Documentation
¶
Overview ¶
Package textcoder lets the v2 json package give a jsontext.Encoder to a method or a function of marshaling, which writes after the output of the encoder of the v2 semantics: jsontext sets these functions, which take its encoder as an any, since this package can't import jsontext.
Index ¶
Constants ¶
const InputRoom = 64
InputRoom is the room which a decoder of a reader leaves after the input in its buffer, for a stream of the decoders of v2 which reads the buffer in place ( see DecoderBuffer ): the padding which the stream keeps after the input in its own buffer, for the nul byte after the input and a block of 64 bytes which a scan reads from any byte of the input, past it from the last bytes.
Variables ¶
var ( // Attach makes enc, a *jsontext.Encoder, write after the output of the place p by its options ( see // Attachment ). // enc is the pointer itself, without the check of a type assertion, which the calls for every value of a // method don't take; the place is given by a pointer, as its fields don't fit the registers of a call. Attach func(enc unsafe.Pointer, p *Attachment) // Detach sets the output of enc, which Attach set up, and its depth and the count of the tokens of its // innermost level in p. Detach func(enc unsafe.Pointer, p *Attachment) // Position returns the levels of enc, a *jsontext.Encoder, the top level first, and the offset in its output // where its next value starts: after the delimiter and the white space before it. Position func(enc any, levels []Level) ([]Level, int64) // Place returns the innermost level of enc, whose count may be the one which Attach was given ( see Attach ), // whether it is the top level, the offset in its output where its next value starts, the Outer which it was // attached with, if it was, and the number of the arrays and the objects open around its next value in the whole // output ( see Attachment.OuterDepth ). Place func(enc any) (inner Level, top bool, offset int64, outer Outer, depth int) // Configure makes enc write by the options until restore is called, which gives it back its own, with the // strings of the values as they are, which the calls which wrote them escaped. Configure func(enc any, opts options.Options) (restore func()) // Invalidate makes enc fail every write until a Reset: MarshalEncode stopped within an object or an array it // opened. Invalidate func(enc any) // Fail records err in enc, an encoder which Attach attached, the first error of a call of MarshalEncode which // failed at it: the value which enc was given the place of fails with it, whatever its method returns. Fail func(enc any, err error) )
var ( // AttachDecoder makes dec, a *jsontext.Decoder, read the input of the place p by its options ( see // DecoderAttachment ). AttachDecoder func(dec unsafe.Pointer, p *DecoderAttachment) // DetachDecoder sets where dec, which AttachDecoder set up, stopped reading in p, and drops its input: a // decoder kept past its call reads nothing more. DetachDecoder func(dec unsafe.Pointer, p *DecoderAttachment) // DecoderPointer returns the JSON pointer of the place of dec, which DetachDecoder detached, which where points // to: the last value which was read (-1), the next value (+1), or the level it is in (0). DecoderPointer func(dec unsafe.Pointer, where int) string // ReleaseDecoder drops the options of dec, which it was attached with, and its outer levels: the decoder, which // is kept for the next call, must not keep the functions of the caller. ReleaseDecoder func(dec unsafe.Pointer) )
var ( // DecoderPlace returns the levels of the whole input around the next value of dec, the top level first, appended // to levels ( see Level ), whose names are copied into the arrays of the names of the entries of levels after its // length, which are reused. DecoderPlace func(dec unsafe.Pointer, levels []Level) []Level // DecoderAt returns the rest of the place of the next value of dec: whether it is a name of an object, the Outer // of the attach of dec, or nil, and the number of the arrays and the objects open around it. DecoderAt func(dec unsafe.Pointer) (bool, Outer, int) // BeginValue reads what dec reads before its next value, which is not a name, as ReadValue reads it: the // white space and the delimiter, which it checks, and returns the offset of the value in the whole input, or // the error which ReadValue returns there, io.EOF after the last value of the input. The value isn't read. BeginValue func(dec unsafe.Pointer) (int64, error) // DecoderInput returns the input of dec from the offset start of the whole input, which its buffer has: the // bytes of its buffer, and its reader, which reads the input after them. DecoderInput func(dec unsafe.Pointer, start int64) ([]byte, io.Reader) // DecoderBuffer returns the buffer of dec, which its input up to len(buf) has been read into, and the offset of // its first byte in the whole input. It reports false if a caller can't read the buffer in place and read