diff options
Diffstat (limited to 'vendor/github.com/beevik/etree')
| -rw-r--r-- | vendor/github.com/beevik/etree/CONTRIBUTORS | 15 | ||||
| -rw-r--r-- | vendor/github.com/beevik/etree/LICENSE | 24 | ||||
| -rw-r--r-- | vendor/github.com/beevik/etree/README.md | 207 | ||||
| -rw-r--r-- | vendor/github.com/beevik/etree/RELEASE_NOTES.md | 212 | ||||
| -rw-r--r-- | vendor/github.com/beevik/etree/etree.go | 1857 | ||||
| -rw-r--r-- | vendor/github.com/beevik/etree/helpers.go | 394 | ||||
| -rw-r--r-- | vendor/github.com/beevik/etree/path.go | 605 |
7 files changed, 3314 insertions, 0 deletions
diff --git a/vendor/github.com/beevik/etree/CONTRIBUTORS b/vendor/github.com/beevik/etree/CONTRIBUTORS new file mode 100644 index 0000000..0bb95df --- /dev/null +++ b/vendor/github.com/beevik/etree/CONTRIBUTORS @@ -0,0 +1,15 @@ +Brett Vickers (beevik) +Felix Geisendörfer (felixge) +Kamil Kisiel (kisielk) +Graham King (grahamking) +Matt Smith (ma314smith) +Michal Jemala (michaljemala) +Nicolas Piganeau (npiganeau) +Chris Brown (ccbrown) +Earncef Sequeira (earncef) +Gabriel de Labachelerie (wuzuf) +Martin Dosch (mdosch) +Hugo Wetterberg (hugowetterberg) +Tobias Theel (nerzal) +Daniel Potapov (dpotapov) +Mikhail Ferapontow (MikhailFerapontow) diff --git a/vendor/github.com/beevik/etree/LICENSE b/vendor/github.com/beevik/etree/LICENSE new file mode 100644 index 0000000..4ebd856 --- /dev/null +++ b/vendor/github.com/beevik/etree/LICENSE @@ -0,0 +1,24 @@ +Copyright 2015-2024 Brett Vickers. All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions +are met: + + 1. Redistributions of source code must retain the above copyright + notice, this list of conditions and the following disclaimer. + + 2. Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + +THIS SOFTWARE IS PROVIDED BY COPYRIGHT HOLDER ``AS IS'' AND ANY +EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR +PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL COPYRIGHT HOLDER OR +CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, +EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR +PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY +OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. diff --git a/vendor/github.com/beevik/etree/README.md b/vendor/github.com/beevik/etree/README.md new file mode 100644 index 0000000..6d15736 --- /dev/null +++ b/vendor/github.com/beevik/etree/README.md @@ -0,0 +1,207 @@ +[](https://godoc.org/github.com/beevik/etree) +[](https://github.com/beevik/etree/actions/workflows/go.yml) + +etree +===== + +The etree package is a lightweight, pure go package that expresses XML in +the form of an element tree. Its design was inspired by the Python +[ElementTree](http://docs.python.org/2/library/xml.etree.elementtree.html) +module. + +Some of the package's capabilities and features: + +* Represents XML documents as trees of elements for easy traversal. +* Imports, serializes, modifies or creates XML documents from scratch. +* Writes and reads XML to/from files, byte slices, strings and io interfaces. +* Performs simple or complex searches with lightweight XPath-like query APIs. +* Auto-indents XML using spaces or tabs for better readability. +* Implemented in pure go; depends only on standard go libraries. +* Built on top of the go [encoding/xml](http://golang.org/pkg/encoding/xml) + package. + +The etree package is compatible with go versions 1.23 and later. + +### Creating an XML document + +The following example creates an XML document from scratch using the etree +package and outputs its indented contents to stdout. +```go +doc := etree.NewDocument() +doc.CreateProcInst("xml", `version="1.0" encoding="UTF-8"`) +doc.CreateProcInst("xml-stylesheet", `type="text/xsl" href="style.xsl"`) + +people := doc.CreateElement("People") +people.CreateComment("These are all known people") + +jon := people.CreateElement("Person") +jon.CreateAttr("name", "Jon") + +sally := people.CreateElement("Person") +sally.CreateAttr("name", "Sally") + +doc.Indent(2) +doc.WriteTo(os.Stdout) +``` + +Output: +```xml +<?xml version="1.0" encoding="UTF-8"?> +<?xml-stylesheet type="text/xsl" href="style.xsl"?> +<People> + <!--These are all known people--> + <Person name="Jon"/> + <Person name="Sally"/> +</People> +``` + +### Reading an XML file + +Suppose you have a file on disk called `bookstore.xml` containing the +following data: + +```xml +<bookstore xmlns:p="urn:schemas-books-com:prices"> + + <book category="COOKING"> + <title lang="en">Everyday Italian</title> + <author>Giada De Laurentiis</author> + <year>2005</year> + <p:price>30.00</p:price> + </book> + + <book category="CHILDREN"> + <title lang="en">Harry Potter</title> + <author>J K. Rowling</author> + <year>2005</year> + <p:price>29.99</p:price> + </book> + + <book category="WEB"> + <title lang="en">XQuery Kick Start</title> + <author>James McGovern</author> + <author>Per Bothner</author> + <author>Kurt Cagle</author> + <author>James Linn</author> + <author>Vaidyanathan Nagarajan</author> + <year>2003</year> + <p:price>49.99</p:price> + </book> + + <book category="WEB"> + <title lang="en">Learning XML</title> + <author>Erik T. Ray</author> + <year>2003</year> + <p:price>39.95</p:price> + </book> + +</bookstore> +``` + +This code reads the file's contents into an etree document. +```go +doc := etree.NewDocument() +if err := doc.ReadFromFile("bookstore.xml"); err != nil { + panic(err) +} +``` + +You can also read XML from a string, a byte slice, or an `io.Reader`. + +### Processing elements and attributes + +This example illustrates several ways to access elements and attributes using +etree selection queries. +```go +root := doc.SelectElement("bookstore") +fmt.Println("ROOT element:", root.Tag) + +for _, book := range root.SelectElementsSeq("book") { + fmt.Println("CHILD element:", book.Tag) + if title := book.SelectElement("title"); title != nil { + lang := title.SelectAttrValue("lang", "unknown") + fmt.Printf(" TITLE: %s (%s)\n", title.Text(), lang) + } + for _, attr := range book.Attr { + fmt.Printf(" ATTR: %s=%s\n", attr.Key, attr.Value) + } +} +``` +Output: +``` +ROOT element: bookstore +CHILD element: book + TITLE: Everyday Italian (en) + ATTR: category=COOKING +CHILD element: book + TITLE: Harry Potter (en) + ATTR: category=CHILDREN +CHILD element: book + TITLE: XQuery Kick Start (en) + ATTR: category=WEB +CHILD element: book + TITLE: Learning XML (en) + ATTR: category=WEB +``` + +### Path queries + +This example uses etree's path functions to select all book titles that fall +into the category of 'WEB'. The double-slash prefix in the path causes the +search for book elements to occur recursively; book elements may appear at any +level of the XML hierarchy. +```go +for _, t := range doc.FindElementsSeq("//book[@category='WEB']/title") { + fmt.Println("Title:", t.Text()) +} +``` + +Output: +``` +Title: XQuery Kick Start +Title: Learning XML +``` + +This example finds the first book element under the root bookstore element and +outputs the tag and text of each of its child elements. +```go +for _, e := range doc.FindElementsSeq("./bookstore/book[1]/*") { + fmt.Printf("%s: %s\n", e.Tag, e.Text()) +} +``` + +Output: +``` +title: Everyday Italian +author: Giada De Laurentiis +year: 2005 +price: 30.00 +``` + +This example finds all books with a price of 49.99 and outputs their titles. +```go +path := etree.MustCompilePath("./bookstore/book[p:price='49.99']/title") +for _, e := range doc.FindElementsPathSeq(path) { + fmt.Println(e.Text()) +} +``` + +Output: +``` +XQuery Kick Start +``` + +Note that this example uses the `FindElementsPathSeq` function, which takes as +an argument a pre-compiled path object. Use precompiled paths when you plan to +search with the same path more than once. + +### Other features + +These are just a few examples of the things the etree package can do. See the +[documentation](http://godoc.org/github.com/beevik/etree) for a complete +description of its capabilities. + +### Contributing + +This project accepts contributions. Just fork the repo and submit a pull +request! diff --git a/vendor/github.com/beevik/etree/RELEASE_NOTES.md b/vendor/github.com/beevik/etree/RELEASE_NOTES.md new file mode 100644 index 0000000..96fe151 --- /dev/null +++ b/vendor/github.com/beevik/etree/RELEASE_NOTES.md @@ -0,0 +1,212 @@ +Release 1.6.0 +============= + +**Changes** + +* Added new iterator versions of existing functions that return slices of + `Element` pointers: `ChildElementsSeq`, `SelectElementsSeq`, + `FindElementsSeq`, and `FindElementsPathSeq`. +* Improved performance of functions that return a single element. +* Because of its use of iterators, this package now requires go 1.23 or later. + +Release 1.5.1 +============= + +**Fixes** + +* Fixed a bug in `InsertChildAt`. + +Release 1.5.0 +============= + +**Changes** + +* Added `Element` function `CreateChild`, which calls a continuation function + after creating and adding a child element. + +**Fixes** + +* Removed a potential conflict between two `ReadSettings` values. When + `AttrSingleQuote` is true, `CanonicalAttrVal` is forced to be false. + +Release 1.4.1 +============= + +**Changes** + +* Minimal go version updated to 1.21. +* Default-initialized CharsetReader causes same result as NewDocument(). +* When reading an XML document, attributes are parsed more efficiently. + +Release v1.4.0 +============== + +**New Features** + +* Add `AutoClose` option to `ReadSettings`. +* Add `ValidateInput` to `ReadSettings`. +* Add `NotNil` function to `Element`. +* Add `NextSibling` and `PrevSibling` functions to `Element`. + +Release v1.3.0 +============== + +**New Features** + +* Add support for double-quotes in filter path queries. +* Add `PreserveDuplicateAttrs` to `ReadSettings`. +* Add `ReindexChildren` to `Element`. + +Release v1.2.0 +============== + +**New Features** + +* Add the ability to write XML fragments using Token WriteTo functions. +* Add the ability to re-indent an XML element as though it were the root of + the document. +* Add a ReadSettings option to preserve CDATA blocks when reading and XML + document. + +Release v1.1.4 +============== + +**New Features** + +* Add the ability to preserve whitespace in leaf elements during indent. +* Add the ability to suppress a document-trailing newline during indent. +* Add choice of XML attribute quoting style (single-quote or double-quote). + +**Removed Features** + +* Removed the CDATA preservation change introduced in v1.1.3. It was + implemented in a way that broke the ability to process XML documents + encoded using non-UTF8 character sets. + +Release v1.1.3 +============== + +* XML reads now preserve CDATA sections instead of converting them to + standard character data. + +Release v1.1.2 +============== + +* Fixed a path parsing bug. +* The `Element.Text` function now handles comments embedded between + character data spans. + +Release v1.1.1 +============== + +* Updated go version in `go.mod` to 1.20 + +Release v1.1.0 +============== + +**New Features** + +* New attribute helpers. + * Added the `Element.SortAttrs` method, which lexicographically sorts an + element's attributes by key. +* New `ReadSettings` properties. + * Added `Entity` for the support of custom entity maps. +* New `WriteSettings` properties. + * Added `UseCRLF` to allow the output of CR-LF newlines instead of the + default LF newlines. This is useful on Windows systems. +* Additional support for text and CDATA sections. + * The `Element.Text` method now returns the concatenation of all consecutive + character data tokens immediately following an element's opening tag. + * Added `Element.SetCData` to replace the character data immediately + following an element's opening tag with a CDATA section. + * Added `Element.CreateCData` to create and add a CDATA section child + `CharData` token to an element. + * Added `Element.CreateText` to create and add a child text `CharData` token + to an element. + * Added `NewCData` to create a parentless CDATA section `CharData` token. + * Added `NewText` to create a parentless text `CharData` + token. + * Added `CharData.IsCData` to detect if the token contains a CDATA section. + * Added `CharData.IsWhitespace` to detect if the token contains whitespace + inserted by one of the document Indent functions. + * Modified `Element.SetText` so that it replaces a run of consecutive + character data tokens following the element's opening tag (instead of just + the first one). +* New "tail text" support. + * Added the `Element.Tail` method, which returns the text immediately + following an element's closing tag. + * Added the `Element.SetTail` method, which modifies the text immediately + following an element's closing tag. +* New element child insertion and removal methods. + * Added the `Element.InsertChildAt` method, which inserts a new child token + before the specified child token index. + * Added the `Element.RemoveChildAt` method, which removes the child token at + the specified child token index. +* New element and attribute queries. + * Added the `Element.Index` method, which returns the element's index within + its parent element's child token list. + * Added the `Element.NamespaceURI` method to return the namespace URI + associated with an element. + * Added the `Attr.NamespaceURI` method to return the namespace URI + associated with an element. + * Added the `Attr.Element` method to return the element that an attribute + belongs to. +* New Path filter functions. + * Added `[local-name()='val']` to keep elements whose unprefixed tag matches + the desired value. + * Added `[name()='val']` to keep elements whose full tag matches the desired + value. + * Added `[namespace-prefix()='val']` to keep elements whose namespace prefix + matches the desired value. + * Added `[namespace-uri()='val']` to keep elements whose namespace URI + matches the desired value. + +**Bug Fixes** + +* A default XML `CharSetReader` is now used to prevent failed parsing of XML + documents using certain encodings. + ([Issue](https://github.com/beevik/etree/issues/53)). +* All characters are now properly escaped according to XML parsing rules. + ([Issue](https://github.com/beevik/etree/issues/55)). +* The `Document.Indent` and `Document.IndentTabs` functions no longer insert + empty string `CharData` tokens. + +**Deprecated** + +* `Element` + * The `InsertChild` method is deprecated. Use `InsertChildAt` instead. + * The `CreateCharData` method is deprecated. Use `CreateText` instead. +* `CharData` + * The `NewCharData` method is deprecated. Use `NewText` instead. + + +Release v1.0.1 +============== + +**Changes** + +* Added support for absolute etree Path queries. An absolute path begins with + `/` or `//` and begins its search from the element's document root. +* Added [`GetPath`](https://godoc.org/github.com/beevik/etree#Element.GetPath) + and [`GetRelativePath`](https://godoc.org/github.com/beevik/etree#Element.GetRelativePath) + functions to the [`Element`](https://godoc.org/github.com/beevik/etree#Element) + type. + +**Breaking changes** + +* A path starting with `//` is now interpreted as an absolute path. + Previously, it was interpreted as a relative path starting from the element + whose + [`FindElement`](https://godoc.org/github.com/beevik/etree#Element.FindElement) + method was called. To remain compatible with this release, all paths + prefixed with `//` should be prefixed with `.//` when called from any + element other than the document's root. +* [**edit 2/1/2019**]: Minor releases should not contain breaking changes. + Even though this breaking change was very minor, it was a mistake to include + it in this minor release. In the future, all breaking changes will be + limited to major releases (e.g., version 2.0.0). + +Release v1.0.0 +============== + +Initial release. diff --git a/vendor/github.com/beevik/etree/etree.go b/vendor/github.com/beevik/etree/etree.go new file mode 100644 index 0000000..bfe1f06 --- /dev/null +++ b/vendor/github.com/beevik/etree/etree.go @@ -0,0 +1,1857 @@ +// Copyright 2015-2019 Brett Vickers. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Package etree provides XML services through an Element Tree +// abstraction. +package etree + +import ( + "bufio" + "bytes" + "encoding/xml" + "errors" + "io" + "iter" + "os" + "slices" + "strings" +) + +const ( + // NoIndent is used with the IndentSettings record to remove all + // indenting. + NoIndent = -1 +) + +// ErrXML is returned when XML parsing fails due to incorrect formatting. +var ErrXML = errors.New("etree: invalid XML format") + +// cdataPrefix is used to detect CDATA text when ReadSettings.PreserveCData is +// true. +var cdataPrefix = []byte("<![CDATA[") + +// ReadSettings determine the default behavior of the Document's ReadFrom* +// functions. +type ReadSettings struct { + // CharsetReader, if non-nil, defines a function to generate + // charset-conversion readers, converting from the provided non-UTF-8 + // charset into UTF-8. If nil, the ReadFrom* functions will use a + // "pass-through" CharsetReader that performs no conversion on the reader's + // data regardless of the value of the "charset" encoding string. Default: + // nil. + CharsetReader func(charset string, input io.Reader) (io.Reader, error) + + // Permissive allows input containing common mistakes such as missing tags + // or attribute values. Default: false. + Permissive bool + + // Preserve CDATA character data blocks when decoding XML (instead of + // converting it to normal character text). This entails additional + // processing and memory usage during ReadFrom* operations. Default: + // false. + PreserveCData bool + + // When an element has two or more attributes with the same name, + // preserve them instead of keeping only one. Default: false. + PreserveDuplicateAttrs bool + + // ValidateInput forces all ReadFrom* functions to validate that the + // provided input is composed of "well-formed"(*) XML before processing it. + // If invalid XML is detected, the ReadFrom* functions return an error. + // Because this option requires the input to be processed twice, it incurs a + // significant performance penalty. Default: false. + // + // (*) Note that this definition of "well-formed" is in the context of the + // go standard library's encoding/xml package. Go's encoding/xml package + // does not, in fact, guarantee well-formed XML as specified by the W3C XML + // recommendation. See: https://github.com/golang/go/issues/68299 + ValidateInput bool + + // Entity to be passed to standard xml.Decoder. Default: nil. + Entity map[string]string + + // When Permissive is true, AutoClose indicates a set of elements to + // consider closed immediately after they are opened, regardless of + // whether an end element is present. Commonly set to xml.HTMLAutoClose. + // Default: nil. + AutoClose []string +} + +// defaultCharsetReader is used by the xml decoder when the ReadSettings +// CharsetReader value is nil. It behaves as a "pass-through", ignoring +// the requested charset parameter and skipping conversion altogether. +func defaultCharsetReader(charset string, input io.Reader) (io.Reader, error) { + return input, nil +} + +// dup creates a duplicate of the ReadSettings object. +func (s *ReadSettings) dup() ReadSettings { + var entityCopy map[string]string + if s.Entity != nil { + entityCopy = make(map[string]string) + for k, v := range s.Entity { + entityCopy[k] = v + } + } + return ReadSettings{ + CharsetReader: s.CharsetReader, + Permissive: s.Permissive, + Entity: entityCopy, + } +} + +// WriteSettings determine the behavior of the Document's WriteTo* functions. +type WriteSettings struct { + // CanonicalEndTags forces the production of XML end tags, even for + // elements that have no child elements. Default: false. + CanonicalEndTags bool + + // CanonicalText forces the production of XML character references for + // text data characters &, <, and >. If false, XML character references + // are also produced for " and '. Default: false. + CanonicalText bool + + // CanonicalAttrVal forces the production of XML character references for + // attribute value characters &, < and ". If false, XML character + // references are also produced for > and '. Ignored when AttrSingleQuote + // is true. Default: false. + CanonicalAttrVal bool + + // AttrSingleQuote causes attributes to use single quotes (attr='example') + // instead of double quotes (attr = "example") when set to true. Default: + // false. + AttrSingleQuote bool + + // UseCRLF causes the document's Indent* functions to use a carriage return + // followed by a linefeed ("\r\n") when outputting a newline. If false, + // only a linefeed is used ("\n"). Default: false. + // + // Deprecated: UseCRLF is deprecated. Use IndentSettings.UseCRLF instead. + UseCRLF bool +} + +// dup creates a duplicate of the WriteSettings object. +func (s *WriteSettings) dup() WriteSettings { + return *s +} + +// IndentSettings determine the behavior of the Document's Indent* functions. +type IndentSettings struct { + // Spaces indicates the number of spaces to insert for each level of + // indentation. Set to etree.NoIndent to remove all indentation. Ignored + // when UseTabs is true. Default: 4. + Spaces int + + // UseTabs causes tabs to be used instead of spaces when indenting. + // Default: false. + UseTabs bool + + // UseCRLF causes newlines to be written as a carriage return followed by + // a linefeed ("\r\n"). If false, only a linefeed character is output + // for a newline ("\n"). Default: false. + UseCRLF bool + + // PreserveLeafWhitespace causes indent functions to preserve whitespace + // within XML elements containing only non-CDATA character data. Default: + // false. + PreserveLeafWhitespace bool + + // SuppressTrailingWhitespace suppresses the generation of a trailing + // whitespace characters (such as newlines) at the end of the indented + // document. Default: false. + SuppressTrailingWhitespace bool +} + +// NewIndentSettings creates a default IndentSettings record. +func NewIndentSettings() *IndentSettings { + return &IndentSettings{ + Spaces: 4, + UseTabs: