOwlhE9c zc0RBs+WEkmov8U)a{`+AXzLp_C!@B1HPOy@(?mPpO>-M+d}*TXuh9hUuR#va)2Q`_T4>`_TFX((J1w;K zgchhh0dpTH{L#u|t*vP34{f|x3vIktYY&=#(8^;iwDMRB?fxAtwDYpHK1)S{to@tzCgy-f7Q3O;6ghP|FML*{J26HroAU+MxMT zWcO@9E$_6^&Ii&4t!IO!Cnp95ZRGQUKzz`CZJ4|w1A{i&dHUM#P}7_CCp7n?op-7I z6)k+w`a9Za{T=Q9sO6mw+WH_JwDmzcEU4w34%&Q>4%&Q>4%&SbI%xH)4yb;G`QMR& zK?kkn1|3j;3C4F~V9-Hp|LPQ@mJd1&sPU_VcD|tw+WCe$XyZjXXyZjXU8wn2 z2W>rv4ro0G%zYsLq1|t$gLc1}4%&FF4%&FF&U`fgpzXKRLECSsvmQ-7+We>v+We@_ ze$@P|gLYn_&O_Avtn&!9ywgGJuj!!m*K|Ikt~b#^8}HXe8}HXeJMTajt-r5}*5B6^ zL2cjZqOBj%MO#0jYmJ(pbYI&!Nwx3oPZ9lCZ18VzD4{bhF4>TVN^S>hl zgC1x+38_BOLp#q%5A8f7J+$+T^n_8{cY0|1#r4qki|a|Fw(s=N=Ku9T^Z!uyf!cR^ znyBqLJrguOsJw=%2et3?K>O2Rd?yA5J+%2$Jtx%su7`HNiXPhiDtc)5P3obwm-RsH zW#sUSL@n?1(8lZaK;!i=_kqG6Z9kh{I%@jULu-%dp|wZ!@=)hr_0Z;1_0Z;1^;*#U zk9K~J9@_aidQ(u-pB~!%FM6Q+UXcBN7`439Lt8(phqiuH?+R*pr-wE_poca;p!XQH zeW!=k9??TNBZkf%R7Cv^Q-ki<4?%uZ$?dT`e^m5K3e^% z{|L3b(?>gxNFTX<>&U>MkJi4`M{D2eqopT(wDhF^6}7z6|AiJ_e^Bdx{lBQ?odH^T zY=BfAgUUMtZq)M4Kn#r!Iu8ySzo7EY05l&7<2x}h7@(b3WT1kY-wg~=-D_Zs>RtmA z)bwD0);=&mYabX`qn39DXzTY3K+G`#Rh2m z*A39tD;s2?`5*0k4+FIGJq)T)&v!CFn=dm!n=dn1hFabkpq(#c0BVmQrEi1nXzJ1C z>kUBj^~mZ^pq6(AXzN7`(8fazZlT^EWQaCjZU`FxLNecw6*WH_qU{GZL|cDv$dB57 zG({?L0w4(D`)8>b+6tXARN%7lxqmM_77tVqh=?^*5k=PFr5Mzn5!!h&M&YRCw^0OYd1r*y|1v`Be;H+1UI)7~}hnioELG>|G_!ujqmUqTz<)blL`Dkp0x_`wOZM@qU zbe>jxLsmZvwY)Py+uv(~w!hb8 z4VwGW`qw6C{cDpesO6mr+WHa`wDy6?J=FfJ3EF&)DcXFFDI;opnWF8-G)3EwX)1== zpEX5`FH=x^AC{h+7#K`J@dxFD+IOa)`_7aosJt^h zgj(L2o<-w>_Gd!XgUUNo@jM2 zXNJ~(G(&4Yn*BiCFJ_JwU*@3sN?3YwVqh=_#UGRpD(}oe?L{abRNk4Rt*RxI4CZL(QJQz4`5$dPr#aesPV;`$^3EJ>KGGaCABpU~>8RzM z`64v`gUVm1`#|NL`BK#K&3pqIA2io)Mpv^~Fpv^~F2&0ZqS)iQ{YXLeR7RfypR;cBj1={$O1={$O zg*R&XWf6s%9xP%|^Mge!n))0x^`P@yk=>Jzx*yA84Vrqi{<6h-)bh>(ZN0q(+IoA7 zov8JJ#T_*BLHm7>-E$wcyt6<%ug(H=J{_|9FR1s8S)%QSu>`HZfu$!W1_n#e{un49 zRNh%4m4Be}&Ju0?u%#(#db6}db-yLrd2*I$_lsMit>3UjTfbrHhFacPqNN8*wDe%< zjauGWqKz+GqV>lut5C~3OSJvomZ1IK(C`3_-&vxqAGT~m^FLbsX^B>UTK1xrca~`9 zzgeQ4|7JNAwY;-L8^5#!jbB3D2P*F@7o(PMmg`XavzAEn@1XL|5|kdH`a%0EEJ5WV zQu?vnjhf#r(bjufqV?