copilot-swe-agent[bot] d936a45bfd Fix SSE specification compliance - add space after colon in all SSE fields
Co-authored-by: appleboy <21979+appleboy@users.noreply.github.com>
2025-11-15 15:22:08 +00:00

389 lines
11 KiB
Go

/**
* Copyright 2025 ByteDance Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package resolver
import (
"reflect"
"sort"
"strings"
"unicode"
"unicode/utf8"
"github.com/bytedance/sonic/internal/encoder/alg"
)
type StdField struct {
name string
nameBytes []byte
nameNonEsc string
nameEscHTML string
tag bool
index []int
typ reflect.Type
omitEmpty bool
omitZero bool
isZero func(reflect.Value) bool
quoted bool
}
type StdStructFields struct {
list []StdField
nameIndex map[string]*StdField
byFoldedName map[string]*StdField
}
func typeFields(t reflect.Type) StdStructFields {
// Anonymous fields to explore at the current level and the next.
current := []StdField{}
next := []StdField{{typ: t}}
// Count of queued names for current level and the next.
var count, nextCount map[reflect.Type]int
// Types already visited at an earlier level.
visited := map[reflect.Type]bool{}
// Fields found.
var fields []StdField
// Buffer to run appendHTMLEscape on field names.
var nameEscBuf []byte
for len(next) > 0 {
current, next = next, current[:0]
count, nextCount = nextCount, map[reflect.Type]int{}
for _, f := range current {
if visited[f.typ] {
continue
}
visited[f.typ] = true
// Scan f.typ for fields to include.
for i := 0; i < f.typ.NumField(); i++ {
sf := f.typ.Field(i)
if sf.Anonymous {
t := sf.Type
if t.Kind() == reflect.Pointer {
t = t.Elem()
}
if !sf.IsExported() && t.Kind() != reflect.Struct {
// Ignore embedded fields of unexported non-struct types.
continue
}
// Do not ignore embedded fields of unexported struct types
// since they may have exported fields.
} else if !sf.IsExported() {
// Ignore unexported non-embedded fields.
continue
}
tag := sf.Tag.Get("json")
if tag == "-" {
continue
}
name, opts := parseTag(tag)
if !isValidTag(name) {
name = ""
}
index := make([]int, len(f.index)+1)
copy(index, f.index)
index[len(f.index)] = i
ft := sf.Type
if ft.Name() == "" && ft.Kind() == reflect.Pointer {
// Follow pointer.
ft = ft.Elem()
}
// Only strings, floats, integers, and booleans can be quoted.
quoted := false
if opts.Contains("string") {
switch ft.Kind() {
case reflect.Bool,
reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64,
reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr,
reflect.Float32, reflect.Float64,
reflect.String:
quoted = true
}
}
// Record found field and index sequence.
if name != "" || !sf.Anonymous || ft.Kind() != reflect.Struct {
tagged := name != ""
if name == "" {
name = sf.Name
}
field := StdField{
name: name,
tag: tagged,
index: index,
typ: ft,
omitEmpty: opts.Contains("omitempty"),
omitZero: opts.Contains("omitzero"),
quoted: quoted,
}
field.nameBytes = []byte(field.name)
// Build nameEscHTML and nameNonEsc ahead of time.
nameEscBuf = alg.HtmlEscape(nameEscBuf[:0], field.nameBytes)
field.nameEscHTML = `"` + string(nameEscBuf) + `":`
field.nameNonEsc = `"` + field.name + `":`
if field.omitZero {
t := sf.Type
// Provide a function that uses a type's IsZero method.
switch {
case t.Kind() == reflect.Interface && t.Implements(isZeroerType):
field.isZero = func(v reflect.Value) bool {
// Avoid panics calling IsZero on a nil interface or
// non-nil interface with nil pointer.
return v.IsNil() ||
(v.Elem().Kind() == reflect.Pointer && v.Elem().IsNil()) ||
v.Interface().(isZeroer).IsZero()
}
case t.Kind() == reflect.Pointer && t.Implements(isZeroerType):
field.isZero = func(v reflect.Value) bool {
// Avoid panics calling IsZero on nil pointer.
return v.IsNil() || v.Interface().(isZeroer).IsZero()
}
case t.Implements(isZeroerType):
field.isZero = func(v reflect.Value) bool {
return v.Interface().(isZeroer).IsZero()
}
case reflect.PointerTo(t).Implements(isZeroerType):
field.isZero = func(v reflect.Value) bool {
if !v.CanAddr() {
// Temporarily box v so we can take the address.
v2 := reflect.New(v.Type()).Elem()
v2.Set(v)
v = v2
}
return v.Addr().Interface().(isZeroer).IsZero()
}
}
}
fields = append(fields, field)
if count[f.typ] > 1 {
// If there were multiple instances, add a second,
// so that the annihilation code will see a duplicate.