false, + UseCRLF: false, + PreserveLeafWhitespace: false, + SuppressTrailingWhitespace: false, + } +} + +type indentFunc func(depth int) string + +func getIndentFunc(s *IndentSettings) indentFunc { + if s.UseTabs { + if s.UseCRLF { + return func(depth int) string { return indentCRLF(depth, indentTabs) } + } else { + return func(depth int) string { return indentLF(depth, indentTabs) } + } + } else { + if s.Spaces < 0 { + return func(depth int) string { return "" } + } else if s.UseCRLF { + return func(depth int) string { return indentCRLF(depth*s.Spaces, indentSpaces) } + } else { + return func(depth int) string { return indentLF(depth*s.Spaces, indentSpaces) } + } + } +} + +// Writer is the interface that wraps the Write* functions called by each token +// type's WriteTo function. +type Writer interface { + io.StringWriter + io.ByteWriter + io.Writer +} + +// A Token is an interface type used to represent XML elements, character +// data, CDATA sections, XML comments, XML directives, and XML processing +// instructions. +type Token interface { + Parent() *Element + Index() int + WriteTo(w Writer, s *WriteSettings) + dup(parent *Element) Token + setParent(parent *Element) + setIndex(index int) +} + +// A Document is a container holding a complete XML tree. +// +// A document has a single embedded element, which contains zero or more child +// tokens, one of which is usually the root element. The embedded element may +// include other children such as processing instruction tokens or character +// data tokens. The document's embedded element is never directly serialized; +// only its children are. +// +// A document also contains read and write settings, which influence the way +// the document is deserialized, serialized, and indented. +type Document struct { + Element + ReadSettings ReadSettings + WriteSettings WriteSettings +} + +// An Element represents an XML element, its attributes, and its child tokens. +type Element struct { + Space, Tag string // namespace prefix and tag + Attr []Attr // key-value attribute pairs + Child []Token // child tokens (elements, comments, etc.) + parent *Element // parent element + index int // token index in parent's children +} + +// An Attr represents a key-value attribute within an XML element. +type Attr struct { + Space, Key string // The attribute's namespace prefix and key + Value string // The attribute value string + element *Element // element containing the attribute +} + +// charDataFlags are used with CharData tokens to store additional settings. +type charDataFlags uint8 + +const ( + // The CharData contains only whitespace. + whitespaceFlag charDataFlags = 1 << iota + + // The CharData contains a CDATA section. + cdataFlag +) + +// CharData may be used to represent simple text data or a CDATA section +// within an XML document. The Data property should never be modified +// directly; use the SetData function instead. +type CharData struct { + Data string // the simple text or CDATA section content + parent *Element + index int + flags charDataFlags +} + +// A Comment represents an XML comment. +type Comment struct { + Data string // the comment's text + parent *Element + index int +} + +// A Directive represents an XML directive. +type Directive struct { + Data string // the directive string + parent *Element + index int +} + +// A ProcInst represents an XML processing instruction. +type ProcInst struct { + Target string // the processing instruction target + Inst string // the processing instruction value + parent *Element + index int +} + +// NewDocument creates an XML document without a root element. +func NewDocument() *Document { + return &Document{ + Element: Element{Child: make([]Token, 0)}, + } +} + +// NewDocumentWithRoot creates an XML document and sets the element 'e' as its +// root element. If the element 'e' is already part of another document, it is +// first removed from its existing document. +func NewDocumentWithRoot(e *Element) *Document { + d := NewDocument() + d.SetRoot(e) + return d +} + +// Copy returns a recursive, deep copy of the document. +func (d *Document) Copy() *Document { + return &Document{ + Element: *(d.Element.dup(nil).(*Element)), + ReadSettings: d.ReadSettings.dup(), + WriteSettings: d.WriteSettings.dup(), + } +} + +// Root returns the root element of the document. It returns nil if there is +// no root element. +func (d *Document) Root() *Element { + for _, t := range d.Child { + if c, ok := t.(*Element); ok { + return c + } + } + return nil +} + +// SetRoot replaces the document's root element with the element 'e'. If the +// document already has a root element when this function is called, then the +// existing root element is unbound from the document. If the element 'e' is +// part of another document, then it is unbound from the other document. +func (d *Document) SetRoot(e *Element) { + if e.parent != nil { + e.parent.RemoveChild(e) + } + + // If there is already a root element, replace it. + p := &d.Element + for i, t := range p.Child { + if _, ok := t.(*Element); ok { + t.setParent(nil) + t.setIndex(-1) + p.Child[i] = e + e.setParent(p) + e.setIndex(i) + return + } + } + + // No existing root element, so add it. + p.addChild(e) +} + +// ReadFrom reads XML from the reader 'r' into this document. The function +// returns the number of bytes read and any error encountered. +func (d *Document) ReadFrom(r io.Reader) (n int64, err error) { + if d.ReadSettings.ValidateInput { + b, err := io.ReadAll(r) + if err != nil { + return 0, err + } + if err := validateXML(bytes.NewReader(b), d.ReadSettings); err != nil { + return 0, err + } + r = bytes.NewReader(b) + } + return d.Element.readFrom(r, d.ReadSettings) +} + +// ReadFromFile reads XML from a local file at path 'filepath' into this +// document. +func (d *Document) ReadFromFile(filepath string) error { + f, err := os.Open(filepath) + if err != nil { + return err + } + defer f.Close() + + _, err = d.ReadFrom(f) + return err +} + +// ReadFromBytes reads XML from the byte slice 'b' into the this document. +func (d *Document) ReadFromBytes(b []byte) error { + if d.ReadSettings.ValidateInput { + if err := validateXML(bytes.NewReader(b), d.ReadSettings); err != nil { + return err + } + } + _, err := d.Element.readFrom(bytes.NewReader(b), d.ReadSettings) + return err +} + +// ReadFromString reads XML from the string 's' into this document. +func (d *Document) ReadFromString(s string) error { + if d.ReadSettings.ValidateInput { + if err := validateXML(strings.NewReader(s), d.ReadSettings); err != nil { + return err + } + } + _, err := d.Element.readFrom(strings.NewReader(s), d.ReadSettings) + return err +} + +// validateXML determines if the data read from the reader 'r' contains +// well-formed XML according to the rules set by the go xml package. +func validateXML(r io.Reader, settings ReadSettings) error { + dec := newDecoder(r, settings) + err := dec.Decode(new(interface{})) + if err != nil { + return err + } + + // If there are any trailing tokens after unmarshalling with Decode(), + // then the XML input didn't terminate properly. + _, err = dec.Token() + if err == io.EOF { + return nil + } + return ErrXML +} + +// newDecoder creates an XML decoder for the reader 'r' configured using +// the provided read settings. +func newDecoder(r io.Reader, settings ReadSettings) *xml.Decoder { + d := xml.NewDecoder(r) + d.CharsetReader = settings.CharsetReader + if d.CharsetReader == nil { + d.CharsetReader = defaultCharsetReader + } + d.Strict = !settings.Permissive + d.Entity = settings.Entity + d.AutoClose = settings.AutoClose + return d +} + +// WriteTo serializes the document out to the writer 'w'. The function returns +// the number of bytes written and any error encountered. +func (d *Document) WriteTo(w io.Writer) (n int64, err error) { + xw := newXmlWriter(w) + b := bufio.NewWriter(xw) + for _, c := range d.Child { + c.WriteTo(b, &d.WriteSettings) + } + err, n = b.Flush(), xw.bytes + return +} + +// WriteToFile serializes the document out to the file at path 'filepath'. +func (d *Document) WriteToFile(filepath string) error { + f, err := os.Create(filepath) + if err != nil { + return err + } + defer f.Close() + _, err = d.WriteTo(f) + return err +} + +// WriteToBytes serializes this document into a slice of bytes. +func (d *Document) WriteToBytes() (b []byte, err error) { + var buf bytes.Buffer + if _, err = d.WriteTo(&buf); err != nil { + return + } + return buf.Bytes(), nil +} + +// WriteToString serializes this document into a string. +func (d *Document) WriteToString() (s string, err error) { + var b []byte + if b, err = d.WriteToBytes(); err != nil { + return + } + return string(b), nil +} + +// Indent modifies the document's element tree by inserting character data +// tokens containing newlines and spaces for indentation. The amount of +// indentation per depth level is given by the 'spaces' parameter. Other than +// the number of spaces, default IndentSettings are used. +func (d *Document) Indent(spaces int) { + s := NewIndentSettings() + s.Spaces = spaces + d.IndentWithSettings(s) +} + +// IndentTabs modifies the document's element tree by inserting CharData +// tokens containing newlines and tabs for indentation. One tab is used per +// indentation level. Other than the use of tabs, default IndentSettings +// are used. +func (d *Document) IndentTabs() { + s := NewIndentSettings() + s.UseTabs = true + d.IndentWithSettings(s) +} + +// IndentWithSettings modifies the document's element tree by inserting +// character data tokens containing newlines and indentation. The behavior +// of the indentation algorithm is configured by the indent settings. +func (d *Document) IndentWithSettings(s *IndentSettings) { + // WriteSettings.UseCRLF is deprecated. Until removed from the package, it + // overrides IndentSettings.UseCRLF when true. + if d.WriteSettings.UseCRLF { + s.UseCRLF = true + } + + d.Element.indent(0, getIndentFunc(s), s) + + if s.SuppressTrailingWhitespace { + d.Element.stripTrailingWhitespace() + } +} + +// Unindent modifies the document's element tree by removing character data +// tokens containing only whitespace. Other than the removal of indentation, +// default IndentSettings are used. +func (d *Document) Unindent() { + s := NewIndentSettings() + s.Spaces = NoIndent + d.IndentWithSettings(s) +} + +// NewElement creates an unparented element with the specified tag (i.e., +// name). The tag may include a namespace prefix followed by a colon. +func NewElement(tag string) *Element { + space, stag := spaceDecompose(tag) + return newElement(space, stag, nil) +} + +// newElement is a helper function that creates an element and binds it to +// a parent element if possible. +func newElement(space, tag string, parent *Element) *Element { + e := &Element{ + Space: space, + Tag: tag, + Attr: make([]Attr, 0), + Child: make([]Token, 0), + parent: parent, + index: -1, + } + if parent != nil { + parent.addChild(e) + } + return e +} + +// Copy creates a recursive, deep copy of the element and all its attributes +// and children. The returned element has no parent but can be parented to a +// another element using AddChild, or added to a document with SetRoot or +// NewDocumentWithRoot. +func (e *Element) Copy() *Element { + return e.dup(nil).