}B(faq6X#IOjwEn&2CDiiH5-t5(g3>?CeU1zamT2dPS)!dM zYl(KAtR>ocvX;+L^S|XcwDbzPp9tCiKT-W}g?8S9720}KD^Aq;3M;hr_f}~6$I2A7 zyt6{f4_0XT!O9KIJ!t!htw8%jk=$<;hU#A{wDCtPwDCu)4Al0W725cb725cb)dbY? z&I+x6ZH3mqwwi_7zOzExk70#Y-&*ZME$^(*&f~X2JCEP$GHUzI8g2f>8u>gUCk6&< z(0mFseSyk5Yvl25MFs|I(D*h|cv_>~r)Q0}U)>tCUmaP!Bx?9rqn$@+jdmWPwK8gX zXN`8Blr`FYQq~%%<()O!eqn3S{$OPHgrk;s)@bX`twHP0q2U24@2rzh>rd-s)bh?6 zt-o!J*59_yLM`vC(e~3>qwS}&E+1$7^&yt8gVE$^(m(fFYA&Y|i- z<()OCJqhE3;vcQQZ#@$=e_5lQH*1Y{-mLXXwD3S%A8L)ZKGb>xYI$dkb{@Jl@_A#9 z3=Gz2`N10PyfJIE{U6q7`#-FYGDsk|@2t_zW3)yqPpu!I`5*0k32U_TC9FT7j>p@e zD5LHH9gp%&F|QN=68_QquozogSP&|25tR^4ch$_ zHfa0RZ9wZckiK*&XyY$7p!Pmedba_M$3Xd@^3Db{UkK%c+IKc+^D{PR>oIK5)??V9t;euITW@TG zHeO=`8m~ciKiYa@TeS7YwrJ~(ZPC^n+oG*EwgvUyk<7P6TW@TO)*rV8^~YiUcVu9& z1+8C0){nN{*cPpRwneL-ZPC^n+oG*EwnbZSY>T$u*cNTQu`SwqW7`Udh_>WybYI$deHok5L z8ed2D?-JDV&JJz<#147Ag%bmV9cVrUn!Z5gogJvYgYrS;ogLbKM7xKm>CFyp{?HCI ze~9d!7ii&wHlA&VHlA(w3AKG^hjw3y9oqTvcHdFkclK!aG24UgV@7h1y&7u!&K_;Q zjXiRI-;sgA9@O7Q)^CLBe|xm^i0#qVPuN?cw(soG<_GQ3<_GN^QPa0Q+IeyIp!4FO z@dIk#+54lGZ}w4We9(LlR6VGDXOFy|*NK6_9<6_4pMjd+?a}rN*`w_jvadi3548JH z?a}rh+t;I(clK!Iqdlm6gt-?K{%GR^_Gs&&?9tXk*`uw8vY)~riQK-kM{9rDqqRTn zH=y}{FKT|a-;bJ~?GK>#XYF62seg^8{taq*XOFfX(;hURj-1~9qLyC{X#F(@Q2z`` zy#qg*dbIT)4ruE?926PUk;BIUZT!dqZT!f=8#O;Wpq&TffOZ~?Ln!L}tOHtlazIN@ z4i%{7odeo_=%2}iW`6OIC? z<((tieg{X;eg|lHfcmqJp!+G2(vzbMYW{Xa>z_NK_0JuZQOi3=wDoC@XzSA)wNcAE zN3`l~d?%R5I<`w^-hRNgrvm6xFLJ4ev`JhJ}+QS-Ya+Wuh2 z1l0WNh<09%BieY8BU*Xtn1Nc}Iij7f=?H3X!rbe~z~Bg4KZWc*wENo~(e@iTqU|?w ztYr{H&d-j0sOiaZ0-Ap&qWK@K{pg6+eso-j+J13F+h5^`w!gv=ZGVL$+WHSiwDlj3 zS5V7GN3`{xj%e#W9iO6>caCW71xK{@g5!78@h>N|{*e<}|H#Q0H9tF9p@yfE4QhBg z*`n?*c0wCJasrJXA;qs#5^8zpgtnf}3ABC=S$#R`{$D4w^$$*<{Ry!21X>RP z<%7yQC(wBXP(G-A=Y+Q2+X?MHQ75$VM<=xL7N>n^;e%HGI)TbtWcQyyE$^Jr&L?z2 zJD>Gc@0`)rr#YjoPjePWE$^Js?o)FHtw)2p4^-YctD=@~&IV|F(EbLfdQf@i44O}Z z@tqhLoI(9}WdGTr=67eb{6#i)a zDQC3tQ)jgCQ)jgCQ|Bzy{O^o5fA5SofA3s_=6|&Ir!%PiiR}L#)cov>Hs0inHs0jC z5w-ksMq6*>3|eo5Z2lp%^oo`roYC@w^I0_iqRj_6gXV*f&A*R2pWuQvf9--cf9=AI znx9?J&L4C^JHN|?mq7-(zH!k+&7Ur4>m^;#)=Rpe^@m*0`a>>`3??Y%qqQGhK;w$7Cva{#|16@xb&cw zcP?o4u?t##>@pFxymLY8f4HFaKU@x=rXLry`9~K}dlu$@M+OELwEev2P7DmLX#34wnNah)E86)Tu4waDu4waDu4v=+u4v=+u9B$boh#b-gDYtK0p>nO z1_oEO{g|%msOisD8#O##(atY()kEEX>WbDLaYbv7xZ0xmA8kF9E82P}S6|frtSj1j zV^_5G#;$1Vja|{!8@r;dH+JnpO%JY9(fo^co`UNP)bh?1?L1po(0R5<<(=y))bU4G zwDXN!