diff --git a/internal/cgen/base/token-public.h b/internal/cgen/base/token-public.h index 9f48d517..1c4c7c29 100644 --- a/internal/cgen/base/token-public.h +++ b/internal/cgen/base/token-public.h @@ -180,6 +180,11 @@ wuffs_base__make_token(uint64_t repr) { #define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED 0x00002 #define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED 0x00004 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF 0x00010 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF 0x00020 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN 0x00040 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN 0x00080 + // The number 300 might be represented as "\x01\x2C", "\x2C\x01\x00\x00" or // "300", which are big-endian, little-endian or text. For binary formats, the // token length discriminates e.g. u16 little-endian vs u32 little-endian. diff --git a/internal/cgen/data.go b/internal/cgen/data.go index dac8f335..3f35fff0 100644 --- a/internal/cgen/data.go +++ b/internal/cgen/data.go @@ -371,8 +371,8 @@ const baseTokenPublicH = "" + "" + "// --------\n\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__UNDEFINED 0x00001\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__NULL 0x00002\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__FALSE 0x00004\n#define WUFFS_BASE__TOKEN__VBD__LITERAL__TRUE 0x00008\n\n " + "" + - "// --------\n\n// For a source string of \"123\" or \"0x9A\", it is valid for a tokenizer to\n// return any one of:\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED.\n//\n// For a source string of \"+123\" or \"-0x9A\", only the first two are valid.\n//\n// For a source string of \"123.\", only the first one is valid.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT 0x00001\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED 0x00002\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED 0x00004\n\n// The number 300 might be represented as \"\\x01\\x2C\", \"\\x2C\\x01\\x00\\x00\" or\n// \"300\", which are big-endian, little-endian or text. For binary formats, the\n// token length discriminates e.g. u16 little-endian vs u32 little-endian.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN 0x00100\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_END" + - "IAN 0x00200\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT 0x00400\n\n" + + "// --------\n\n// For a source string of \"123\" or \"0x9A\", it is valid for a tokenizer to\n// return any one of:\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED.\n// - WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED.\n//\n// For a source string of \"+123\" or \"-0x9A\", only the first two are valid.\n//\n// For a source string of \"123.\", only the first one is valid.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_FLOATING_POINT 0x00001\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED 0x00002\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED 0x00004\n\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF 0x00010\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF 0x00020\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN 0x00040\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN 0x00080\n\n// The number 300 might be represented as \"\\x01\\x2C\", \"\\x2C\\x01\\x00\\x00\" or\n// \"300\", which are big-endian, li" + + "ttle-endian or text. For binary formats, the\n// token length discriminates e.g. u16 little-endian vs u32 little-endian.