(*Element) +} + +// FullTag returns the element e's complete tag, including namespace prefix if +// present. +func (e *Element) FullTag() string { + if e.Space == "" { + return e.Tag + } + return e.Space + ":" + e.Tag +} + +// NamespaceURI returns the XML namespace URI associated with the element. If +// the element is part of the XML default namespace, NamespaceURI returns the +// empty string. +func (e *Element) NamespaceURI() string { + if e.Space == "" { + return e.findDefaultNamespaceURI() + } + return e.findLocalNamespaceURI(e.Space) +} + +// findLocalNamespaceURI finds the namespace URI corresponding to the +// requested prefix. +func (e *Element) findLocalNamespaceURI(prefix string) string { + for _, a := range e.Attr { + if a.Space == "xmlns" && a.Key == prefix { + return a.Value + } + } + + if e.parent == nil { + return "" + } + + return e.parent.findLocalNamespaceURI(prefix) +} + +// findDefaultNamespaceURI finds the default namespace URI of the element. +func (e *Element) findDefaultNamespaceURI() string { + for _, a := range e.Attr { + if a.Space == "" && a.Key == "xmlns" { + return a.Value + } + } + + if e.parent == nil { + return "" + } + + return e.parent.findDefaultNamespaceURI() +} + +// namespacePrefix returns the namespace prefix associated with the element. +func (e *Element) namespacePrefix() string { + return e.Space +} + +// name returns the tag associated with the element. +func (e *Element) name() string { + return e.Tag +} + +// ReindexChildren recalculates the index values of the element's child +// tokens. This is necessary only if you have manually manipulated the +// element's `Child` array. +func (e *Element) ReindexChildren() { + for i := 0; i < len(e.Child); i++ { + e.Child[i].setIndex(i) + } +} + +// Text returns all character data immediately following the element's opening +// tag. +func (e *Element) Text() string { + if len(e.Child) == 0 { + return "" + } + + text := "" + for _, ch := range e.Child { + if cd, ok := ch.(*CharData); ok { + if text == "" { + text = cd.Data + } else { + text += cd.Data + } + } else if _, ok := ch.(*Comment); ok { + // ignore + } else { + break + } + } + return text +} + +// SetText replaces all character data immediately following an element's +// opening tag with the requested string. +func (e *Element) SetText(text string) { + e.replaceText(0, text, 0) +} + +// SetCData replaces all character data immediately following an element's +// opening tag with a CDATA section. +func (e *Element) SetCData(text string) { + e.replaceText(0, text, cdataFlag) +} + +// Tail returns all character data immediately following the element's end +// tag. +func (e *Element) Tail() string { + if e.Parent() == nil { + return "" + } + + p := e.Parent() + i := e.Index() + + text := "" + for _, ch := range p.Child[i+1:] { + if cd, ok := ch.(*CharData); ok { + if text == "" { + text = cd.Data + } else { + text += cd.Data + } + } else { + break + } + } + return text +} + +// SetTail replaces all character data immediately following the element's end +// tag with the requested string. +func (e *Element) SetTail(text string) { + if e.Parent() == nil { + return + } + + p := e.Parent() + p.replaceText(e.Index()+1, text, 0) +} + +// replaceText is a helper function that replaces a series of chardata tokens +// starting at index i with the requested text. +func (e *Element) replaceText(i int, text string, flags charDataFlags) { + end := e.findTermCharDataIndex(i) + + switch { + case end == i: + if text != "" { + // insert a new chardata token at index i + cd := newCharData(text, flags, nil) + e.InsertChildAt(i, cd) + } + + case end == i+1: + if text == "" { + // remove the chardata token at index i + e.RemoveChildAt(i) + } else { + // replace the first and only character token at index i + cd := e.Child[i].(*CharData) + cd.Data, cd.flags = text, flags + } + + default: + if text == "" { + // remove all chardata tokens starting from index i + copy(e.Child[i:], e.Child[end:]) + removed := end - i + e.Child = e.Child[:len(e.Child)-removed] + for j := i; j < len(e.Child); j++ { + e.Child[j].setIndex(j) + } + } else { + // replace the first chardata token at index i and remove all + // subsequent chardata tokens + cd := e.Child[i].(*CharData) + cd.Data, cd.flags = text, flags + copy(e.Child[i+1:], e.Child[end:]) + removed := end - (i + 1) + e.Child = e.Child[:len(e.Child)-removed] + for j := i + 1; j < len(e.Child); j++ { + e.Child[j].setIndex(j) + } + } + } +} + +// findTermCharDataIndex finds the index of the first child token that isn't +// a CharData token. It starts from the requested start index. +func (e *Element) findTermCharDataIndex(start int) int { + for i := start; i < len(e.Child); i++ { + if _, ok := e.Child[i].(*CharData); !ok { + return i + } + } + return len(e.Child) +} + +// CreateElement creates a new element with the specified tag (i.e., name) and +// adds it as the last child of element 'e'. The tag may include a prefix +// followed by a colon. +func (e *Element) CreateElement(tag string) *Element { + space, stag := spaceDecompose(tag) + return newElement(space, stag, e) +} + +// CreateChild performs the same task as CreateElement but calls a +// continuation function after the child element is created, allowing +// additional actions to be performed on the child element before returning. +// +// This method of element creation is particularly useful when building nested +// XML documents from code. For example: +// +// org := doc.CreateChild("organization", func(e *Element) { +// e.CreateComment("Mary") +// e.CreateChild("person", func(e *Element) { +// e.CreateAttr("name", "Mary") +// e.CreateAttr("age", "30") +// e.CreateAttr("hair", "brown") +// }) +// }) +func (e *Element) CreateChild(tag string, cont func(e *Element)) *Element { + child := e.CreateElement(tag) + cont(child) + return child +} + +// AddChild adds the token 't' as the last child of the element. If token 't' +// was already the child of another element, it is first removed from its +// parent element. +func (e *Element) AddChild(t Token) { + if t.Parent() != nil { + t.Parent().RemoveChild(t) + } + e.addChild(t) +} + +// InsertChild inserts the token 't' into this element's list of children just +// before the element's existing child token 'ex'. If the existing element +// 'ex' does not appear in this element's list of child tokens, then 't' is +// added to the end of this element's list of child tokens. If token 't' is +// already the child of another element, it is first removed from the other +// element's list of child tokens. +// +// Deprecated: InsertChild is deprecated. Use InsertChildAt instead. +func (e *Element) InsertChild(ex Token, t Token) { + if ex == nil || ex.Parent() != e { + e.AddChild(t) + return + } + + if t.Parent() != nil { + t.Parent().RemoveChild(t) + } + + t.setParent(e) + + i := ex.Index() + e.Child = append(e.Child, nil) + copy(e.Child[i+1:], e.Child[i:]) + e.Child[i] = t + + for j := i; j < len(e.Child); j++ { + e.Child[j].setIndex(j) + } +} + +// InsertChildAt inserts the token 't' into this element's list of child +// tokens just before the requested 'index'. If the index is greater than or +// equal to the length of the list of child tokens, then the token 't' is +// added to the end of the list of child tokens. +func (e *Element) InsertChildAt(index int, t Token) { + if index >= len(e.Child) { + e.AddChild(t) + return + } + + if t.Parent() != nil { + if t.Parent() == e && t.Index() < index { + index-- + } + t.Parent().RemoveChild(t) + } + + t.setParent(e) + + e.Child = append(e.Child, nil) + copy(e.Child[index+1:], e.Child[index:]) + e.Child[index] = t + + for j := index; j < len(e.Child); j++ { + e.Child[j].setIndex(j) + } +} + +// RemoveChild attempts to remove the token 't' from this element's list of +// child tokens. If the token 't' was a child of this element, then it is +// removed and returned. Otherwise, nil is returned. +func (e *Element) RemoveChild(t Token) Token { + if t.Parent() != e { + return nil + } + return e.RemoveChildAt(t.Index()) +} + +// RemoveChildAt removes the child token appearing in slot 'index' of this +// element's list of child tokens. The removed child token is then returned. +// If the index is out of bounds, no child is removed and nil is returned. +func (e *Element) RemoveChildAt(index int) Token { + if index >= len(e.Child) { + return nil + } + + t := e.Child[index] + for j := index + 1; j < len(e.Child); j++ { + e.Child[j].setIndex(j - 1) + } + e.Child = append(e.Child[:index], e.Child[index+1:]...) + t.setIndex(-1) + t.setParent(nil) + return t +} + +// autoClose analyzes the stack's top element and the current token to decide +// whether the top element should be closed. +func (e *Element) autoClose(stack *stack[*Element], t xml.Token, tags []string) { + if stack.empty() { + return + } + + top := stack.peek() + + for _, tag := range tags { + if strings.EqualFold(tag, top.FullTag()) { + if e, ok := t.(xml.EndElement); !ok || + !strings.EqualFold(e.Name.Space, top.Space) || + !strings.EqualFold(e.Name.Local, top.Tag) { + stack.pop() + } + break + } + } +} + +// ReadFrom reads XML from the reader 'ri' and stores the result as a new +// child of this element. +func (e *Element) readFrom(ri io.Reader, settings ReadSettings) (n int64, err error) { + var r xmlReader + var pr *xmlPeekReader + if settings.PreserveCData { + pr = newXmlPeekReader(ri) + r = pr + } else { + r = newXmlSimpleReader(ri) + } + + attrCheck := make(map[xml.Name]int) + dec := newDecoder(r, settings) + + var stack stack[*Element] + stack.push(e) + for { + if pr != nil { + pr.PeekPrepare(dec.InputOffset(), len(cdataPrefix)) + } + + t, err := dec.RawToken() + + if settings.Permissive && settings.AutoClose != nil { + e.autoClose(&stack, t, settings.AutoClose) + } + + switch { + case err == io.EOF: + if len(stack.data) != 1 { + return r.Bytes(), ErrXML + } + return r.Bytes(), nil + case err != nil: + return r.Bytes(), err + case stack.empty(): + return r.Bytes(), ErrXML + } + + top := stack.peek() + + switch t := t.