(atw^MLXZv6>a^zE86;b*ITIVUstsDh$~uq#Pttq`_2vRegikqd@7Rr-2_p` z@7&P(<8Gk&8CZI9VqkCs_0OSvP2beeQ-qFdrpzh~$`-xhAy8S}44{g1^JKB1EceM5T?r7`v-O<+TyNjXv-yN+yb_bQmNcqiO z4Rw6V9qqh&ceMS#?w+XaJ9o78$?j!(bgxsqqSe$ zt5MglxuebJxbHv>fA^iJ`Pm(<|L%^~e|JBO7Cw(r%WHSg`LoF3`3$wZb4S}x=?>aY ziLCxN>iC@p+Ws>S(0U-G^zMPQz8F;Ad4T4(Ve*O$3?88IA|&%Y(DpBSpzUAuK-<6Q z;eZ-G9%%F99-#6U*?ce5^3DToyxs#e{)DXFAGN&mKs&$F19W~TvidsI^3DV8d_xc9 z{l$(93?69ptw$GXe)B+Ue|n&`KRqU*mUkX#^N}8)`AB5<%|tElJkZXQ^8lSE2aO+4 zdFQblb^Oj_BWii)fmA+%+IJqv>ph$p7(CF{!+Y#U&F>y)=PP@F&R0fu?|HQFK)c`5 z19YD!vifVN<(&uGdV3Gh`eLa6K;x$#k5TiR#}hRB(AFP#fYuuz+xHrEJ&GsVd61r< z^B+k2w*zdX_UU!IPr<(((m`7@qq=g)Wsp}7aG|L+Ov z|0BCU3AMcQM7xi}6YV|@&k_blbUk@#}>azg~{0<((H=d(#WlzC;R7FK^WH&I|4Q3op=qIArw!sO6m(+Wepw z+WeqbJ!*O9h1Q?&0+nwt|2r};c%kj@_3B3Re?Mw{8}yHCjr`Tiu3`&OWqZ(f_x_@MDrn17TQ7`#B`8;tM7z~F_pf86UJYX0&p#5F)_-{Oq4w9jby4Hj8+88>l6$-jP~GE=mVdm_@{hM8 zYWd|IfMz~u|0lBf!D#By_HTNF=IfEwr=rf!dZVou@kU!O;@yv$pS{uQZ*R2v+j}Nz zf7u&t|Cu)^JtDhjKWh8V8*M(v8*M(v`wD9N)CX<+#Rt^?g{3Da1_mF{cnp*eD(`$i z=Vd_opz_WK?LKoKQPlM2BZ(URK4|lqJ~F7`Wf8Y~^THg7fr3W8SdVu-gk%7Snw0;9AK77*9{ExQ&$OmoxkxwpadFO*R zAL#>{k3=@V6t%qbK|62L2Xx*h)P11#oli3Z517lq%)r3Jz~D0hwY>8IogWKUz`_7( z-}!*ncf$Cf_y_ghk?o&{n%{lU`rkfi{coRjXyLI5Ej^&^zx3IPTHg7fwGVti?E{$m zK;e(pKk_+@W*^#l4?dvv1<3w6g?j#-588TTAJFcfI zc!2s-erWAaKX26h=7-iF@k8s6_ywVscYbK)u^(D_>=%Vv-ua=e@9+by?|`}wRNnbz zpq6ib#b|s`e;%qHRNnc4+Pg5m69a=E+Icp9ji}|FA6kFi53RrMHxVs7(8kmK(8kmK zW}udLerW0650w64?sH^d@IxC9^;?E!-zwDn=?A($896o+K+x{?MJ_VsO6VG+Wfvh+WfvhKWcjLM>~(xAMHF&e_7P> z&R+-B{r;f)H1+@JMF8(;TF8(;U2N6pXvXyb|gXyb|gnW*hwf3)+F{L%VP z{@tkMoj=-r0{&?C3HUESov-jmn?LbKJ0HUzG@k-ZU!d~NAG970$_KUY{6XuJk;CUM zYI^g3h!&n`<1PMA(834pd@X;p{Tu%8P|G`iwDYe0(ayW}|BPDR1)$xx7XUin63P7m zN~q;s0NVX30igR;VE%VxU{qcZjsO4P%T6-h_)c!#B?|0Pl zD-dlzZy?%!-at+?^=Ru$13~Lck<1U2W-vl7?*h@rj{?!gj{*Zw^K&3t{T+x_e+NdP z_Gbgp+Jk{;?ZLnX)bcJ6ZNGUS+J5uEsi^(gK(zG|fync9P7DlzNb_}|@-7fGUk6KX ziVO^aXyeU+*HF`2AliAEfoSDp;9a!vLEGOQh_=5w@HuLE7l=0B7l=0B7x)IXybD5$ z-ypR34U$GJ?}E_!*Fm8EHO&8x3=BbN>mh>B)@E-OmPfAE>+wazUN12=YVYgU+*rst1*KL7?;1VSFbBh9I=_yMm%o z^Lr55c|}2J=M@EIqo#)-wEe+Bp#5yf{wqQ)?}E_ATZ53>pP=wZ%MU>fX!fCgVe(0C@2eIX*K`7H#kd<{V>Uqhr( z%exS?{TLx=`!Pb4QOmm!wEb=&Nc-JDKOaCFD^bd2NBLhPS+WAEx z`Do#fHr^D1Hr^CchT6UhL2D0&ptT1>(C%joLF-S2p!KIh7NEB8LeTDK2|-)G8FB`- z{0c#9FNC1A7eel$r3bY6=n%B|=#bZ_(fZ4wX#M5TOw|5tDB5{Kp$n1HzY_yP zDDwG2AU^W=i6R3-DBAc*DBAc*DBAc*DBAc*=vuV!K|8NM6m9=T=yufdE);G3Stw|I z1ycSA-Gf@*g`%a0P_*;|`W-dDg`$nugrbetg#JS< z@50d5LxiENhX`XsP2XW?