\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_BIG_ENDIAN 0x00100\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_BINARY_LITTLE_ENDIAN 0x00200\n#define WUFFS_BASE__TOKEN__VBD__NUMBER__FORMAT_TEXT 0x00400\n\n" + "" + "// --------\n\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 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" + diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index b615ecc3..d242f9a7 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -1939,6 +1939,11 @@ wuffs_base__make_token(uint64_t repr) { #define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_SIGNED 0x00002 #define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_INTEGER_UNSIGNED 0x00004 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_INF 0x00010 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_INF 0x00020 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_NEG_NAN 0x00040 +#define WUFFS_BASE__TOKEN__VBD__NUMBER__CONTENT_POS_NAN 0x00080 + // The number 300 might be represented as "\x01\x2C", "\x2C\x01\x00\x00" or // "300", which are big-endian, little-endian or text. For binary formats, the // token length discriminates e.g. u16 little-endian vs u32 little-endian. @@ -6023,6 +6028,7 @@ struct wuffs_json__decoder__struct { uint32_t p_decode_tokens[1]; uint32_t p_decode_leading[1]; uint32_t p_decode_comment[1]; + uint32_t p_decode_inf_nan[1]; uint32_t p_decode_trailing_new_line[1]; } private_impl; @@ -6037,6 +6043,9 @@ struct wuffs_json__decoder__struct { struct { uint32_t v_link_prev; } s_decode_comment[1]; + struct { + uint32_t v_neg; + } s_decode_inf_nan[1]; } private_data; #ifdef __cplusplus @@ -20295,10 +20304,10 @@ static const uint8_t // WUFFS_BASE__POTENTIALLY_UNUSED = { 15, 15, 15, 15, 15, 15, 15, 15, 15, 0, 0, 15, 15, 0, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 0, 15, 1, 15, - 15, 15, 15, 15, 15, 15, 15, 15, 2, 4, 15, 12, 4, 4, 4, 4, 4, 4, + 15, 15, 15, 15, 15, 15, 15, 11, 2, 4, 15, 12, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, - 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, - 15, 7, 15, 8, 15, 15, 15, 15, 15, 15, 15, 15, 9, 15, 15, 15, 15, 15, + 15, 11, 15, 15, 15, 15, 11, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, + 15, 7, 15, 8, 15, 15, 15, 15, 15, 15, 15, 15, 9, 15, 15, 11, 15, 15, 15, 15, 11, 15, 15, 15, 15, 15, 10, 15, 15, 15, 15, 15, 15, 5, 15, 6, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, @@ -20375,6 +20384,11 @@ wuffs_json__decoder__decode_comment(wuffs_json__decoder* self, wuffs_base__token_buffer* a_dst, wuffs_base__io_buffer* a_src); +static wuffs_base__status // +wuffs_json__decoder__decode_inf_nan(wuffs_json__decoder* self, + wuffs_base__token_buffer* a_dst, + wuffs_base__io_buffer* a_src); + static wuffs_base__status // wuffs_json__decoder__decode_trailing_new_line(wuffs_json__decoder* self, wuffs_base__token_buffer* a_dst, @@ -21374,6 +21388,27 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, } } if (v_number_status == 1) { + if (self->private_impl.f_quirk_enabled_allow_inf_nan_numbers) { + if (a_dst) { + a_dst->meta.wi = ((size_t)(iop_a_dst - a_dst->data.ptr)); + } + if (a_src) { + a_src->meta.ri = ((size_t)(iop_a_src - a_src->data.ptr)); + } + WUFFS_BASE__COROUTINE_SUSPENSION_POINT(16); + status = + wuffs_json__decoder__decode_inf_nan(self, a_dst, a_src); + if (a_dst) { + iop_a_dst = a_dst->data.ptr + a_dst->meta.wi; + } + if (a_src) { + iop_a_src = a_src->data.ptr + a_src->meta.ri; + } + if (status.repr) { + goto suspend; + } + goto label__3__break; + } status = wuffs_base__make_status(wuffs_json__error__bad_input); goto exit; } else if (v_number_status == 2) { @@ -21383,11 +21418,11 @@ 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(16); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(17); 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(17); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(18); } } } @@ -21521,7 +21556,7 @@ 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(19); goto label__outer__continue; } } else if (v_class == 10) { @@ -21542,7 +21577,7 @@ 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(20); goto label__outer__continue; } } else if (v_class == 11) { @@ -21563,9 +21598,29 @@ 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(20); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(21); goto label__outer__continue; } + if (self->private_impl.f_quirk_enabled_allow_inf_nan_numbers) { + if (a_dst) { + a_dst->meta.wi = ((size_t)(iop_a_dst - a_dst->data.ptr)); + } + if (a_src) { + a_src->meta.ri = ((size_t)(iop_a_src - a_src->data.ptr)); + } + WUFFS_BASE__COROUTINE_SUSPENSION_POINT(22); + status = wuffs_json__decoder__decode_inf_nan(self, a_dst, a_src); + if (a_dst) { + iop_a_dst = a_dst->data.ptr + a_dst->meta.wi; + } + if (a_src) { + iop_a_src = a_src->data.ptr + a_src->meta.ri; + } + if (status.repr) { + goto suspend; + } + goto label__goto_parsed_a_leaf_value__break; + } } else if (v_class == 12) { if (self->private_impl.f_quirk_enabled_allow_comment_block || self->private_impl.f_quirk_enabled_allow_comment_line) { @@ -21575,7 +21630,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, if (a_src) { a_src->meta.ri = ((size_t)(iop_a_src - a_src->data.ptr)); } - WUFFS_BASE__COROUTINE_SUSPENSION_POINT(21); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT(23); status = wuffs_json__decoder__decode_comment(self, a_dst, a_src); if (a_dst) { iop_a_dst = a_dst->data.ptr + a_dst->meta.wi; @@ -21606,7 +21661,7 @@ wuffs_json__decoder__decode_tokens(wuffs_json__decoder* self, if (a_src) { a_src->meta.ri = ((size_t)(iop_a_src - a_src->data.ptr)); } - WUFFS_BASE__COROUTINE_SUSPENSION_POINT(22); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT(24); status = wuffs_json__decoder__decode_trailing_new_line(self, a_dst, a_src); if (a_dst) { @@ -22155,6 +22210,174 @@ exit: return status; } +// -------- func json.decoder.decode_inf_nan + +static wuffs_base__status // +wuffs_json__decoder__decode_inf_nan(wuffs_json__decoder* self, + wuffs_base__token_buffer* a_dst, + wuffs_base__io_buffer* a_src) { + wuffs_base__status status = wuffs_base__make_status(NULL); + + uint32_t v_c4 = 0; + uint32_t v_neg = 0; + + wuffs_base__token* iop_a_dst = NULL; + wuffs_base__token* io0_a_dst WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + wuffs_base__token* io1_a_dst WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + wuffs_base__token* io2_a_dst WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + if (a_dst) { + io0_a_dst = a_dst->data.ptr; + io1_a_dst = io0_a_dst + a_dst->meta.wi; + iop_a_dst = io1_a_dst; + io2_a_dst = io0_a_dst + a_dst->data.len; + if (a_dst->meta.closed) { + io2_a_dst = iop_a_dst; + } + } + uint8_t* iop_a_src = NULL; + uint8_t* io0_a_src WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + uint8_t* io1_a_src WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + uint8_t* io2_a_src WUFFS_BASE__POTENTIALLY_UNUSED = NULL; + if (a_src) { + io0_a_src = a_src->data.ptr; + io1_a_src = io0_a_src + a_src->meta.ri; + iop_a_src = io1_a_src; + io2_a_src = io0_a_src + a_src->meta.wi; + } + + uint32_t coro_susp_point = self->private_impl.p_decode_inf_nan[0]; + if (coro_susp_point) { + v_neg = self->private_data.s_decode_inf_nan[0].v_neg; + } + switch (coro_susp_point) { + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_0; + + label__0__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(1); + goto label__0__continue; + } + if (((uint64_t)(io2_a_src - iop_a_src)) <= 2) { + 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(2); + goto label__0__continue; + } + v_c4 = ((uint32_t)(wuffs_base__load_u24le__no_bounds_check(iop_a_src))); + if ((v_c4 | 2105376) == 6712937) { + if (((uint64_t)(io2_a_src - iop_a_src)) > 7) { + if ((wuffs_base__load_u64le__no_bounds_check(iop_a_src) | + 2314885530818453536) == 8751735898823356009) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)(10485792)) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(8)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + (iop_a_src += 8, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(NULL); + goto ok; + } + } else if (!