(type) { + case xml.StartElement: + e := newElement(t.Name.Space, t.Name.Local, top) + if settings.PreserveDuplicateAttrs || len(t.Attr) < 2 { + for _, a := range t.Attr { + e.addAttr(a.Name.Space, a.Name.Local, a.Value) + } + } else { + for _, a := range t.Attr { + if i, contains := attrCheck[a.Name]; contains { + e.Attr[i].Value = a.Value + } else { + attrCheck[a.Name] = e.addAttr(a.Name.Space, a.Name.Local, a.Value) + } + } + clear(attrCheck) + } + stack.push(e) + case xml.EndElement: + if top.Tag != t.Name.Local || top.Space != t.Name.Space { + return r.Bytes(), ErrXML + } + stack.pop() + case xml.CharData: + data := string(t) + var flags charDataFlags + if pr != nil { + peekBuf := pr.PeekFinalize() + if bytes.Equal(peekBuf, cdataPrefix) { + flags = cdataFlag + } else if isWhitespace(data) { + flags = whitespaceFlag + } + } else { + if isWhitespace(data) { + flags = whitespaceFlag + } + } + newCharData(data, flags, top) + case xml.Comment: + newComment(string(t), top) + case xml.Directive: + newDirective(string(t), top) + case xml.ProcInst: + newProcInst(t.Target, string(t.Inst), top) + } + } +} + +// SelectAttr finds an element attribute matching the requested 'key' and, if +// found, returns a pointer to the matching attribute. The function returns +// nil if no matching attribute is found. The key may include a namespace +// prefix followed by a colon. +func (e *Element) SelectAttr(key string) *Attr { + space, skey := spaceDecompose(key) + for i, a := range e.Attr { + if spaceMatch(space, a.Space) && skey == a.Key { + return &e.Attr[i] + } + } + return nil +} + +// SelectAttrValue finds an element attribute matching the requested 'key' and +// returns its value if found. If no matching attribute is found, the function +// returns the 'dflt' value instead. The key may include a namespace prefix +// followed by a colon. +func (e *Element) SelectAttrValue(key, dflt string) string { + space, skey := spaceDecompose(key) + for _, a := range e.Attr { + if spaceMatch(space, a.Space) && skey == a.Key { + return a.Value + } + } + return dflt +} + +// ChildElements returns all elements that are children of this element. +func (e *Element) ChildElements() []*Element { + return slices.Collect(e.ChildElementsSeq()) +} + +// ChildElementsSeq returns an iterator over all child elements of this +// element. +func (e *Element) ChildElementsSeq() iter.Seq[*Element] { + return func(yield func(*Element) bool) { + for _, t := range e.Child { + if c, ok := t.(*Element); ok { + if !yield(c) { + return + } + } + } + } +} + +// SelectElement returns the first child element with the given 'tag' (i.e., +// name). The function returns nil if no child element matching the tag is +// found. The tag may include a namespace prefix followed by a colon. +func (e *Element) SelectElement(tag string) *Element { + for element := range e.SelectElementsSeq(tag) { + return element + } + return nil +} + +// SelectElements returns a slice of all child elements with the given 'tag' +// (i.e., name). The tag may include a namespace prefix followed by a colon. +func (e *Element) SelectElements(tag string) []*Element { + return slices.Collect(e.SelectElementsSeq(tag)) +} + +// SelectElementsSeq returns an iterator over all child elements with the +// given 'tag' (i.e., name). The tag may include a namespace prefix followed +// by a colon. +func (e *Element) SelectElementsSeq(tag string) iter.Seq[*Element] { + return func(yield func(*Element) bool) { + space, stag := spaceDecompose(tag) + for _, t := range e.Child { + if c, ok := t.(*Element); ok && spaceMatch(space, c.Space) && stag == c.Tag { + if !yield(c) { + return + } + } + } + } +} + +// FindElement returns the first element matched by the XPath-like 'path' +// string. The function returns nil if no child element is found using the +// path. It panics if an invalid path string is supplied. +func (e *Element) FindElement(path string) *Element { + return e.FindElementPath(MustCompilePath(path)) +} + +// FindElementPath returns the first element matched by the 'path' object. The +// function returns nil if no element is found using the path. +func (e *Element) FindElementPath(path Path) *Element { + for element := range path.traverse(e) { + return element + } + return nil +} + +// FindElements returns a slice of elements matched by the XPath-like 'path' +// string. The function returns nil if no child element is found using the +// path. It panics if an invalid path string is supplied. +func (e *Element) FindElements(path string) []*Element { + return slices.Collect(e.FindElementsSeq(path)) +} + +// FindElementsSeq returns an iterator over elements matched by the XPath-like +// 'path' string. This function uses Go's iterator support for +// memory-efficient traversal. It panics if an invalid path string is +// supplied. +func (e *Element) FindElementsSeq(path string) iter.Seq[*Element] { + return e.FindElementsPathSeq(MustCompilePath(path)) +} + +// FindElementsPath returns a slice of elements matched by the 'path' object. +func (e *Element) FindElementsPath(path Path) []*Element { + return slices.Collect(e.FindElementsPathSeq(path)) +} + +// FindElementsPathSeq returns an iterator over elements matched by the 'path' +// object. +func (e *Element) FindElementsPathSeq(path Path) iter.Seq[*Element] { + return path.traverse(e) +} + +// NotNil returns the receiver element if it isn't nil; otherwise, it returns +// an unparented element with an empty string tag. This function simplifies +// the task of writing code to ignore not-found results from element queries. +// For example, instead of writing this: +// +// if e := doc.SelectElement("enabled"); e != nil { +// e.SetText("true") +// } +// +// You could write this: +// +// doc.SelectElement("enabled").NotNil().SetText("true") +func (e *Element) NotNil() *Element { + if e == nil { + return NewElement("") + } + return e +} + +// GetPath returns the absolute path of the element. The absolute path is the +// full path from the document's root. +func (e *Element) GetPath() string { + path := []string{} + for seg := e; seg != nil; seg = seg.Parent() { + if seg.Tag != "" { + path = append(path, seg.Tag) + } + } + + // Reverse the path. + for i, j := 0, len(path)-1; i < j; i, j = i+1, j-1 { + path[i], path[j] = path[j], path[i] + } + + return "/" + strings.Join(path, "/") +} + +// GetRelativePath returns the path of this element relative to the 'source' +// element. If the two elements are not part of the same element tree, then +// the function returns the empty string. +func (e *Element) GetRelativePath(source *Element) string { + var path []*Element + + if source == nil { + return "" + } + + // Build a reverse path from the element toward the root. Stop if the + // source element is encountered. + var seg *Element + for seg = e; seg != nil && seg != source; seg = seg.Parent() { + path = append(path, seg) + } + + // If we found the source element, reverse the path and compose the + // string. + if seg == source { + if len(path) == 0 { + return "." + } + parts := []string{} + for i := len(path) - 1; i >= 0; i-- { + parts = append(parts, path[i].Tag) + } + return "./" + strings.Join(parts, "/") + } + + // The source wasn't encountered, so climb from the source element toward + // the root of the tree until an element in the reversed path is + // encountered. + + findPathIndex := func(e *Element, path []*Element) int { + for i, ee := range path { + if e == ee { + return i + } + } + return -1 + } + + climb := 0 + for seg = source; seg != nil; seg = seg.Parent() { + i := findPathIndex(seg, path) + if i >= 0 { + path = path[:i] // truncate at found segment + break + } + climb++ + } + + // No element in the reversed path was encountered, so the two elements + // must not be part of the same tree. + if seg == nil { + return "" + } + + // Reverse the (possibly truncated) path and prepend ".." segments to + // climb. + parts := []string{} + for i := 0; i < climb; i++ { + parts = append(parts, "..") + } + for i := len(path) - 1; i >= 0; i-- { + parts = append(parts, path[i].Tag) + } + return strings.Join(parts, "/") +} + +// IndentWithSettings modifies the element and its child tree by inserting +// character data tokens containing newlines and indentation. The behavior of +// the indentation algorithm is configured by the indent settings. Because +// this function indents the element as if it were at the root of a document, +// it is most useful when called just before writing the element as an XML +// fragment using WriteTo. +func (e *Element) IndentWithSettings(s *IndentSettings) { + e.indent(1, getIndentFunc(s), s) +} + +// indent recursively inserts proper indentation between an XML element's +// child tokens. +func (e *Element) indent(depth int, indent indentFunc, s *IndentSettings) { + e.stripIndent(s) + n := len(e.Child) + if n == 0 { + return + } + + oldChild := e.Child + e.Child = make([]Token, 0, n*2+1) + isCharData, firstNonCharData := false, true + for _, c := range oldChild { + // Insert NL+indent before child if it's not character data. + // Exceptions: when it's the first non-character-data child, or when + // the child is at root depth. + _, isCharData = c.(*CharData) + if !isCharData { + if !firstNonCharData || depth > 0 { + s := indent(depth) + if s != "" { + newCharData(s, whitespaceFlag, e) + } + } + firstNonCharData = false + } + + e.addChild(c) + + // Recursively process child elements. + if ce, ok := c.(*Element); ok { + ce.indent(depth+1, indent, s) + } + } + + // Insert NL+indent before the last child. + if !isCharData { + if !firstNonCharData || depth > 0 { + s := indent(depth - 1) + if s != "" { + newCharData(s, whitespaceFlag, e) + } + } + } +} + +// stripIndent removes any previously inserted indentation. +func (e *Element) stripIndent(s *IndentSettings) { + // Count the number of non-indent child tokens + n := len(e.Child) + for _, c := range e.Child { + if cd, ok := c.(*CharData); ok && cd.IsWhitespace() { + n-- + } + } + if n == len(e.Child) { + return + } + if n == 0 && len(e.Child) == 1 && s.PreserveLeafWhitespace { + return + } + + // Strip out indent CharData + newChild := make([]Token, n) + j := 0 + for _, c := range e.Child { + if cd, ok := c.(*CharData); ok && cd.IsWhitespace() { + continue + } + newChild[j] = c + newChild[j].setIndex(j) + j++ + } + e.Child = newChild +} + +// stripTrailingWhitespace removes any trailing whitespace CharData tokens +// from the element's children. +func (e *Element) stripTrailingWhitespace() { + for i := len(e.Child) - 1; i >= 0; i-- { + if cd, ok := e.Child[i].(*CharData); !ok || !cd.IsWhitespace() { + e.Child = e.Child[:i+1] + return + } + } +} + +// dup duplicates the element. +func (e *Element) dup(parent *Element) Token { + ne := &Element{ + Space: e.Space, + Tag: e.Tag, + Attr: make([]Attr, len(e.Attr)), + Child: make([]Token, len(e.Child)), + parent: parent, + index: e.index, + } + for i, t := range e.Child { + ne.Child[i] = t.dup(ne) + } + copy(ne.Attr, e.Attr) + return ne +} + +// NextSibling returns this element's next sibling element. It returns nil if +// there is no next sibling element. +func (e *Element) NextSibling() *Element { + if e.parent == nil { + return nil + } + for i := e.index + 1; i < len(e.parent.Child); i++ { + if s, ok := e.parent.Child[i].(*Element); ok { + return s + } + } + return nil +} + +// PrevSibling returns this element's preceding sibling element. It returns +// nil if there is no preceding sibling element. +func (e *Element) PrevSibling() *Element { + if e.parent == nil { + return nil + } + for i := e.index - 1; i >= 0; i-- { + if s, ok := e.parent.Child[i].