_e+I=%3o-C04ndoL{QsvVRESDT^Q2+KA`qp7-&2ksvk5z zABMKRJxm9+ybD7+?<)+nKLaWJ!)#E)Bg_FcJ%>4=j(3GQqn3AJX#J-!Q2z<$K1T+I zFtq({VS#A&p`Fhj25QeDyDtLuzU?ry_G%bfdo>L0e9bVl{&^T$|2(W3H9v=;ozEVI zc0PO9T-5R|3~fFr3~fFrY#r)-+hJ(^>oBzbb=WS{@-7Um{0&1Zf5Xn8jz5N>jUR=f zjUR=5M$OM*Xys`bT6r4w7xljFaJ2EqaJ2Eqa9PyyE*z~q4M!_a!;Mhy+YU$DZxaqG zA7SaqiGd*;wEqUm2et3QvysX}MFxg&wE4#HT-5XyjyAp%jyAp%UV<7v;b`mQ!fR3U zLwFr(c^8h>o()H9&xSXnmUrRnQS)0k=)P&B^dG(nwY&>QJC7(Fd3~BA14B4yJqNOW zwDX9<(as|ZM>~%w{3vR97ml_cC>(7+Q21HY@-7^0KSnrcKL#wmK<>MZTHb{}MJ?~b zLGy`F^`P=D9Mu1T@j>yAmY%~uqn39OX#JrGwEj>8GirK>U`GxA2u{@ajo?Bp?;_Cl z*F}K#*TLN9$iNUGj;3D%%|5jISR>HxV~vnU%|8)n=if$v&U-|P-v|>l|D%m3M1aN< zkkz}O_Gcr|=9eNs^GnF;Gf>OB2(iklnu*H9tq7&HqH8&HqFkM=kFn(8jYO(8jYOo}#*t{}8%Ki9qeC_y?R?`%wE4?Or1m$ceHV## zzHubl`Nol;^E_exQDR_-1htQ0d{F$OmFJOY=Nm_!MlHW0(atxHL_6O&6778BNYHpN zvj5P|H;zO*-#8NOeB(&a`eS7MXy+S8qMdIX2|Di#S^X>2_FW`edn6LAJremBwY-Z$ z+aDT*wm&qA7j^wwlnH8hMWN-ND09^EE($F_M4{z}C=WFCX!l)4fyQHz{2LXCTHZyW zwP&Ny+OtvFsPnT?XycnvXycnvvr)^tD75q$1xlaD?pcaDeiwx{{uPBb{uOl!wY-Z$ z+y5Jdw*NQkKI-^gG}`*OXwdi)QhJXDt(SxHLFHYvENc5M8m+#JmPf5$qtV9uqe0{S zNdAphLk*v3wDp+LXzMYf4N%*6(P;CZ(P;CZ(Wa>FyJ)oiXVGZ$kI^Zp?Yn5S^RA;o z<3-T$0FB>8qn&pcU4Z6)wDTLHLHm=D{a=pSzKcd1pN|HO&m*g^N3GAJ(ayt(2Azik zbswm_i|$2j&qdEb<0Fk%fZBJ_p!36^`a%8KXteQ)=w+z+J$f^0{*4BmCxPtVZD`?v z);@?vYac}KMQz_jqox07Q2K|t&yj&48nmAo**>)Md7{zI=ZOZ*$04h~gqnY%(b^Z$ zXzh#W=V<;%n?H?48{dfjj@mwoK^xDF0p&L&_ryq}mR~Vw`{!fO#us9AP}6G+T6r3S zR-VR~qn39uX!8j%X!8j%o~Yw@F=+j@7_|Oc3|fCJ2CcsqgVtY*sYC5A$DoyuF=*vu z4BGm;7_|O-3|jv^W*6#xj4^2Y6JtR4UBJ?l69YpG(t0~kc^3oPuK|-+WMGIvYd^;P zK~3MWX#Jg7wEj*k+Ib4GX!T_*T74PIi(1~rqOF&X1+9-pNxi0uVm;B+dn1+ijtmU3Xzj^ZUo`)tjlagCjlaf*pq6*BXy9Ic^8W| z9ubQ?{s3}c7HatxTZ+a9&2PfuQ;C5gwgJh1P7DmOXyXI1&8Yc37HvOrEZTnL*vV+& zfwq4n7H$7X>@3vsE*5RSTrARlIZ%BQ3))YD96l@1>_fY6Ar|evh1m5BBFO74Vh^Lm z_fgdTa_liQ|D)}Xi$zvKEpV){1AthAL5u%%ey$V@!2@g_$-othlLp*3c6j{F^YJQ7HJ6}5vqqSe-%hAFEt-g;(tMB9MP|Le`wDqU)p!KIP_kzM7EkDHfpxK939>=4V z$MKU<@9T+2yH742?LN8qHE8}vn=gw8jRzpd?;h0s@bPH?LRNntwY-Z*+rJSH z8gE5b{{bz$(E8)?X#MecwDUm|(AG02psi<4;783r325U}325U}325U}326QC1hoEm zf(L^fa($D4Hs6{6nomITZ$c(&d6$5eo)XZ~Q$ihs8H)L6^Pvf#{s=5RIWaIKfaXV` zd{B9p06Gr=$_JHq325!dgk7lVEdlL5-vqS%miQOmmowDV~a z(9WkxxPTU4Xy?TzfX+ig4$uFn!V0M+SyOwEZ86+^GIfL~DO0qP0I0 zg;2}8L~+#mGf@(?yh)ToE$hsiD>hsiD>hsiD>hsi4myzHxX?