You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

encode.go 14 kB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561
  1. //
  2. // Copyright (c) 2011-2019 Canonical Ltd
  3. //
  4. // Licensed under the Apache License, Version 2.0 (the "License");
  5. // you may not use this file except in compliance with the License.
  6. // You may obtain a copy of the License at
  7. //
  8. // http://www.apache.org/licenses/LICENSE-2.0
  9. //
  10. // Unless required by applicable law or agreed to in writing, software
  11. // distributed under the License is distributed on an "AS IS" BASIS,
  12. // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13. // See the License for the specific language governing permissions and
  14. // limitations under the License.
  15. package yaml
  16. import (
  17. "encoding"
  18. "fmt"
  19. "io"
  20. "reflect"
  21. "regexp"
  22. "sort"
  23. "strconv"
  24. "strings"
  25. "time"
  26. "unicode/utf8"
  27. )
  28. type encoder struct {
  29. emitter yaml_emitter_t
  30. event yaml_event_t
  31. out []byte
  32. flow bool
  33. indent int
  34. doneInit bool
  35. }
  36. func newEncoder() *encoder {
  37. e := &encoder{}
  38. yaml_emitter_initialize(&e.emitter)
  39. yaml_emitter_set_output_string(&e.emitter, &e.out)
  40. yaml_emitter_set_unicode(&e.emitter, true)
  41. return e
  42. }
  43. func newEncoderWithWriter(w io.Writer) *encoder {
  44. e := &encoder{}
  45. yaml_emitter_initialize(&e.emitter)
  46. yaml_emitter_set_output_writer(&e.emitter, w)
  47. yaml_emitter_set_unicode(&e.emitter, true)
  48. return e
  49. }
  50. func (e *encoder) init() {
  51. if e.doneInit {
  52. return
  53. }
  54. if e.indent == 0 {
  55. e.indent = 4
  56. }
  57. e.emitter.best_indent = e.indent
  58. yaml_stream_start_event_initialize(&e.event, yaml_UTF8_ENCODING)
  59. e.emit()
  60. e.doneInit = true
  61. }
  62. func (e *encoder) finish() {
  63. e.emitter.open_ended = false
  64. yaml_stream_end_event_initialize(&e.event)
  65. e.emit()
  66. }
  67. func (e *encoder) destroy() {
  68. yaml_emitter_delete(&e.emitter)
  69. }
  70. func (e *encoder) emit() {
  71. // This will internally delete the e.event value.
  72. e.must(yaml_emitter_emit(&e.emitter, &e.event))
  73. }
  74. func (e *encoder) must(ok bool) {
  75. if !ok {
  76. msg := e.emitter.problem
  77. if msg == "" {
  78. msg = "unknown problem generating YAML content"
  79. }
  80. failf("%s", msg)
  81. }
  82. }
  83. func (e *encoder) marshalDoc(tag string, in reflect.Value) {
  84. e.init()
  85. var node *Node
  86. if in.IsValid() {
  87. node, _ = in.Interface().(*Node)
  88. }
  89. if node != nil && node.Kind == DocumentNode {
  90. e.nodev(in)
  91. } else {
  92. yaml_document_start_event_initialize(&e.event, nil, nil, true)
  93. e.emit()
  94. e.marshal(tag, in)
  95. yaml_document_end_event_initialize(&e.event, true)
  96. e.emit()
  97. }
  98. }
  99. func (e *encoder) marshal(tag string, in reflect.Value) {
  100. tag = shortTag(tag)
  101. if !in.IsValid() || in.Kind() == reflect.Ptr && in.IsNil() {
  102. e.nilv()
  103. return
  104. }
  105. iface := in.Interface()
  106. switch value := iface.(type) {
  107. case *Node:
  108. e.nodev(in)
  109. return
  110. case time.Time:
  111. e.timev(tag, in)
  112. return
  113. case *time.Time:
  114. e.timev(tag, in.Elem())
  115. return
  116. case time.Duration:
  117. e.stringv(tag, reflect.ValueOf(value.String()))
  118. return
  119. case Marshaler:
  120. v, err := value.MarshalYAML()
  121. if err != nil {
  122. fail(err)
  123. }
  124. if v == nil {
  125. e.nilv()
  126. return
  127. }
  128. e.marshal(tag, reflect.ValueOf(v))
  129. return
  130. case encoding.TextMarshaler:
  131. text, err := value.MarshalText()
  132. if err != nil {
  133. fail(err)
  134. }
  135. in = reflect.ValueOf(string(text))
  136. case nil:
  137. e.nilv()
  138. return
  139. }
  140. switch in.Kind() {
  141. case reflect.Interface:
  142. e.marshal(tag, in.Elem())
  143. case reflect.Map:
  144. e.mapv(tag, in)
  145. case reflect.Ptr:
  146. e.marshal(tag, in.Elem())
  147. case reflect.Struct:
  148. e.structv(tag, in)
  149. case reflect.Slice, reflect.Array:
  150. e.slicev(tag, in)
  151. case reflect.String:
  152. e.stringv(tag, in)
  153. case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
  154. e.intv(tag, in)
  155. case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
  156. e.uintv(tag, in)
  157. case reflect.Float32, reflect.Float64:
  158. e.floatv(tag, in)
  159. case reflect.Bool:
  160. e.boolv(tag, in)
  161. default:
  162. panic("cannot marshal type: " + in.Type().String())
  163. }
  164. }
  165. func (e *encoder) mapv(tag string, in reflect.Value) {
  166. e.mappingv(tag, func() {
  167. keys := keyList(in.MapKeys())
  168. sort.Sort(keys)
  169. for _, k := range keys {
  170. e.marshal("", k)
  171. e.marshal("", in.MapIndex(k))
  172. }
  173. })
  174. }
  175. func (e *encoder) fieldByIndex(v reflect.Value, index []int) (field reflect.Value) {
  176. for _, num := range index {
  177. for {
  178. if v.Kind() == reflect.Ptr {
  179. if v.IsNil() {
  180. return reflect.Value{}
  181. }
  182. v = v.Elem()
  183. continue
  184. }
  185. break
  186. }
  187. v = v.Field(num)
  188. }
  189. return v
  190. }
  191. func (e *encoder) structv(tag string, in reflect.Value) {
  192. sinfo, err := getStructInfo(in.Type())
  193. if err != nil {
  194. panic(err)
  195. }
  196. e.mappingv(tag, func() {
  197. for _, info := range sinfo.FieldsList {
  198. var value reflect.Value
  199. if info.Inline == nil {
  200. value = in.Field(info.Num)
  201. } else {
  202. value = e.fieldByIndex(in, info.Inline)
  203. if !value.IsValid() {
  204. continue
  205. }
  206. }
  207. if info.OmitEmpty && isZero(value) {
  208. continue
  209. }
  210. e.marshal("", reflect.ValueOf(info.Key))
  211. e.flow = info.Flow
  212. e.marshal("", value)
  213. }
  214. if sinfo.InlineMap >= 0 {
  215. m := in.Field(sinfo.InlineMap)
  216. if m.Len() > 0 {
  217. e.flow = false
  218. keys := keyList(m.MapKeys())
  219. sort.Sort(keys)
  220. for _, k := range keys {
  221. if _, found := sinfo.FieldsMap[k.String()]; found {
  222. panic(fmt.Sprintf("cannot have key %q in inlined map: conflicts with struct field", k.String()))
  223. }
  224. e.marshal("", k)
  225. e.flow = false
  226. e.marshal("", m.MapIndex(k))
  227. }
  228. }
  229. }
  230. })
  231. }
  232. func (e *encoder) mappingv(tag string, f func()) {
  233. implicit := tag == ""
  234. style := yaml_BLOCK_MAPPING_STYLE
  235. if e.flow {
  236. e.flow = false
  237. style = yaml_FLOW_MAPPING_STYLE
  238. }
  239. yaml_mapping_start_event_initialize(&e.event, nil, []byte(tag), implicit, style)
  240. e.emit()
  241. f()
  242. yaml_mapping_end_event_initialize(&e.event)
  243. e.emit()
  244. }
  245. func (e *encoder) slicev(tag string, in reflect.Value) {
  246. implicit := tag == ""
  247. style := yaml_BLOCK_SEQUENCE_STYLE
  248. if e.flow {
  249. e.flow = false
  250. style = yaml_FLOW_SEQUENCE_STYLE
  251. }
  252. e.must(yaml_sequence_start_event_initialize(&e.event, nil, []byte(tag), implicit, style))
  253. e.emit()
  254. n := in.Len()
  255. for i := 0; i < n; i++ {
  256. e.marshal("", in.Index(i))
  257. }
  258. e.must(yaml_sequence_end_event_initialize(&e.event))
  259. e.emit()
  260. }
  261. // isBase60 returns whether s is in base 60 notation as defined in YAML 1.1.