(*Element); ok { + return s + } + } + return nil +} + +// Parent returns this element's parent element. It returns nil if this +// element has no parent. +func (e *Element) Parent() *Element { + return e.parent +} + +// Index returns the index of this element within its parent element's +// list of child tokens. If this element has no parent, then the function +// returns -1. +func (e *Element) Index() int { + return e.index +} + +// WriteTo serializes the element to the writer w. +func (e *Element) WriteTo(w Writer, s *WriteSettings) { + w.WriteByte('<') + w.WriteString(e.FullTag()) + for _, a := range e.Attr { + w.WriteByte(' ') + a.WriteTo(w, s) + } + if len(e.Child) > 0 { + w.WriteByte('>') + for _, c := range e.Child { + c.WriteTo(w, s) + } + w.Write([]byte{'<', '/'}) + w.WriteString(e.FullTag()) + w.WriteByte('>') + } else { + if s.CanonicalEndTags { + w.Write([]byte{'>', '<', '/'}) + w.WriteString(e.FullTag()) + w.WriteByte('>') + } else { + w.Write([]byte{'/', '>'}) + } + } +} + +// setParent replaces this element token's parent. +func (e *Element) setParent(parent *Element) { + e.parent = parent +} + +// setIndex sets this element token's index within its parent's Child slice. +func (e *Element) setIndex(index int) { + e.index = index +} + +// addChild adds a child token to the element e. +func (e *Element) addChild(t Token) { + t.setParent(e) + t.setIndex(len(e.Child)) + e.Child = append(e.Child, t) +} + +// CreateAttr creates an attribute with the specified 'key' and 'value' and +// adds it to this element. If an attribute with same key already exists on +// this element, then its value is replaced. The key may include a namespace +// prefix followed by a colon. +func (e *Element) CreateAttr(key, value string) *Attr { + space, skey := spaceDecompose(key) + + for i, a := range e.Attr { + if space == a.Space && skey == a.Key { + e.Attr[i].Value = value + return &e.Attr[i] + } + } + + i := e.addAttr(space, skey, value) + return &e.Attr[i] +} + +// addAttr is a helper function that adds an attribute to an element. Returns +// the index of the added attribute. +func (e *Element) addAttr(space, key, value string) int { + a := Attr{ + Space: space, + Key: key, + Value: value, + element: e, + } + e.Attr = append(e.Attr, a) + return len(e.Attr) - 1 +} + +// RemoveAttr removes the first attribute of this element whose key matches +// 'key'. It returns a copy of the removed attribute if a match is found. If +// no match is found, it returns nil. The key may include a namespace prefix +// followed by a colon. +func (e *Element) RemoveAttr(key string) *Attr { + space, skey := spaceDecompose(key) + for i, a := range e.Attr { + if space == a.Space && skey == a.Key { + e.Attr = append(e.Attr[0:i], e.Attr[i+1:]...) + return &Attr{ + Space: a.Space, + Key: a.Key, + Value: a.Value, + element: nil, + } + } + } + return nil +} + +// SortAttrs sorts this element's attributes lexicographically by key. +func (e *Element) SortAttrs() { + slices.SortFunc(e.Attr, func(a, b Attr) int { + if v := strings.Compare(a.Space, b.Space); v != 0 { + return v + } + return strings.Compare(a.Key, b.Key) + }) +} + +// FullKey returns this attribute's complete key, including namespace prefix +// if present. +func (a *Attr) FullKey() string { + if a.Space == "" { + return a.Key + } + return a.Space + ":" + a.Key +} + +// Element returns a pointer to the element containing this attribute. +func (a *Attr) Element() *Element { + return a.element +} + +// NamespaceURI returns the XML namespace URI associated with this attribute. +// The function returns the empty string if the attribute is unprefixed or +// if the attribute is part of the XML default namespace. +func (a *Attr) NamespaceURI() string { + if a.Space == "" { + return "" + } + return a.element.findLocalNamespaceURI(a.Space) +} + +// WriteTo serializes the attribute to the writer. +func (a *Attr) WriteTo(w Writer, s *WriteSettings) { + w.WriteString(a.FullKey()) + if s.AttrSingleQuote { + w.WriteString(`='`) + } else { + w.WriteString(`="`) + } + var m escapeMode + if s.CanonicalAttrVal && !s.AttrSingleQuote { + m = escapeCanonicalAttr + } else { + m = escapeNormal + } + escapeString(w, a.Value, m) + if s.AttrSingleQuote { + w.WriteByte('\'') + } else { + w.WriteByte('"') + } +} + +// NewText creates an unparented CharData token containing simple text data. +func NewText(text string) *CharData { + return newCharData(text, 0, nil) +} + +// NewCData creates an unparented XML character CDATA section with 'data' as +// its content. +func NewCData(data string) *CharData { + return newCharData(data, cdataFlag, nil) +} + +// NewCharData creates an unparented CharData token containing simple text +// data. +// +// Deprecated: NewCharData is deprecated. Instead, use NewText, which does the +// same thing. +func NewCharData(data string) *CharData { + return newCharData(data, 0, nil) +} + +// newCharData creates a character data token and binds it to a parent +// element. If parent is nil, the CharData token remains unbound. +func newCharData(data string, flags charDataFlags, parent *Element) *CharData { + c := &CharData{ + Data: data, + parent: nil, + index: -1, + flags: flags, + } + if parent != nil { + parent.addChild(c) + } + return c +} + +// CreateText creates a CharData token containing simple text data and adds it +// to the end of this element's list of child tokens. +func (e *Element) CreateText(text string) *CharData { + return newCharData(text, 0, e) +} + +// CreateCData creates a CharData token containing a CDATA section with 'data' +// as its content and adds it to the end of this element's list of child +// tokens. +func (e *Element) CreateCData(data string) *CharData { + return newCharData(data, cdataFlag, e) +} + +// CreateCharData creates a CharData token containing simple text data and +// adds it to the end of this element's list of child tokens. +// +// Deprecated: CreateCharData is deprecated. Instead, use CreateText, which +// does the same thing. +func (e *Element) CreateCharData(data string) *CharData { + return e.CreateText(data) +} + +// SetData modifies the content of the CharData token. In the case of a +// CharData token containing simple text, the simple text is modified. In the +// case of a CharData token containing a CDATA section, the CDATA section's +// content is modified. +func (c *CharData) SetData(text string) { + c.Data = text + if isWhitespace(text) { + c.flags |= whitespaceFlag + } else { + c.flags &= ^whitespaceFlag + } +} + +// IsCData returns true if this CharData token is contains a CDATA section. It +// returns false if the CharData token contains simple text. +func (c *CharData) IsCData() bool { + return (c.flags & cdataFlag) != 0 +} + +// IsWhitespace returns true if this CharData token contains only whitespace. +func (c *CharData) IsWhitespace() bool { + return (c.flags & whitespaceFlag) != 0 +} + +// Parent returns this CharData token's parent element, or nil if it has no +// parent. +func (c *CharData) Parent() *Element { + return c.parent +} + +// Index returns the index of this CharData token within its parent element's +// list of child tokens. If this CharData token has no parent, then the +// function returns -1. +func (c *CharData) Index() int { + return c.index +} + +// WriteTo serializes character data to the writer. +func (c *CharData) WriteTo(w Writer, s *WriteSettings) { + if c.IsCData() { + w.WriteString(`<![CDATA[`) + w.WriteString(c.Data) + w.WriteString(`]]>`) + } else { + var m escapeMode + if s.CanonicalText { + m = escapeCanonicalText + } else { + m = escapeNormal + } + escapeString(w, c.Data, m) + } +} + +// dup duplicates the character data. +func (c *CharData) dup(parent *Element) Token { + return &CharData{ + Data: c.Data, + flags: c.flags, + parent: parent, + index: c.index, + } +} + +// setParent replaces the character data token's parent. +func (c *CharData) setParent(parent *Element) { + c.parent = parent +} + +// setIndex sets the CharData token's index within its parent element's Child +// slice. +func (c *CharData) setIndex(index int) { + c.index = index +} + +// NewComment creates an unparented comment token. +func NewComment(comment string) *Comment { + return newComment(comment, nil) +} + +// NewComment creates a comment token and sets its parent element to 'parent'. +func newComment(comment string, parent *Element) *Comment { + c := &Comment{ + Data: comment, + parent: nil, + index: -1, + } + if parent != nil { + parent.addChild(c) + } + return c +} + +// CreateComment creates a comment token using the specified 'comment' string +// and adds it as the last child token of this element. +func (e *Element) CreateComment(comment string) *Comment { + return newComment(comment, e) +} + +// dup duplicates the comment. +func (c *Comment) dup(parent *Element) Token { + return &Comment{ + Data: c.Data, + parent: parent, + index: c.index, + } +} + +// Parent returns comment token's parent element, or nil if it has no parent. +func (c *Comment) Parent() *Element { + return c.parent +} + +// Index returns the index of this Comment token within its parent element's +// list of child tokens. If this Comment token has no parent, then the +// function returns -1. +func (c *Comment) Index() int { + return c.index +} + +// WriteTo serialies the comment to the writer. +func (c *Comment) WriteTo(w Writer, s *WriteSettings) { + w.WriteString("<!--") + w.WriteString(c.Data) + w.WriteString("-->") +} + +// setParent replaces the comment token's parent. +func (c *Comment) setParent(parent *Element) { + c.parent = parent +} + +// setIndex sets the Comment token's index within its parent element's Child +// slice. +func (c *Comment) setIndex(index int) { + c.index = index +} + +// NewDirective creates an unparented XML directive token. +func NewDirective(data string) *Directive { + return newDirective(data, nil) +} + +// newDirective creates an XML directive and binds it to a parent element. If +// parent is nil, the Directive remains unbound. +func newDirective(data string, parent *Element) *Directive { + d := &Directive{ + Data: data, + parent: nil, + index: -1, + } + if parent != nil { + parent.addChild(d) + } + return d +} + +// CreateDirective creates an XML directive token with the specified 'data' +// value and adds it as the last child token of this element. +func (e *Element) CreateDirective(data string) *Directive { + return newDirective(data, e) +} + +// dup duplicates the directive. +func (d *Directive) dup(parent *Element) Token { + return &Directive{ + Data: d.Data, + parent: parent, + index: d.index, + } +} + +// Parent returns directive token's parent element, or nil if it has no +// parent. +func (d *Directive) Parent() *Element { + return d.parent +} + +// Index returns the index of this Directive token within its parent element's +// list of child tokens. If this Directive token has no parent, then the +// function returns -1. +func (d *Directive) Index() int { + return d.index +} + +// WriteTo serializes the XML directive to the writer. +func (d *Directive) WriteTo(w Writer, s *WriteSettings) { + w.WriteString("<!") + w.WriteString(d.Data) + w.WriteString(">") +} + +// setParent replaces the