=Um|jO3<`g={d|dN z`}q>l_VXp8&95cqqTa`th_+uk5wu?#DSafO-S?V^)?Q6SYp*8GLfwy*h<08`BHDVc zM6~lh6VcXZC8Dj*O1y%4K1w3m`3i|>9iXz`npkJ`RVK|5a~1$lhL ziGd*nG(G}NU!d|X1$6%uln-hTq@b-QPg#wc9#VForiYYWpk3=A1DF{YSQr>m(B}VA z(B}VA(B}VA&Z6l@JC7sfB5Ha`L7U%8d5ESTblwJX_&-Hm&y<3e9#YWKL&`_g^Mq2- z%Ewet`G{nGDh~rQa(YNbYyYK!+JB(nVPIhZjZdY5);A;T7e>v`sc7xNRJ8VBsuY@f zwEGHD(e5iqRYWasQqlH@q=NQ`AiEDOy{GD<*@rehl!`V!lxl|BzDh+aKT|>bjbZjV TF)*Zp%2Q-MXgn0k7jOaq^U4xf literal 28384 zcmZR807fv?;2_Arz`(@7z`z3LD=;ZAFfed3Fff3`85kgZSq26M4kUSX1_lPGeg#Jc z1_liT-$9Xqfk6w+JVOQs26iO<)=)cP<~uMjFlZym=b_nGfn+|&zA7~H8yOfFSdsL1 zGB7Z3Bk`vp^n=|yoq>UY3oM@i_WL3<{u+dO2PFmuhP6n1Ck6(F4GatnykPy1@Hv3y zp5t)yAo7=y^n=`g6;1se1_lNpB=bI^n$HLdYfug7#Pfvrdu=D{+ z4@_w3iwTszkkSK_Flu^Wl0Z!lOm3*@fhhoCpMxU<15*%ccrrzzrU#}()bL>fr7uvJ zGB7lN(gRaIJifr`feDnpkkSVeD7>KRLFs|170G^(e>+g)i)jX$d*;K<2d4+7)kyk5 z?gynWr2NOUg@J)V6cOG`XVA^_(*sK`YIE7#LV4BJn}~pMn}+EKAVb1B!o87{UCv9Z5gP{+(#*LFEf5 z3}NceqnUr7fq_925nrsJvJVu-Q1z@bsQzUImCs1&ffbY%jF9vfqLvS=pzr|23Cz5? zsOf=q6KeUu`U=TEP7DmJuMvDudSHEv;DgEs*7vCXW&MU~KO4wtNa=wMRMsP<2R38W z@_`Lh9zx?2R6elTqnghKDo;RWK>fq!hZ;U?(Fpf}$_F-3`hui4aC%?^r6*)QC_TaK zQ(|CXLn>b!oER9`K;em0U$BAdE2Q+#2Fia(>46QDeqiP~GBB`#>MvyVXyMNW3V&qt z522<9ws)xI1KWR8|FVPhBBcj*P+Y5l0v%HSvwNV%2RkVLfz&|5kG%vnJ+Oo7Ur?C? zlRt5=WME)Fh8kb&pz;Q(d|svnddI6&zQWCsHSI6ZLKp_UIE zpzwmK2h|rGpz;>RcVb}R@IlSL9H8`t6u%tFaPuJU2h~q7{f-O_9H8`stRB=pLW)0* zE;RFJLEC6B|A6Xqr1F8|8k+iNsPzZO4+aJXP`d@HpA%GGAjJnKC_jPPeK7ef)bzkv z&%nT7iAYbJpz;moA14L|&RvN32c-wjJqSLieBj)N8eg0s_aoa6Di4v$2hJ;~?FTN9 z98!AV0+ppO`yCk=xIkfn6rNn5`UojKaDmbXQhMM5m7h@kp!C26N_$A@fy)dYf8g}N z1!~_y)q~Ol7bq-Yd?yA5E*I4B=L$jfFBhmhf|w6Z4_xU;;RAAiCYt&J)bQs5wO2rS z92!4dQyCZ-ED-U*bpf?};ChH!K5&7`XQcGS4RSkD`+ytNA3%x^Zcuw4seIrLMs*K& z25No5y#UESP7Dm(3lV%!df;A+;Dgcw_fk~468-R|16zv^~iKYCA*X3mpGEp!9>x2Za|@Jt#f!fZFyj zz7qoj4=6n$#TO5#tp&0Z8h<>XJ_y7-aC+bY)n73EjtmSup!PAc`f$|z%mYd<$mUlv zFfbsM_dKBf5GYT;?Awf*K6pUoHBx%u0hJ%f`a$&zQhe}&+J~Ti6x2LkJ=FBT>&(Ew zV1-DJymd(N;l#ke3o1XL{sW~4UQqoB<%7}#Z!