(a_src && a_src->meta.closed)) { + status = wuffs_base__make_status(wuffs_base__suspension__short_read); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(3); + goto label__0__continue; + } + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)(10485792)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + (iop_a_src += 3, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(NULL); + goto ok; + } else if ((v_c4 | 2105376) == 7233902) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)(10485888)) << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(3)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + (iop_a_src += 3, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(NULL); + goto ok; + } else if ((v_c4 & 255) == 43) { + v_neg = 0; + } else if ((v_c4 & 255) == 45) { + v_neg = 1; + } else { + status = wuffs_base__make_status(wuffs_json__error__bad_input); + goto exit; + } + if (((uint64_t)(io2_a_src - iop_a_src)) <= 3) { + 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(4); + goto label__0__continue; + } + v_c4 = (wuffs_base__load_u32le__no_bounds_check(iop_a_src) >> 8); + if ((v_c4 | 2105376) == 6712937) { + if (((uint64_t)(io2_a_src - iop_a_src)) > 8) { + if ((wuffs_base__load_u64le__no_bounds_check(iop_a_src + 1) | + 2314885530818453536) == 8751735898823356009) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)((10485760 | (((uint32_t)(32)) >> v_neg)))) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(9)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + (iop_a_src += 9, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(NULL); + goto ok; + } + } else if (!(a_src && a_src->meta.closed)) { + status = wuffs_base__make_status(wuffs_base__suspension__short_read); + WUFFS_BASE__COROUTINE_SUSPENSION_POINT_MAYBE_SUSPEND(5); + goto label__0__continue; + } + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)((10485760 | (((uint32_t)(32)) >> v_neg)))) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(4)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + (iop_a_src += 4, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(NULL); + goto ok; + } else if ((v_c4 | 2105376) == 7233902) { + *iop_a_dst++ = wuffs_base__make_token( + (((uint64_t)((10485760 | (((uint32_t)(128)) >> v_neg)))) + << WUFFS_BASE__TOKEN__VALUE_MINOR__SHIFT) | + (((uint64_t)(4)) << WUFFS_BASE__TOKEN__LENGTH__SHIFT)); + (iop_a_src += 4, wuffs_base__make_empty_struct()); + status = wuffs_base__make_status(NULL); + goto ok; + } + status = wuffs_base__make_status(wuffs_json__error__bad_input); + goto exit; + } + + goto ok; + ok: + self->private_impl.p_decode_inf_nan[0] = 0; + goto exit; + } + + goto suspend; +suspend: + self->private_impl.p_decode_inf_nan[0] = + wuffs_base__status__is_suspension(&status) ? coro_susp_point : 0; + self->private_data.s_decode_inf_nan[0].v_neg = v_neg; + + goto exit; +exit: + if (a_dst) { + a_dst->meta.wi = ((size_t)(iop_a_dst - a_dst->data.ptr)); + } + if (a_src) { + a_src->meta.ri = ((size_t)(iop_a_src - a_src->data.ptr)); + } + + return status; +} + // -------- func json.decoder.decode_trailing_new_line static wuffs_base__status // diff --git a/std/json/common_consts.wuffs b/std/json/common_consts.wuffs index 8a43b711..f92383de 100644 --- a/std/json/common_consts.wuffs +++ b/std/json/common_consts.wuffs @@ -196,7 +196,7 @@ pri const lut_chars array[256] base.u8 = [ // - 0x08 (bitmask 0x0100) is CLASS_CLOSE_SQUARE_BRACKET. // - 0x09 (bitmask 0x0200) is CLASS_FALSE. // - 0x0A (bitmask 0x0400) is CLASS_TRUE. -// - 0x0B (bitmask 0x0800) is CLASS_NULL. +// - 0x0B (bitmask 0x0800) is CLASS_NULL_NAN_INF. // - 0x0C (bitmask 0x1000) is CLASS_COMMENT. // - 0x0D (bitmask 0x2000) is reserved. // - 0x0E (bitmask 0x4000) is reserved. @@ -236,16 +236,16 @@ pri const lut_classes array[256] base.u8[..= 0x0F] = [ 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x10 ..= 0x17. 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x18 ..= 0x1F. 0x00, 0x0F, 0x01, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x20 ..= 0x27. ' ', '"'. - 0x0F, 0x0F, 0x0F, 0x0F, 0x02, 0x04, 0x0F, 0x0C, // 0x28 ..= 0x2F. ',', '-', '/'. + 0x0F, 0x0F, 0x0F, 0x0B, 0x02, 0x04, 0x0F, 0x0C, // 0x28 ..= 0x2F. '+', ',', '-', '/'. 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, // 0x30 ..= 0x37. '0'-'7'. 