more // into it ( see FetchInput ): the array of a bytes.Buffer, of a decoder of a bytes.Buffer, which the decoder // doesn't own, or a buffer without InputRoom bytes of room after its input, which a decoder of a reader leaves // it. DecoderBuffer func(dec unsafe.Pointer) ([]byte, int64, bool) // FetchInput reads more of the input of dec into its buffer, which DecoderBuffer gave a caller, after what has // been read, n bytes at most, by one read of its reader, keeping the bytes from the offset keep of the whole // input, which the buffer may move to its start, and returns the buffer as DecoderBuffer does, the number of // the bytes read and the error of the read, which dec doesn't keep. The state of dec is the caller's until it // ends the value by EndValue, with the rest, or PutBack. FetchInput func(dec unsafe.Pointer, keep int64, n int) ([]byte, int64, int, error) // EndValue makes dec go on after the next value, which was read from start to end, offsets of the whole // input: in its buffer, which may have dropped the start of the value for a caller which read more into it // ( see FetchInput ), or, if rest is set, from rest, the input after the value which was read from its reader, // which it copies. The value is the last value which dec read, from start, whose place a decoder attached at a // value reports to the caller ( see DetachDecoder ). EndValue func(dec unsafe.Pointer, start, end int64, rest []byte) // PutBack makes dec read in, the input from the offset start of the whole input, which a caller read from its // reader, as the next input: dec is where it was before its next value, which the caller didn't decode. PutBack func(dec unsafe.Pointer, start int64, in []byte) // FailDecoder makes every later read of dec fail with err, which the read returns as it is. FailDecoder func(dec unsafe.Pointer, err error) // DecoderFailed reports whether FailDecoder made every read of dec fail. DecoderFailed func(dec unsafe.Pointer) bool // SkipToEnd reads the input of the place p as UnmarshalPlaceAt reads it, up to the end of the value of the // level of the place at depth, which has count tokens before the value: the levels deeper are read to their // ends. It returns the offset of the end of the value, and what it read from the reader of p and didn't use, // or the error of the input. SkipToEnd func(p *DecoderAttachment, depth int, count int64) (int64, []byte, error) )
The decoding of the next value of a *jsontext.Decoder by the v2 json package ( UnmarshalDecode ), which reads the value itself from the input of the decoder, and then makes the decoder go on after it.
var ( ReadError func(err error) error IsIOError func(err error) bool )
ReadError returns err, the error of a reader of the input, as jsontext reports it, and IsIOError reports whether err is the error of a reader or of a writer, which a method passes on as it is.
var ( TakeSkipper func(allowDuplicateNames, allowInvalidUTF8 bool) unsafe.Pointer ReleaseSkipper func(skipper unsafe.Pointer) ValidValue func(skipper unsafe.Pointer, v []byte) bool )
A skipper is a scanner of jsontext which reads the values that a call of the v2 json package skips, by the options of the names of an object twice and of invalid UTF-8, without writing them: TakeSkipper returns one, which the call keeps for all its values, and ReleaseSkipper puts it back.
ValidValue reports whether v, a JSON value valid by the grammar, has the text which jsontext reads by the options of the skipper: valid UTF-8 unless invalid UTF-8 is allowed, and no object with a name twice unless that is.
AfterTopLevelError returns the syntax error of the first byte of in, at the offset base of the input, which follows the top-level value, as jsontext reports it.
var ConfigureDecoder func(dec unsafe.Pointer, opts *options.Config) (restore func())
ConfigureDecoder makes the Decoder at dec read by the options of opts, which are its own and the ones of a call, until restore is called, which gives it back its own.
var UnmarshalPlace func(in []byte, opts *options.Config, stop int64, read Read) (string, error)
UnmarshalPlace reads in, the input of an unmarshaling of the v2 json package, by the options, as encoding/json/v2 reads it, up to the value which starts at stop, which it reads as read says: it returns the syntax error it meets first, a *jsontext.SyntacticError, or else the JSON pointer of the value. With stop -1, it reads the whole input and returns its first syntax error, or nil if it has none.
var UnmarshalPlaceAt func(p *DecoderAttachment, stop int64, read Read) (string, error)
UnmarshalPlaceAt is UnmarshalPlace for a part of the input, p.In, which is read from the place p as a decoder attached there reads it ( see DecoderAttachment ), with the offsets of the whole input: it reads up to the value which starts at the offset stop, or, with stop -1, up to the first syntax error, and returns "" and nil if In ends before either. The levels of the place may end in the part, as no method reads it.