>Cq@q)@fWcxw&DX0v9**_CCJ@A71 zYe?k-??oj09T^ySLH#$R@Z`OPnjUx`q2?D}P=0~B4^%$zen2e`cz>hv`9Nh2R6Qs? z@PQgUFuoH510N{-km8F^64k$apgucNdf)@~4`KQp85sCL;f>^eK2YC591-4p{;1^- zUp#95B5B#9=2APl4 zp9ZA|eo%gb+3&=_AOH$Kr1T&F3OgkC2!PTDQhE>owP#@Z9T^w|Kyibt9+W;nZ6v7s z1zb_ngFqkH)KL~=#3Xq?m z>IFe6JAgKKV^N$k)gJ2ON{z2t~AgDbGlUHP55G+H@FM_qG_6xS7rU$_u1_lOD z9)j5qsy~p@gCMB=0W;r`fk6;dejtUXASitxr3XP!`G{0s2tGz~FG&9r=$IQcJqUu@ z>yY>Z*B^qQ`~$O3iGe}zCqh4{JuL{TpODgn5U72FJs+(cvIWjN^ zfx;eHJt#j4Aj$_JP+mkfA2hxMDs!OWBLwR2Af*Q(P0@)9$pODmxfXW+Cm_y@3 z1XNZanJ)qwpFtXb5COG6km6qiH2wxoPLTejND6BGAp**upfU<(J}5kq%4?Ao3=9mQ zGzyakmB&csg9xa-ic~&`N+X4j69a=NX#4{jKA`jCeG#PeB?ijRNbLhLP=5v~ zKEy!vKd4O&jSn$U{}w4dh=KAKsEq@Ye~RQECk6&Fkb9xw14<8KFOcLF85qPs^)-@z z#X#fn$o7Ns2U2E&+g4!ppkJ`Tw2c=D@eW39NaZq`RWWG44ZbV8C;voMcl@H>e z_B&KRsQn-wfm$AjBgPBB=|dcCJW3o?K120`(t|iCJtD=IIH*01>>jl7C~;7E2h;D! zz#xt`9wiQ{?~%;|jb|hCLFE}zc`ptMf28yv{vEY{Apy!uNa;aBih+Rv)JK4ZuLP)k zMv4y!(0BtV55wd^`3EUINPy}aP#S~D-$sgWCk6%y(D)J5f1varaSusek%2+t0c!g} z;yG%4ApuGsNa;ZWl>U*@gCwYI1NjMRza%JqK)=e_)ZKAlA!Sfr1+97NA<5Hs67oa z51d~mLH#S3en$odNl<$U+5MAI+l!K*@f~FILFEf5%|pXO3KZ8!=|Kt4<7b!hRbu%z9fXW=0{7TgJgA}N|1C8Oq%y(p9kOqw( zBAG7@YOf%r2We1wfRrAjk>V4S9;89_5hT9A=|LJ)ej@WhkC=X_&p^4!0AI4H2w^;Pl#u$G9NUa zhZLW(p!P0OdXNS6pFn*aSa^W)FH(At1KEX?p5#E~1yX#-f$|rqjDhNx%SJ69LP`(vps@|8eo%Ul2iXOQFL3$b-gvVfq~z8010qHL`k8dj!-+g@%_rD1U3Lv+D`Y13yDE}bEhXSZQhLj!@K=a2) z<%2>o>Ug;VD1E~GSb|D1y=lR6i&^D1ypHP#8hS z9~4354Km*ZseV*qU{C~=w^03{^q>gJD@gIB2+Ci`?g7S9Pew`)N}%x(nE8$j3`(H!ASCmZLFEZjdQb+X1rUasuMApm z1Jw^o56Xz~LvZ<^49Yu@_=E64kh7*c#FgU0(oZFp#Sses}JDLtrwI)I=tYnVK;f1DT?R6yfPQ2&9_g9@ns z2IYg&gGv-?`Je(CFF~?j1ynzQ?1Y&Qs(+Evg9>PU5mI_kK~8Uu3=Ar0=}iTc-jLFR z3aCCqN)IZa@(1cZPOtv2 z4Q>2I4K&_@6klqf_(ygRXubnt9ymRygWLi$&yj&a9i$gYy*g+-5@|e29W;IiQUeWt zbx`{TslHGLwP%sigF2}F0je`$>Ou1ZNa;ZxG+zWtV=#Hp`aGogPzSa5kjA6bLFpeU zJ*b2Flb|{is$TTLFqvYlvY7{VERF6 zAEXAx2aT5@r3Wogd4!Z6v_RuUNa;Zfxju4aV9-LVkF-GL5y(w2^FiwekkW$|X#5`P zK2Unl0+mOQ_yVT~ttQm|rxvLH2UQO$AGAQ@bud22|DgFIr1;W8n@`a~q%UxK&;qSb zfawRhAC#Vu!w1x!LN*VSo{;&V@js;etqn2@seI4|#T~>yVEeQ|;fa(Uv_bs|r1>aq z(0D6Sd}xE}7o_x{4az?tGoj(B4Jx0J(t|drKLc_D149F-eXR`|KY;Q<=|TGqBL9KX zgEnZs5XryVpz$hX`$6R?QhLwaQX5LH!4$_|gHb??6fqI-v3# zslTZU3R|S~pbM&xkjfWbQ2!DsK6F9-XC(7=LG?RQde8-x&q(P(7gT=0{Nu#HpbMJ6 zf$~A=K^IhiL;0ZepbKi>B887GsDFuUKPdf!