0x04, 0x04, 0x03, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x38 ..= 0x3F. '8'-'9', ':'. 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x40 ..= 0x47. - 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x48 ..= 0x4F. + 0x0F, 0x0B, 0x0F, 0x0F, 0x0F, 0x0F, 0x0B, 0x0F, // 0x48 ..= 0x4F. 'I', 'N'. 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, // 0x50 ..= 0x57. 0x0F, 0x0F, 0x0F, 0x07, 0x0F, 0x08, 0x0F, 0x0F, // 0x58 ..= 0x5F. '[', ']'. 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x09, 0x0F, // 0x60 ..= 0x67. 'f'. - 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0F, 0x0B, 0x0F, // 0x68 ..= 0x6F. 'n'. + 0x0F, 0x0B, 0x0F, 0x0F, 0x0F, 0x0F, 0x0B, 0x0F, // 0x68 ..= 0x6F. 'i', 'n'. 0x0F, 0x0F, 0x0F, 0x0F, 0x0A, 0x0F, 0x0F, 0x0F, // 0x70 ..= 0x77. 't'. 0x0F, 0x0F, 0x0F, 0x05, 0x0F, 0x06, 0x0F, 0x0F, // 0x78 ..= 0x7F. '{', '}'. diff --git a/std/json/decode_json.wuffs b/std/json/decode_json.wuffs index 566811d6..3d99554a 100644 --- a/std/json/decode_json.wuffs +++ b/std/json/decode_json.wuffs @@ -888,6 +888,10 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { } if number_status == 1 { + if this.quirk_enabled_allow_inf_nan_numbers { + this.decode_inf_nan?(dst: args.dst, src: args.src) + break + } return "#bad input" } else if number_status == 2 { return "#unsupported number length" @@ -1061,7 +1065,7 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { continue.outer } - } else if class == 0x0B { // 0x0B is CLASS_NULL. + } else if class == 0x0B { // 0x0B is CLASS_NULL_NAN_INF. match = args.src.match7(a: 0x6C_6C75_6E04) // 4 bytes "null". if match == 0 { args.dst.write_fast_token!( @@ -1079,6 +1083,11 @@ pub func decoder.decode_tokens?(dst: base.token_writer, src: base.io_reader) { continue.outer } + if this.quirk_enabled_allow_inf_nan_numbers { + this.decode_inf_nan?(dst: args.dst, src: args.src) + break.goto_parsed_a_leaf_value + } + } else if class == 0x0C { // 0x0C is CLASS_COMMENT. if this.quirk_enabled_allow_comment_block or this.quirk_enabled_allow_comment_line { this.decode_comment?(dst: args.dst, src: args.src) @@ -1463,6 +1472,111 @@ pri func decoder.decode_comment?(dst: base.token_writer, src: base.io_reader) { return "#bad input" } +pri func decoder.decode_inf_nan?(dst: base.token_writer, src: base.io_reader) { + var c4 : base.u32 + var neg : base.u32[..= 1] + + while true { + if args.dst.available() <= 0 { + yield? base."$short write" + continue + } + if args.src.available() <= 2 { + if args.src.is_closed() { + return "#bad input" + } + yield? base."$short read" + continue + } + + // Bitwise or'ing with 0x20 converts upper case ASCII to lower case. + + c4 = args.src.peek_u24le_as_u32() + if (c4 | 0x20_2020) == 0x66_6E69 { + if args.src.available() > 7 { + if (args.src.peek_u64le() | 0x2020_2020_2020_2020) == 0x7974_696E_6966_6E69 { + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0xA0_0020, + link: 0x0, + length: 8) + args.src.skip32_fast!(actual: 8, worst_case: 8) + return ok + } + } else if not args.src.is_closed() { + yield? base."$short read" + continue + } + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0xA0_0020, + link: 0x0, + length: 3) + args.src.skip32_fast!(actual: 3, worst_case: 3) + return ok + + } else if (c4 | 0x20_2020) == 0x6E_616E { + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0xA0_0080, + link: 0x0, + length: 3) + args.src.skip32_fast!(actual: 3, worst_case: 3) + return ok + } else if (c4 & 0xFF) == 0x2B { // 0x2B is '+'. + neg = 0 + } else if (c4 & 0xFF) == 0x2D { // 0x2D is '-'. + neg = 1 + } else { + return "#bad input" + } + + if args.src.available() <= 3 { + if args.src.is_closed() { + return "#bad input" + } + yield? base."$short read" + continue + } + + c4 = args.src.peek_u32le() >> 8 + if (c4 | 0x20_2020) == 0x66_6E69 { + if args.src.available() > 8 { + if (args.src.peek_u64le_at(offset: 1) | 0x2020_2020_2020_2020) == 0x7974_696E_6966_6E69 { + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0xA0_0000 | ((0x20 as base.u32) >> neg), + link: 0x0, + length: 9) + args.src.skip32_fast!