  262. //
  263. // The base 60 float notation in YAML 1.1 is a terrible idea and is unsupported
  264. // in YAML 1.2 and by this package, but these should be marshalled quoted for
  265. // the time being for compatibility with other parsers.
  266. func isBase60Float(s string) (result bool) {
  267. // Fast path.
  268. if s == "" {
  269. return false
  270. }
  271. c := s[0]
  272. if !(c == '+' || c == '-' || c >= '0' && c <= '9') || strings.IndexByte(s, ':') < 0 {
  273. return false
  274. }
  275. // Do the full match.
  276. return base60float.MatchString(s)
  277. }
  278. // From http://yaml.org/type/float.html, except the regular expression there
  279. // is bogus. In practice parsers do not enforce the "\.[0-9_]*" suffix.
  280. var base60float = regexp.MustCompile(`^[-+]?[0-9][0-9_]*(?::[0-5]?[0-9])+(?:\.[0-9_]*)?$`)
  281. // isOldBool returns whether s is bool notation as defined in YAML 1.1.
  282. //
  283. // We continue to force strings that YAML 1.1 would interpret as booleans to be
  284. // rendered as quotes strings so that the marshalled output valid for YAML 1.1
  285. // parsing.
  286. func isOldBool(s string) (result bool) {
  287. switch s {
  288. case "y", "Y", "yes", "Yes", "YES", "on", "On", "ON",
  289. "n", "N", "no", "No", "NO", "off", "Off", "OFF":
  290. return true
  291. default:
  292. return false
  293. }
  294. }
  295. func (e *encoder) stringv(tag string, in reflect.Value) {
  296. var style yaml_scalar_style_t
  297. s := in.String()
  298. canUsePlain := true
  299. switch {
  300. case !utf8.ValidString(s):
  301. if tag == binaryTag {
  302. failf("explicitly tagged !!binary data must be base64-encoded")
  303. }
  304. if tag != "" {
  305. failf("cannot marshal invalid UTF-8 data as %s", shortTag(tag))
  306. }
  307. // It can't be encoded directly as YAML so use a binary tag
  308. // and encode it as base64.
  309. tag = binaryTag
  310. s = encodeBase64(s)
  311. case tag == "":
  312. // Check to see if it would resolve to a specific
  313. // tag when encoded unquoted. If it doesn't,
  314. // there's no need to quote it.
  315. rtag, _ := resolve("", s)
  316. canUsePlain = rtag == strTag && !(isBase60Float(s) || isOldBool(s))
  317. }
  318. // Note: it's possible for user code to emit invalid YAML
  319. // if they explicitly specify a tag and a string containing
  320. // text that's incompatible with that tag.