+zgFRT~PlBDLsIy8|3{SdZ4g{neWKJ zpa=3hlKFa|wmDLI&;yOVBBcjCr130Jde8&ae;_lV?FT(j{}P!GDt}@2DKRkUfzlt0 z@5I2M2kK8C#g`svJPX-9i1BZ5de8%bl<^`QD5)Te{S zrvYfZ3fX+n_&F#I!{kBjZKU$S0F*zG`Jnz2QhG1|wO2rS0;=B-6jn&_VF*$S@)JxR zG~bMr9t=V28<6@JhM@U6n17rY7z{z}5hx#&9t=VAAy7UjJs5)OTO|J)g3>p#{h;}G zr1W428qY^c4~C%eSfun|2&z9}<~uSl7=qfr$mWCE{~$lY`~zBl0E%-MA8GsolpYL0 z^Q(~f0@n{lp!OOvAJpD~*{8(7U<4Z9g7KXg7>q#WFH(FNfx;5WJw~AMP>6Zp^k4+) zAHeiGGB6l{#?z72gXWq*a}ZGX8iCqN$mWB}Z%~?q$%EQANbzq3%736Zg~@~J8)QDH zzXj?8z~n*s2`N5|K>Z`6{);ilE~NBe3`!rMG83xb7&Kl6^N$k)gE4aZ48#YuM_}@b z3=GDg^o-4+V^I1*N)N`M_8C%oFb36^Na?}&BvSjtk%7S&)ZRrlAC!KO(t|N5 zKOm(CW6=C6)P11zU<~RnLgEXY9*jZtFQ^R+Eia5g<6Thopz^^ORNlb&P7Dm7;sANS zo(U-ZAi2i`l-D8VgVTcvXzl@~-;sgA1XOk-s|WcVDf~@9wj59Xlp1t=es9?U`IBU1R7gUT8t`^`b) z{YdG-9JD?GDLt5j){7#g2XoN)0?d3z1_pCb{e^5kXuJg}J(z>$e~{9HIjH>t)elM! z=AiThi7#+^FbDOgkolnTO{jWM`CtxeAH(<{|AWSFkiy>_G~R{m9?<+R#5{0%Fb9n< z!1ROM4+?K&^`QP5C=Wuz+X6KHj%2Guc(t`zP{Vk|X1l4b8fK*;PF)&zy`m<2~fzpE|Qu+X; z2TM@;K=Q98s4aqIza^-Dh?E{ILHA@0I-az6DoE|Je?OUY!!V=UzgQ^Fm2TRcW0E`dvKPY`5#g`>o z{bdQNzaZv=(}N|bJcj9aWMHra&Hp32A5@A@O#`~c*C zP<;jp6KH<12Fr59xLLGyb^?L8Y% zc!J^ps@?|FRz&iz4XCb1N)I-m{v#;NVd_ESDM;~Q1ImA(F)WxoX#5rwhcG^9ycSgE z!}w_9Q8u9R4eCEodayz64=OS+*nrxvNdC0}^#_sdM;ni_1*t_!54NE41}Qz*g33Cm zeW3Ee7Bn6J(gSs$Ehv8=r3YKI@hDqR`vs~WlpbtB<55WQZwnd^K<0zy1EA_b>A@D1 zzF>SO1_oPD`x+^}Y(ewU$nF8P{~_jq(}OK?f8CLR!4{Msk=29NPa*X$Y(eEWviYF) zG*W!pg4Pcqr3X-TjC|gO9Y_u-J=lTTil9CWG`#FU`3WgL>_F>9k2ep4;d?yA5drNjNdp#C^0 z&qL$O9#p;{n-4k<0I9yP2jxFdoWsVah!0A?FnL7=1_!kI)B#kVBHIs2Ur6b} z0hE4_(t{(Y{y|C)j-a+T%zQ@%21n3(f28nq1mzv1^xz2U|0AUbN048j`a$Wz5tM#F zbuzSn=?GdY0*OCx`fvobZ=vcz>A?{xJ%PsS9YOs+r1)|K)n~}=0i_p+dEoTm2wKko zGY{l`P=6g+J!m`_G{yl9Z%5F16UgR+$`7Rc>xec!=?LmSBBc*U(0B||dT<2gH;|v9 z<~xDPU!?eO0+qi=^CM26`7fmO-~^iQ0kvUa`qAe1oIvA=Q2&9-2PaVf1?`AaCXKJs{h5(9$^sJ#Q#4=NvA zL2d_yA=JIDps+)7k1MF`hnNRW53Zo}2h;D!z~Bm+FF;lgny^8de{comS7h@+0cWhl(OXzQa~LF0Ew<&P^UKZ5)ORqqBWyO7!s zZlLjYn17rY7~DYZRVW{n9^61{Z=ifodT>L_FK(dxf|MTIK=UJ@FooF%s;`mKgBxf* z9w|MzfyQHD<~uSlxPj&ak;2mrG@g%?9^63vN2Ks^1N9G~`a$Wz4bQhd3A%4cNvfa-UMdEoTm1{%+S=?A$VG=7ex-W?=| z6#nj@u?Qsd-9hasr25?*6dy?G!5!58K}t{Vp!t5J^xzK4pP(`h8s6@p_9{|*xP$5+ zWb;A$eURM)8s7)ixlsKcpt1twX6X2&2dMuG<%80L2Po}8`JnXR0jjHz{ObW~+alTT z0ZLCu>A?fkzC%h69-#dVNa?`?H2wrL-;sgA1C$?