(actual: 9, worst_case: 9) + return ok + } + } else if not args.src.is_closed() { + yield? base."$short read" + continue + } + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0xA0_0000 | ((0x20 as base.u32) >> neg), + link: 0x0, + length: 4) + args.src.skip32_fast!(actual: 4, worst_case: 4) + return ok + + } else if (c4 | 0x20_2020) == 0x6E_616E { + args.dst.write_fast_token!( + value_major: 0, + value_minor: 0xA0_0000 | ((0x80 as base.u32) >> neg), + link: 0x0, + length: 4) + args.src.skip32_fast!(actual: 4, worst_case: 4) + return ok + } + + return "#bad input" + } +} + pri func decoder.decode_trailing_new_line?(dst: base.token_writer, src: base.io_reader) { var c : base.u8 var whitespace_length : base.u32[..= 0xFFFE] diff --git a/test/c/std/json.c b/test/c/std/json.c index f1bfc1ab..3cce8fd5 100644 --- a/test/c/std/json.c +++ b/test/c/std/json.c @@ -1397,6 +1397,82 @@ test_wuffs_json_decode_quirk_allow_final_comma() { return NULL; } +const char* // +test_wuffs_json_decode_quirk_allow_inf_nan_numbers() { + CHECK_FOCUS(__func__); + + struct { + // want has 2 bytes, one for each possible q: + // - q&1 sets WUFFS_JSON__QUIRK_ALLOW_INF_NAN_NUMBERS. + // An 'X', '+' or '-' means that decoding should succeed (and consume the + // entire input), succeed (without consuming the entire input) or fail. + const char* want; + const char* str; + } test_cases[] = { + {.want = "-X", .str = "InFiniTy"}, + {.want = "-X", .str = "[+inf, -infinity, +nan,-NaN,NAN]"}, + {.want = "-X", .str = "inf"}, + {.want = "-+", .str = "infinit"}, + {.want = "-+", .str = "infiQity"}, + {.want = "-+", .str = "nana"}, + {.want = "--", .str = "+-inf"}, + {.want = "--", .str = "-+inf"}, + {.want = "--", .str = "[infinit,"}, + {.want = "--", .str = "[infiQity,"}, + {.want = "--", .str = "[nana,"}, + {.want = "--", .str = "∞"}, + }; + + int tc; + for (tc = 0; tc < WUFFS_TESTLIB_ARRAY_SIZE(test_cases); tc++) { + int q; + for (q = 0; q < 2; q++) { + wuffs_json__decoder dec; + CHECK_STATUS("initialize", wuffs_json__decoder__initialize( + &dec, sizeof dec, WUFFS_VERSION, + WUFFS_INITIALIZE__DEFAULT_OPTIONS)); + wuffs_json__decoder__set_quirk_enabled( + &dec, WUFFS_JSON__QUIRK_ALLOW_INF_NAN_NUMBERS, q & 1); + + wuffs_base__token_buffer tok = + wuffs_base__make_token_buffer_writer(global_have_token_slice); + wuffs_base__io_buffer src = wuffs_base__make_io_buffer_reader( + wuffs_base__make_slice_u8((void*)(test_cases[tc].str), + strlen(test_cases[tc].str)), + true); + const char* have = + wuffs_json__decoder__decode_tokens(&dec, &tok, &src).repr; + const char* want = + (test_cases[tc].want[q] != '-') ? NULL : wuffs_json__error__bad_input; + if (have != want) { + RETURN_FAIL("tc=%d, q=%d: decode_tokens: have \"%s\", want \"%s\"", tc, + q, have, want); + } + + size_t total_length = 0; + while (tok.meta.ri < tok.meta.wi) { + total_length += wuffs_base__token__length(&tok.data.ptr[tok.meta.ri++]); + } + if (total_length != src.meta.ri) { + RETURN_FAIL("tc=%d, q=%d: total_length: have %zu, want %zu", tc, q, + total_length, src.meta.ri); + } + if (test_cases[tc].want[q] == 'X') { + if (total_length != src.data.len) { + RETURN_FAIL("tc=%d, q=%d: total_length: have %zu, want %zu", tc, q, + total_length, src.data.len); + } + } else if (test_cases[tc].want[q] == '+') { + if (total_length >= src.data.len) { + RETURN_FAIL("tc=%d, q=%d: total_length: have %zu, want < %zu", tc, q, + total_length, src.data.len); + } + } + } + } + return NULL; +} + const char* // test_wuffs_json_decode_quirk_allow_comment_etc() { CHECK_FOCUS(__func__); @@ -2115,6 +2191,7 @@ proc tests[] = { test_wuffs_json_decode_quirk_allow_backslash_x, test_wuffs_json_decode_quirk_allow_comment_etc, test_wuffs_json_decode_quirk_allow_final_comma, + test_wuffs_json_decode_quirk_allow_inf_nan_numbers, test_wuffs_json_decode_quirk_allow_leading_etc, test_wuffs_json_decode_quirk_allow_trailing_etc, test_wuffs_json_decode_quirk_replace_invalid_utf_8,