directive token's parent. +func (d *Directive) setParent(parent *Element) { + d.parent = parent +} + +// setIndex sets the Directive token's index within its parent element's Child +// slice. +func (d *Directive) setIndex(index int) { + d.index = index +} + +// NewProcInst creates an unparented XML processing instruction. +func NewProcInst(target, inst string) *ProcInst { + return newProcInst(target, inst, nil) +} + +// newProcInst creates an XML processing instruction and binds it to a parent +// element. If parent is nil, the ProcInst remains unbound. +func newProcInst(target, inst string, parent *Element) *ProcInst { + p := &ProcInst{ + Target: target, + Inst: inst, + parent: nil, + index: -1, + } + if parent != nil { + parent.addChild(p) + } + return p +} + +// CreateProcInst creates an XML processing instruction token with the +// specified 'target' and instruction 'inst'. It is then added as the last +// child token of this element. +func (e *Element) CreateProcInst(target, inst string) *ProcInst { + return newProcInst(target, inst, e) +} + +// dup duplicates the procinst. +func (p *ProcInst) dup(parent *Element) Token { + return &ProcInst{ + Target: p.Target, + Inst: p.Inst, + parent: parent, + index: p.index, + } +} + +// Parent returns processing instruction token's parent element, or nil if it +// has no parent. +func (p *ProcInst) Parent() *Element { + return p.parent +} + +// Index returns the index of this ProcInst token within its parent element's +// list of child tokens. If this ProcInst token has no parent, then the +// function returns -1. +func (p *ProcInst) Index() int { + return p.index +} + +// WriteTo serializes the processing instruction to the writer. +func (p *ProcInst) WriteTo(w Writer, s *WriteSettings) { + w.WriteString("<?") + w.WriteString(p.Target) + if p.Inst != "" { + w.WriteByte(' ') + w.WriteString(p.Inst) + } + w.WriteString("?>") +} + +// setParent replaces the processing instruction token's parent. +func (p *ProcInst) setParent(parent *Element) { + p.parent = parent +} + +// setIndex sets the processing instruction token's index within its parent +// element's Child slice. +func (p *ProcInst) setIndex(index int) { + p.index = index +} diff --git a/vendor/github.com/beevik/etree/helpers.go b/vendor/github.com/beevik/etree/helpers.go new file mode 100644 index 0000000..ea789b6 --- /dev/null +++ b/vendor/github.com/beevik/etree/helpers.go @@ -0,0 +1,394 @@ +// Copyright 2015-2019 Brett Vickers. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package etree + +import ( + "io" + "strings" + "unicode/utf8" +) + +type stack[E any] struct { + data []E +} + +func (s *stack[E]) empty() bool { + return len(s.data) == 0 +} + +func (s *stack[E]) push(value E) { + s.data = append(s.data, value) +} + +func (s *stack[E]) pop() E { + value := s.data[len(s.data)-1] + var empty E + s.data[len(s.data)-1] = empty + s.data = s.data[:len(s.data)-1] + return value +} + +func (s *stack[E]) peek() E { + return s.data[len(s.data)-1] +} + +type queue[E any] struct { + data []E + head, tail int +} + +func (f *queue[E]) add(value E) { + if f.len()+1 >= len(f.data) { + f.grow() + } + f.data[f.tail] = value + if f.tail++; f.tail == len(f.data) { + f.tail = 0 + } +} + +func (f *queue[E]) remove() E { + value := f.data[f.head] + var empty E + f.data[f.head] = empty + if f.head++; f.head == len(f.data) { + f.head = 0 + } + return value +} + +func (f *queue[E]) len() int { + if f.tail >= f.head { + return f.tail - f.head + } + return len(f.data) - f.head + f.tail +} + +func (f *queue[E]) grow() { + c := len(f.data) * 2 + if c == 0 { + c = 4 + } + buf, count := make([]E, c), f.len() + if f.tail >= f.head { + copy(buf[:count], f.data[f.head:f.tail]) + } else { + hindex := len(f.data) - f.head + copy(buf[:hindex], f.data[f.head:]) + copy(buf[hindex:count], f.data[:f.tail]) + } + f.data, f.head, f.tail = buf, 0, count +} + +// xmlReader provides the interface by which an XML byte stream is +// processed and decoded. +type xmlReader interface { + Bytes() int64 + Read(p []byte) (n int, err error) +} + +// xmlSimpleReader implements a proxy reader that counts the number of +// bytes read from its encapsulated reader. +type xmlSimpleReader struct { + r io.Reader + bytes int64 +} + +func newXmlSimpleReader(r io.Reader) xmlReader { + return &xmlSimpleReader{r, 0} +} + +func (xr *xmlSimpleReader) Bytes() int64 { + return xr.bytes +} + +func (xr *xmlSimpleReader) Read(p []byte) (n int, err error) { + n, err = xr.r.Read(p) + xr.bytes += int64(n) + return n, err +} + +// xmlPeekReader implements a proxy reader that counts the number of +// bytes read from its encapsulated reader. It also allows the caller to +// "peek" at the previous portions of the buffer after they have been +// parsed. +type xmlPeekReader struct { + r io.Reader + bytes int64 // total bytes read by the Read function + buf []byte // internal read buffer + bufSize int // total bytes used in the read buffer + bufOffset int64 // total bytes read when buf was last filled + window []byte // current read buffer window + peekBuf []byte // buffer used to store data to be peeked at later + peekOffset int64 // total read offset of the start of the peek buffer +} + +func newXmlPeekReader(r io.Reader) *xmlPeekReader { + buf := make([]byte, 4096) + return &xmlPeekReader{ + r: r, + bytes: 0, + buf: buf, + bufSize: 0, + bufOffset: 0, + window: buf[0:0], + peekBuf: make([]byte, 0), + peekOffset: -1, + } +} + +func (xr *xmlPeekReader) Bytes() int64 { + return xr.bytes +} + +func (xr *xmlPeekReader) Read(p []byte) (n int, err error) { + if len(xr.window) == 0 { + err = xr.fill() + if err != nil { + return 0, err + } + if len(xr.window) == 0 { + return 0, nil + } + } + + if len(xr.window) < len(p) { + n = len(xr.window) + } else { + n = len(p) + } + + copy(p, xr.window) + xr.window = xr.window[n:] + xr.bytes += int64(n) + + return n, err +} + +func (xr *xmlPeekReader) PeekPrepare(offset int64, maxLen int) { + if maxLen > cap(xr.peekBuf) { + xr.peekBuf = make([]byte, 0, maxLen) + } + xr.peekBuf = xr.peekBuf[0:0] + xr.peekOffset = offset + xr.updatePeekBuf() +} + +func (xr *xmlPeekReader) PeekFinalize() []byte { + xr.updatePeekBuf() + return xr.peekBuf +} + +func (xr *xmlPeekReader) fill() error { + xr.bufOffset = xr.bytes + xr.bufSize = 0 + n, err := xr.r.Read(xr.buf) + if err != nil { + xr.window, xr.bufSize = xr.buf[0:0], 0 + return err + } + xr.window, xr.bufSize = xr.buf[:n], n + xr.updatePeekBuf() + return nil +} + +func (xr *xmlPeekReader) updatePeekBuf() { + peekRemain := cap(xr.peekBuf) - len(xr.peekBuf) + if xr.peekOffset >= 0 && peekRemain > 0 { + rangeMin := xr.peekOffset + rangeMax := xr.peekOffset + int64(cap(xr.peekBuf)) + bufMin := xr.bufOffset + bufMax := xr.bufOffset + int64(xr.bufSize) + if rangeMin < bufMin { + rangeMin = bufMin + } + if rangeMax > bufMax { + rangeMax = bufMax + } + if rangeMax > rangeMin { + rangeMin -= xr.bufOffset + rangeMax -= xr.bufOffset + if int(rangeMax-rangeMin) > peekRemain { + rangeMax = rangeMin + int64(peekRemain) + } + xr.peekBuf = append(xr.peekBuf, xr.buf[rangeMin:rangeMax]...) + } + } +} + +// xmlWriter implements a proxy writer that counts the number of +// bytes written by its encapsulated writer. +type xmlWriter struct { + w io.Writer + bytes int64 +} + +func newXmlWriter(w io.Writer) *xmlWriter { + return &xmlWriter{w: w} +} + +func (xw *xmlWriter) Write(p []byte) (n int, err error) { + n, err = xw.w.Write(p) + xw.bytes += int64(n) + return n, err +} + +// isWhitespace returns true if the byte slice contains only +// whitespace characters. +func isWhitespace(s string) bool { + for i := 0; i < len(s); i++ { + if c := s[i]; c != ' ' && c != '\t' && c != '\n' && c != '\r' { + return false + } + } + return true +} + +// spaceMatch returns true if namespace a is the empty string +// or if namespace a equals namespace b. +func spaceMatch(a, b string) bool { + switch { + case a == "": + return true + default: + return a == b + } +} + +// spaceDecompose breaks a namespace:tag identifier at the ':' +// and returns the two parts. +func spaceDecompose(str string) (space, key string) { + colon := strings.IndexByte(str, ':') + if colon == -1 { + return "", str + } + return str[:colon], str[colon+1:] +} + +// Strings used by indentCRLF and indentLF +const ( + indentSpaces = "\r\n " + indentTabs = "\r\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t" +) + +// indentCRLF returns a CRLF newline followed by n copies of the first +// non-CRLF character in the source string. +func indentCRLF(n int, source string) string { + switch { + case n < 0: + return source[:2] + case n < len(source)-1: + return source[:n+2] + default: + return source + strings.Repeat(source[2:3], n-len(source)+2) + } +} + +// indentLF returns a LF newline followed by n copies of the first non-LF +// character in the source string. +func indentLF(n int, source string) string { + switch { + case n < 0: + return source[1:2] + case n < len(source)-1: + return source[1 : n+2] + default: + return source[1:] + strings.Repeat(source[2:3], n-len(source)+2) + } +} + +// nextIndex returns the index of the next occurrence of byte ch in s, +// starting from offset. It returns -1 if the byte is not found. +func nextIndex(s string, ch byte, offset int) int { + switch i := strings.IndexByte(s[offset:], ch); i { + case -1: + return -1 + default: + return offset + i + } +} + +// isInteger returns true if the string s contains an integer. +func isInteger(s string) bool { + for i := 0; i < len(s); i++ { + if (s[i] < '0' || s[i] > '9') && !(i == 0 && s[i] == '-') { + return false + } + } + return true +} + +type escapeMode byte + +const ( + escapeNormal escapeMode = iota + escapeCanonicalText + escapeCanonicalAttr +) + +// escapeString writes an escaped version of a string to the writer. +func escapeString(w Writer, s string, m escapeMode) { + var esc []byte + last := 0 + for i := 0; i < len(s); { + r, width := utf8.DecodeRuneInString(s[i:]) + i += width + switch r { + case '&': + esc = []byte("&") + case '<': + esc = []byte("<") + case '>': + if m == escapeCanonicalAttr { + continue + } + esc = []byte(">") + case '\'': + if m != escapeNormal { + continue + } + esc = []byte("'") + case '"': + if m == escapeCanonicalText { + continue + } + esc = []byte(""") + case '\t': + if m != escapeCanonicalAttr { + continue + } + esc = []byte("	") + case '\n': + if m != escapeCanonicalAttr { + continue + } + esc = []byte("
") + case '\r': + if m == escapeNormal { + continue + } + esc = []byte("