-%?Fh)Na?`?)E`7j4<4ZO0@V*n z4<4ZP2}tb+4^a6Fi9c}q@IXpGp!DDYN#va2gT^*^~s^};R!0gk1xh z&j7xDL#Bb<8w&uRbS9}K2mz{1&!Z>@;FRC zXgmq#A14L|U(otKC?Av_e39!%MFs|6wEEExq!!73KTupDr3XJydP7PNexUItr1ani zsxLuifa)tp1_nP+`GORlexULaDLwds(i6xIsQdju<4sWgp!DE}=nsP9-w(9j2@-!0 zKB)f(RS!xJexUjb#&=?1@B_`)BBcjEP<@5$9#DG@Vjeg>_<_cAVERGs2i0H5>OuQi zKz$;ZdqM3PWb;AuYoI<7OdeEzAjQ8wDDIK`>knG*0uqC&_Xq740Ij)%@j>l3r1zASGLFpj?xxeVdzz_f`uaM#^08~FCy9d;MgO~?S z4*{V4RxtBG?g!O>$m&7m6HA)q~PQ5NLf0jPJz25Cj_kMvAW> z(0ntpdqCscNa-O6xqNYCUq4+ewEU!?d529>{{JP5Tv7_@#5DLn*(>Q|)tA{ebb6%1-`K*I-= z9)gkUM@0sPV6^%%7*szZ+Yd@VNa-OMRNf$^hY(O&0AZ;6LO^bV+6QV+hJe;5A%$lM zDE%O%hY(O&M@kPNp!OG3KPWwffaYf*@dYj)LO}Dc$b8Uz2vj{NJ%k{ycXMK32mzHh zNbwZ{%3sLt0hK2Z^T6dp2&jDnGtZHMAp|u4jI171pCgqIA)xXC*?drWjZ{8_fW}{t z;y(nG-;n(a>aQTBhY-+sHd6m26vRe~k5EwAfmC0Fg4VBq!WV8mq2m_@jr1BsPw7(M)f8g{H1`02zdQf@@1H}c5 z@5I0m268J>e1(D16OwzvK;;p{JaBpl1NEn2`W+bpk?ao#m7hrIAskd5 zfx;MOA833WDLsUP(hJOdM+Sy)PPKG9Z2|r z<3AiUo&2i0##^+g0IK9JHw1Zcbx$-fbx`UfdJM1cB}$ofJ3X{7jw z0JUF1X#nP4P<@W<9?*CtQhJC4r4N{YoER7)LGvL{J}5mzg49C!p!5(4DqE2J8wnck zN3uT>ZGBWEs6T|19wI^G_ekj>614sWX1*f>LnLT?3fX*6e;O%0M53*aibN{EKAJtD zA8mb9BxwBrQuz=G%CAV~M}gR&HWM_yqCnw`lpdl$;fc%#t@lSt4^g1}iqyY|0<}+& z;v)(a9!TwnC{X>5lpdl$^B+j6kpMx zzAuvf(V+SqseFhAr6;8H5DjX-Bc+FEQ2PdEz9R!eG-y2sviYF-6jtyCW^njNgT}Lv z@=G*mz67ctlpdl%^S6-r0;h*)ME?-NN2(7&TdF9a4D3fYJ+6 zdWZqF7eRF()V(oi<#i0Gyhf@|VnFEw$vrWk`~j*HVfsPtK?;u;r11t&dWZpyH^Af- z85m+f}fjtmU3pz;LSd{Fy` z6_FldLF-+R(nBnA{RYwxnooqp7dZZ7LE~4*d{BCV*{8(75DV%L!1zuK46&g44de%y zdqMLp$nF7+XF<#Zr-xWj`hn>O*$)~YMOF`5&j@PMz{~@!FF-aQ)L#RQIYH&)K;em$ z9^ycHk<`b5;ua}A#DV%-pgtN*Kj?e`r1*#f)jvq-Ee@1_Kyd<74{ARm^-tnJ^)1Xl zP7Dljp!Ew-J}5oJA+HZqWMGH`^?#838wX17$o7N!LrCc%4m7@nlpf+iZbwQF@u2Y( zsC}URMLei~fMk9=XuKRLJ;a0NOOVn-JShD@^@GZXcu@Zm5?|o-5DzL}k@=wcXsCKn z`4A5(uV8#928MXh_!d(5$AiizWcPsb8^khANC1^DNd8R#^+%D?LjtJ0L`n|{ zp!rv1_k!k^kF*4b=UKpt={TACw*vL1iZ- z{J`lU5!4<+=7Ywwpz1;CArTaQFuoH5Ln5gBLW-|MwEjXOsJ{R)51bwnk<$~%{b=Lu ziJ9`ja3xz|05DA0fL3G#&vl4;(+q zp!kRBcVu8l29?jq>OtcXNb#8r3V&qtLFX5M+FUUEK=}(PJtTv|57|CY`GJ%kl0oG$ z()eNusJumrj}%Z{fK)!DfZPwtV^I52K<+~-A5xI(KPLu;6y*6r5Fb<@!u+emz>oqO zPebxg3Mjom;y(ej43=9fP3JeSkYzzzxP~02z1SXJs92giFn33e& z!E6u#F(1SSg$c-fkUKo#?sH&>VPIhJM&g6Q!WSB5AoU=AFvut{hS(p1@UMUq02l@l AD*ylh