Functions ¶
This section is empty.
Types ¶
type Attachment ¶
type Attachment struct {
// Out is the output before the place, at the offset Base of the whole output, or, if Out is empty, Base is
// the offset of the value after its delimiter: the delimiter which the encoder writes before the value, whose
// length Attach sets in Skip, is not a part of the output then, which the caller writes to another encoder.
Out []byte
Base int64
// Object and Count are the innermost level of the place, an open object or array, or the top level if Top is
// set: Count has the parity of the place, a value or a name, and is 0 only if the level has no token yet. The
// levels around it, and its name and count, Outer gives when the encoder needs them.
Object bool
Top bool
Count int64
Outer Outer
// Opts are the options of the encoder, which are the ones it was attached with the last time if Same is set.
Opts *options.Config
Same bool
// Skip, and Depth and Count of the innermost level, which Detach sets after the writes, Attach sets.
Skip int
Depth int
// OuterDepth is the number of the arrays and the objects which are open around the place in the whole output,
// which the encoder counts its levels after: it fails to begin one past the deepest level, as the encoder of the
// whole output does.
OuterDepth int
// Failed is the error of a call which failed at the encoder, which Detach sets ( see Fail ).
Failed error
}
Attachment is the place of the output where an encoder is attached ( see Attach ), and, after Attach and Detach, where it is.
type DecoderAttachment ¶
type DecoderAttachment struct {
// In is the input which the decoder reads from Pos: the end of the token before the value, so that it reads the
// delimiter before the value as its own. In[0] is at the offset Base of the whole input, which the offsets of
// the decoder count from. R is the reader of the input after In, which the decoder reads when it needs more, or
// nil if In has all of it: the whole input of the call, or of the value. If Own is set, In is a buffer of the
// caller, which the decoder moves its input in and reads into, rather than one it makes, and which is In after
// DetachDecoder, grown if the decoder needed more room, for the next attach.
In []byte
Pos int
Base int64
R io.Reader
Own bool
// Object and Count are the innermost level of the place, an open object or array, or the top level if Top is
// set: Count has the parity of the place, which is a value, and is 0 only if the level has no token before
// it. The levels around it, and its name and its count, Outer gives when the decoder needs them.
Object bool
Top bool
Count int64
Outer Outer
// OuterDepth is the number of the arrays and the objects which are open around the place in the whole input,
// which the decoder counts its levels after: it fails to begin one past the deepest level, as the decoder of
// the whole input does.
OuterDepth int
// Opts are the options of the decoder, which are the ones it was attached with the last time if Same is set.
Opts *options.Config
Same bool
// Depth and Count of the innermost level, Pos, End, the offset in the whole input where the decoder stopped
// reading, and PrevStart, the one of the start of the last token or value which was read, DetachDecoder sets
// after the reads; Depth is the one of the place after AttachDecoder. If the decoder read R, Rest is what it
// read and didn't use, which follows End, and Pos is not a place of In.
Depth int
End int64
PrevStart int64
Rest []byte
}
DecoderAttachment is the place of the input where a decoder is attached ( see AttachDecoder ), and, after AttachDecoder and DetachDecoder, where it stopped reading.
type Level ¶
type Level struct {
Object bool
// Count is the number of the tokens of the level written before the value of the next level, which is open,
// or, at the innermost level, before the next token: the elements of an array, the names and the values of an
// object.
Count int64
// Name is the last name of an object, unquoted: of the value which is open or next, if Count is odd. The names
// before it are not kept: a method writes a single value, which is checked for duplicate names in itself, and
// the place of a value is the last name of each object. It is nil where the level has no name, and empty but
// not nil for the name "".
Name []byte
}
Level is a level of the output: the top level, or an open object or array.
type Outer ¶
type Outer interface {
Levels() []Level
}
Outer gives the levels of the place where an encoder was attached, the top level first ( see Attach ): they are found only when the encoder needs them, for a JSON pointer or the stack, which most writes don't.
Source Files