// It only cares about the distinction between 1 and 2,
// so don't bother generating any more copies.
fields = append(fields, fields[len(fields)-1])
}
continue
}
// Record new anonymous struct to explore in next round.
nextCount[ft]++
if nextCount[ft] == 1 {
next = append(next, StdField{name: ft.Name(), index: index, typ: ft})
}
}
}
}
sort.Slice(fields, func(i, j int) bool {
a, b := fields[i], fields[j]
// sort field by name, breaking ties with depth, then
// breaking ties with "name came from json tag", then
// breaking ties with index sequence.
if c := strings.Compare(a.name, b.name); c != 0 {
return c < 0
}
if len(a.index) != len(b.index) {
return len(a.index) < len(b.index)
}
if a.tag != b.tag {
if a.tag {
return true
}
return false
}
return compare(a.index, b.index) < 0
})
// Delete all fields that are hidden by the Go rules for embedded fields,
// except that fields with JSON tags are promoted.
// The fields are sorted in primary order of name, secondary order
// of field index length. Loop over names; for each name, delete
// hidden fields by choosing the one dominant field that survives.
out := fields[:0]
for advance, i := 0, 0; i < len(fields); i += advance {
// One iteration per name.
// Find the sequence of fields with the name of this first field.
fi := fields[i]
name := fi.name
for advance = 1; i+advance < len(fields); advance++ {
fj := fields[i+advance]
if fj.name != name {
break
}
}
if advance == 1 { // Only one field with this name
out = append(out, fi)
continue
}
dominant, ok := dominantField(fields[i : i+advance])
if ok {
out = append(out, dominant)
}
}
fields = out
sort.Slice(fields, func(i, j int) bool {
a, b := fields[i], fields[j]
return compare(a.index, b.index) < 0
})
exactNameIndex := make(map[string]*StdField, len(fields))
foldedNameIndex := make(map[string]*StdField, len(fields))
for i, field := range fields {
exactNameIndex[field.name] = &fields[i]
// For historical reasons, first folded match takes precedence.
if _, ok := foldedNameIndex[string(foldName(field.nameBytes))]; !ok {
foldedNameIndex[string(foldName(field.nameBytes))] = &fields[i]
}
}
return StdStructFields{fields, exactNameIndex, foldedNameIndex}
}
func compare(s1, s2 []int) int {
for i, v1 := range s1 {
if i >= len(s2) {
return +1
}
v2 := s2[i]
if v1 != v2 {
return v1 - v2
}
}
if len(s1) < len(s2) {
return -1
}
return 0
}
type isZeroer interface {
IsZero() bool
}
var isZeroerType = reflect.TypeOf((*isZeroer)(nil)).Elem()
// tagOptions is the string following a comma in a struct field's "json"
// tag, or the empty string. It does not include the leading comma.
type tagOptions string
// parseTag splits a struct field's json tag into its name and
// comma-separated options.
func parseTag(tag string) (string, tagOptions) {
tag, opt, _ := strings.Cut(tag, ",")
return tag, tagOptions(opt)
}
// Contains reports whether a comma-separated list of options
// contains a particular substr flag. substr must be surrounded by a
// string boundary or commas.
func (o tagOptions) Contains(optionName string) bool {
if len(o) == 0 {
return false
}
s := string(o)
for s != "" {
var name string
name, s, _ = strings.Cut(s, ",")
if name == optionName {
return true
}
}
return false
}
func isValidTag(s string) bool {
if s == "" {
return false
}
for _, c := range s {
switch {
case strings.ContainsRune("!#$%&()*+-./:;<=>?@[]^_{|}~ ", c):
// Backslash and quote chars are reserved, but
// otherwise any punctuation chars are allowed
// in a tag name.
case !unicode.IsLetter(c) && !unicode.IsDigit(c):
return false
}
}
return true
}
// dominantField looks through the fields, all of which are known to
// have the same name, to find the single field that dominates the
// others using Go's embedding rules, modified by the presence of
// JSON tags. If there are multiple top-level fields, the boolean
// will be false: This condition is an error in Go and we skip all
// the fields.
func dominantField(fields []StdField) (StdField, bool) {
// The fields are sorted in increasing index-length order, then by presence of tag.
// That means that the first field is the dominant one. We need only check
// for error cases: two fields at top level, either both tagged or neither tagged.
if len(fields) > 1 && len(fields[0].index) == len(fields[1].index) && fields[0].tag == fields[1].tag {
return StdField{}, false
}
return fields[0], true
}
// foldName returns a folded string such that foldName(x) == foldName(y)
// is identical to bytes.EqualFold(x, y).
func foldName(in []byte) []byte {
// This is inlinable to take advantage of "function outlining".
var arr [32]byte // large enough for most JSON names
return appendFoldedName(arr[:0], in)
}
func appendFoldedName(out, in []byte) []byte {
for i := 0; i < len(in); {
// Handle single-byte ASCII.
if c := in[i]; c < utf8.RuneSelf {
if 'a' <= c && c <= 'z' {
c -= 'a' - 'A'
}
out = append(out, c)
i++
continue
}
// Handle multi-byte Unicode.
r, n := utf8.DecodeRune(in[i:])
out = utf8.AppendRune(out, foldRune(r))
i += n
}
return out
}
// foldRune is returns the smallest rune for all runes in the same fold set.
func foldRune(r rune) rune {
for {
r2 := unicode.SimpleFold(r)
if r2 <= r {
return r2
}
r = r2
}
}