  321. switch {
  322. case strings.Contains(s, "\n"):
  323. if e.flow {
  324. style = yaml_DOUBLE_QUOTED_SCALAR_STYLE
  325. } else {
  326. style = yaml_LITERAL_SCALAR_STYLE
  327. }
  328. case canUsePlain:
  329. style = yaml_PLAIN_SCALAR_STYLE
  330. default:
  331. style = yaml_DOUBLE_QUOTED_SCALAR_STYLE
  332. }
  333. e.emitScalar(s, "", tag, style, nil, nil, nil, nil)
  334. }
  335. func (e *encoder) boolv(tag string, in reflect.Value) {
  336. var s string
  337. if in.Bool() {
  338. s = "true"
  339. } else {
  340. s = "false"
  341. }
  342. e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE, nil, nil, nil, nil)
  343. }
  344. func (e *encoder) intv(tag string, in reflect.Value) {
  345. s := strconv.FormatInt(in.Int(), 10)
  346. e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE, nil, nil, nil, nil)
  347. }
  348. func (e *encoder) uintv(tag string, in reflect.Value) {
  349. s := strconv.FormatUint(in.Uint(), 10)
  350. e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE, nil, nil, nil, nil)
  351. }
  352. func (e *encoder) timev(tag string, in reflect.Value) {
  353. t := in.Interface().(time.Time)
  354. s := t.Format(time.RFC3339Nano)
  355. e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE, nil, nil, nil, nil)
  356. }
  357. func (e *encoder) floatv(tag string, in reflect.Value) {
  358. // Issue #352: When formatting, use the precision of the underlying value
  359. precision := 64
  360. if in.Kind() == reflect.Float32 {
  361. precision = 32
  362. }
  363. s := strconv.FormatFloat(in.Float(), 'g', -1, precision)
  364. switch s {
  365. case "+Inf":
  366. s = ".inf"
  367. case "-Inf":
  368. s = "-.inf"
  369. case "NaN":
  370. s = ".nan"
  371. }
  372. e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE, nil, nil, nil, nil)
  373. }
  374. func (e *encoder) nilv() {
  375. e.emitScalar("null", "", "", yaml_PLAIN_SCALAR_STYLE, nil, nil, nil, nil)
  376. }
  377. func (e *encoder) emitScalar(value, anchor, tag string, style yaml_scalar_style_t, head, line, foot, tail []byte) {
  378. // TODO Kill this function. Replace all initialize calls by their underlining Go literals.
  379. implicit := tag == ""
  380. if !implicit {
  381. tag = longTag(tag)
  382. }
  383. e.must(yaml_scalar_event_initialize(&e.event, []byte(anchor), []byte(tag), []byte(value), implicit, implicit, style))
  384. e.event.head_comment = head
  385. e.event.line_comment = line
  386. e.event.foot_comment = foot
  387. e.event.tail_comment = tail
  388. e.emit()
  389. }
  390. func (e *encoder) nodev(in reflect.Value) {
  391. e.node(in.Interface().(*Node), "")
  392. }
  393. func (e *encoder) node(node *Node, tail string) {
  394. // If the tag was not explicitly requested, and dropping it won't change the
  395. // implicit tag of the value, don't include it in the presentation.
  396. var tag = node.Tag
  397. var stag = shortTag(tag)
  398. var rtag string
  399. var forceQuoting bool
  400. if tag != "" && node.Style&TaggedStyle == 0 {
  401. if node.Kind == ScalarNode {
  402. if stag == strTag && node.Style&(SingleQuotedStyle|DoubleQuotedStyle|LiteralStyle|FoldedStyle) != 0 {
  403. tag = ""
  404. } else {
  405. rtag, _ = resolve("", node.Value)
  406. if rtag == stag {
  407. tag = ""
  408. } else if stag == strTag {
  409. tag = ""
  410. forceQuoting = true
  411. }
  412. }
  413. } else {
  414. switch node.Kind {
  415. case MappingNode:
  416. rtag = mapTag
  417. case SequenceNode:
  418. rtag = seqTag
  419. }
  420. if rtag == stag {
  421. tag = ""
  422. }
  423. }
  424. }
  425. switch node.Kind {
  426. case DocumentNode:
  427. yaml_document_start_event_initialize(&e.event, nil, nil, true)
  428. e.event.head_comment = []byte(node.HeadComment)
  429. e.emit()
  430. for _, node := range node.Content {
  431. e.node(node, "")
  432. }
  433. yaml_document_end_event_initialize(&e.event, true)
  434. e.event.foot_comment = []byte(node.FootComment)
  435. e.emit()
  436. case SequenceNode:
  437. style := yaml_BLOCK_SEQUENCE_STYLE
  438. if node.Style&FlowStyle != 0 {
  439. style = yaml_FLOW_SEQUENCE_STYLE
  440. }
  441. e.must(yaml_sequence_start_event_initialize(&e.event, []byte(node.Anchor), []byte(tag), tag == "", style))
  442. e.event.head_comment = []byte(node.HeadComment)
  443. e.emit()
  444. for _, node := range node.Content {
  445. e.node(node, "")
  446. }
  447. e.must(yaml_sequence_end_event_initialize(&e.event))
  448. e.event.line_comment = []byte(node.LineComment)
  449. e.event.foot_comment = []byte(node.FootComment)
  450. e.emit()
  451. case MappingNode:
  452. style := yaml_BLOCK_MAPPING_STYLE
  453. if node.Style&FlowStyle != 0 {
  454. style = yaml_FLOW_MAPPING_STYLE
  455. }
  456. yaml_mapping_start_event_initialize(&e.event, []byte(node.Anchor), []byte(tag), tag == "", style)
  457. e.event.tail_comment = []byte(tail)
  458. e.event.head_comment = []byte(node.HeadComment)
  459. e.emit()
  460. // The tail logic below moves the foot comment of prior keys to the following key,
  461. // since the value for each key may be a nested structure and the foot needs to be
  462. // processed only the entirety of the value is streamed. The last tail is processed
  463. // with the mapping end event.
  464. var tail string
  465. for i := 0; i+1 < len(node.Content); i += 2 {
  466. k := node.Content[i]
  467. foot := k.FootComment
  468. if foot != "" {
  469. kopy := *k
  470. kopy.FootComment = ""
  471. k = &kopy
  472. }
  473. e.node(k, tail)
  474. tail = foot
  475. v := node.Content[i+1]
  476. e.node(v, "")
  477. }
  478. yaml_mapping_end_event_initialize(&e.event)
  479. e.event.tail_comment = []byte(tail)
  480. e.event.line_comment = []byte(node.LineComment)
  481. e.event.foot_comment = []byte(node.FootComment)
  482. e.emit()
  483. case AliasNode:
  484. yaml_alias_event_initialize(&e.event, []byte(node.Value))
  485. e.event.head_comment = []byte(node.HeadComment)
  486. e.event.line_comment = []byte(node.LineComment)
  487. e.event.foot_comment = []byte(node.FootComment)
  488. e.emit()
  489. case ScalarNode:
  490. value := node.Value
  491. if !utf8.ValidString(value) {
  492. if tag == binaryTag {
  493. failf("explicitly tagged !!binary data must be base64-encoded")
  494. }
  495. if tag != "" {
  496. failf("cannot marshal invalid UTF-8 data as %s", shortTag(tag))
  497. }
  498. // It can't be encoded directly as YAML so use a binary tag
  499. // and encode it as base64.
  500. tag = binaryTag
  501. value = encodeBase64(value)
  502. }
  503. style := yaml_PLAIN_SCALAR_STYLE
  504. switch {
  505. case node.Style&DoubleQuotedStyle != 0:
  506. style = yaml_DOUBLE_QUOTED_SCALAR_STYLE
  507. case node.Style&SingleQuotedStyle != 0:
  508. style = yaml_SINGLE_QUOTED_SCALAR_STYLE
  509. case node.Style&LiteralStyle != 0:
  510. style = yaml_LITERAL_SCALAR_STYLE
  511. case node.Style&FoldedStyle != 0:
  512. style = yaml_FOLDED_SCALAR_STYLE
  513. case strings.Contains(value, "\n"):
  514. style = yaml_LITERAL_SCALAR_STYLE
  515. case forceQuoting:
  516. style = yaml_DOUBLE_QUOTED_SCALAR_STYLE
  517. }
  518. e.emitScalar(value, node.Anchor, tag, style, []byte(node.HeadComment), []byte(node.LineComment), []byte(node.FootComment), []byte(tail))
  519. }
  520. }