") + default: + if !isInCharacterRange(r) || (r == 0xFFFD && width == 1) { + esc = []byte("\uFFFD") + break + } + continue + } + w.WriteString(s[last : i-width]) + w.Write(esc) + last = i + } + w.WriteString(s[last:]) +} + +func isInCharacterRange(r rune) bool { + return r == 0x09 || + r == 0x0A || + r == 0x0D || + r >= 0x20 && r <= 0xD7FF || + r >= 0xE000 && r <= 0xFFFD || + r >= 0x10000 && r <= 0x10FFFF +} diff --git a/vendor/github.com/beevik/etree/path.go b/vendor/github.com/beevik/etree/path.go new file mode 100644 index 0000000..21760d3 --- /dev/null +++ b/vendor/github.com/beevik/etree/path.go @@ -0,0 +1,605 @@ +// Copyright 2015-2019 Brett Vickers. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package etree + +import ( + "iter" + "strconv" + "strings" +) + +/* +A Path is a string that represents a search path through an etree starting +from the document root or an arbitrary element. Paths are used with the +Element object's Find* methods to locate and return desired elements. + +A Path consists of a series of slash-separated "selectors", each of which may +be modified by one or more bracket-enclosed "filters". Selectors are used to +traverse the etree from element to element, while filters are used to narrow +the list of candidate elements at each node. + +Although etree Path strings are structurally and behaviorally similar to XPath +strings (https://www.w3.org/TR/1999/REC-xpath-19991116/), they have a more +limited set of selectors and filtering options. + +The following selectors are supported by etree paths: + + . Select the current element. + .. Select the parent of the current element. + * Select all child elements of the current element. + / Select the root element when used at the start of a path. + // Select all descendants of the current element. + tag Select all child elements with a name matching the tag. + +The following basic filters are supported: + + [@attrib] Keep elements with an attribute named attrib. + [@attrib='val'] Keep elements with an attribute named attrib and value matching val. + [tag] Keep elements with a child element named tag. + [tag='val'] Keep elements with a child element named tag and text matching val. + [n] Keep the n-th element, where n is a numeric index starting from 1. + +The following function-based filters are supported: + + [text()] Keep elements with non-empty text. + [text()='val'] Keep elements whose text matches val. + [local-name()='val'] Keep elements whose un-prefixed tag matches val. + [name()='val'] Keep elements whose full tag exactly matches val. + [namespace-prefix()] Keep elements with non-empty namespace prefixes. + [namespace-prefix()='val'] Keep elements whose namespace prefix matches val. + [namespace-uri()] Keep elements with non-empty namespace URIs. + [namespace-uri()='val'] Keep elements whose namespace URI matches val. + +Below are some examples of etree path strings. + +Select the bookstore child element of the root element: + + /bookstore + +Beginning from the root element, select the title elements of all descendant +book elements having a 'category' attribute of 'WEB': + + //book[@category='WEB']/title + +Beginning from the current element, select the first descendant book element +with a title child element containing the text 'Great Expectations': + + .//book[title='Great Expectations'][1] + +Beginning from the current element, select all child elements of book elements +with an attribute 'language' set to 'english': + + ./book/*[@language='english'] + +Beginning from the current element, select all child elements of book elements +containing the text 'special': + + ./book/*[text()='special'] + +Beginning from the current element, select all descendant book elements whose +title child element has a 'language' attribute of 'french': + + .//book/title[@language='french']/.. + +Beginning from the current element, select all descendant book elements +belonging to the http://www.w3.org/TR/html4/ namespace: + + .//book[namespace-uri()='http://www.w3.org/TR/html4/'] +*/ +type Path struct { + segments []segment +} + +// ErrPath is returned by path functions when an invalid etree path is provided. +type ErrPath string + +// Error returns the string describing a path error. +func (err ErrPath) Error() string { + return "etree: " + string(err) +} + +// CompilePath creates an optimized version of an XPath-like string that +// can be used to query elements in an element tree. +func CompilePath(path string) (Path, error) { + var comp compiler + segments := comp.parsePath(path) + if comp.err != ErrPath("") { + return Path{nil}, comp.err + } + return Path{segments}, nil +} + +// MustCompilePath creates an optimized version of an XPath-like string that +// can be used to query elements in an element tree. Panics if an error +// occurs. Use this function to create Paths when you know the path is +// valid (i.e., if it's hard-coded). +func MustCompilePath(path string) Path { + p, err := CompilePath(path) + if err != nil { + panic(err) + } + return p +} + +// traverse follows the path from the element e, yielding elements that match +// the path's selectors and filters using iterators. +func (p Path) traverse(e *Element) iter.Seq[*Element] { + pather := newPather() + return func(yield func(*Element) bool) { + pather.queue.add(node{e, p.segments}) + for pather.queue.len() > 0 { + if cont := pather.eval(pather.queue.remove(), yield); !cont { + return + } + } + } +} + +// A segment is a portion of a path between "/" characters. +// It contains one selector and zero or more [filters]. +type segment struct { + sel selector + filters []filter +} + +func (seg *segment) apply(e *Element, p *pather) { + seg.sel.apply(e, p) + for _, f := range seg.filters { + f.apply(p) + } +} + +// A selector selects XML elements for consideration by the +// path traversal. +type selector interface { + apply(e *Element, p *pather) +} + +// A filter pares down a list of candidate XML elements based +// on a path filter in [brackets]. +type filter interface { + apply(p *pather) +} + +// A node represents an element and the remaining path segments that +// should be applied against it by the pather. +type node struct { + e *Element + segments []segment +} + +// A pather is helper object that traverses an element tree using +// a Path object. It collects and deduplicates all elements matching +// the path query. +type pather struct { + queue queue[node] + results []*Element + inResults map[*Element]bool + candidates []*Element + scratch []*Element // used by filters +} + +// newPather creates a new pather instance. +func newPather() *pather { + return &pather{ + results: make([]*Element, 0), + inResults: make(map[*Element]bool), + candidates: make([]*Element, 0), + scratch: make([]*Element, 0), + } +} + +// eval evaluates the current path node by applying the remaining path's +// selector rules against the node's element, yielding results via iterator. +// Returns false if early termination is requested. +func (p *pather) eval(n node, yield func(*Element) bool) bool { + p.candidates = p.candidates[:0] + seg, remain := n.segments[0], n.segments[1:] + seg.apply(n.e, p) + + if len(remain) == 0 { + for _, c := range p.candidates { + if in := p.inResults[c]; !in { + p.inResults[c] = true + if !yield(c) { + return false + } + } + } + } else { + for _, c := range p.candidates { + p.queue.add(node{c, remain}) + } + } + return true +} + +// A compiler generates a compiled path from a path string. +type compiler struct { + err ErrPath +} + +// parsePath parses an XPath-like string describing a path +// through an element tree and returns a slice of segment +// descriptors. +func (c *compiler) parsePath(path string) []segment { + // If path ends with //, fix it + if strings.HasSuffix(path, "//") { + path += "*" + } + + var segments []segment + + // Check for an absolute path + if strings.HasPrefix(path, "/") { + segments = append(segments, segment{new(selectRoot), []filter{}}) + path = path[1:] + } + + // Split path into segments + for _, s := range splitPath(path) { + segments = append(segments, c.parseSegment(s)) + if c.err != ErrPath("") { + break + } + } + return segments +} + +func splitPath(path string) []string { + var pieces []string + start := 0 + inquote := false + var quote byte + for i := 0; i+1 <= len(path); i++ { + if !inquote { + if path[i] == '\'' || path[i] == '"' { + inquote, quote = true, path[i] + } else if path[i] == '/' { + pieces = append(pieces, path[start:i]) + start = i + 1 + } + } else if path[i] == quote { + inquote = false + } + } + return append(pieces, path[start:]) +} + +// parseSegment parses a path segment between / characters. +func (c *compiler) parseSegment(path string) segment { + pieces := strings.Split(path, "[") + seg := segment{ + sel: c.parseSelector(pieces[0]), + filters: []filter{}, + } + for i := 1; i < len(pieces); i++ { + fpath := pieces[i] + if len(fpath) == 0 || fpath[len(fpath)-1] != ']' { + c.err = ErrPath("path has invalid filter [brackets].") + break + } + seg.filters = append(seg.filters, c.parseFilter(fpath[:len(fpath)-1])) + } + return seg +} + +// parseSelector parses a selector at the start of a path segment. +func (c *compiler) parseSelector(path string) selector { + switch path { + case ".": + return new(selectSelf) + case "..": + return new(selectParent) + case "*": + return new(selectChildren) + case "": + return new(selectDescendants) + default: + return newSelectChildrenByTag(path) + } +} + +var fnTable = map[string]func(e *Element) string{ + "local-name": (*Element).name, + "name": (*Element).FullTag, + "namespace-prefix": (*Element).namespacePrefix, + "namespace-uri": (*Element).NamespaceURI, + "text": (*Element).Text, +} + +// parseFilter parses a path filter contained within [brackets]. +func (c *compiler) parseFilter(path string) filter { + if len(path) == 0 { + c.err = ErrPath("path contains an empty filter expression.") + return nil + } + + // Filter contains [@attr='val'], [@attr="val"], [fn()='val'], + // [fn()="val"], [tag='val'] or [tag="val"]? + eqindex := strings.IndexByte(path, '=') + if eqindex >= 0 && eqindex+1 < len(path) { + quote := path[eqindex+1] + if quote == '\'' || quote == '"' { + rindex := nextIndex(path, quote, eqindex+2) + if rindex != len(path)-1 { + c.err = ErrPath("path has mismatched filter quotes.") + return nil + } + + key := path[:eqindex] + value := path[eqindex+2 : rindex] + + switch { + case key[0] == '@': + return newFilterAttrVal(key[1:], value) + case strings.HasSuffix(key, "()"): + name := key[:len(key)-2] + if fn, ok := fnTable[name]; ok { + return newFilterFuncVal(fn, value) + } + c.err = ErrPath("path has unknown function " + name) + return nil + default: + return newFilterChildText(key, value) + } + } + } + + // Filter contains [@attr], [N], [tag] or [fn()] + switch { + case path[0] == '@': + return newFilterAttr(path[1:]) + case strings.HasSuffix(path, "()"): + name := path[:len(path)-2] + if fn, ok := fnTable[name]; ok { + return newFilterFunc(fn) + } + c.err = ErrPath("path has unknown function " + name) + return nil + case isInteger(path): + pos, _ := strconv.Atoi(path) + switch { + case pos > 0: + return newFilterPos(pos - 1) + default: + return newFilterPos(pos) + } + default: + return newFilterChild(path) + } +} + +// selectSelf selects the current element into the candidate list. +type selectSelf struct{} + +func (s *selectSelf) apply(e *Element, p *pather) { + p.candidates = append(p.candidates, e) +} + +// selectRoot selects the element's root node. +type selectRoot struct{} + +func (s *selectRoot) apply(e *Element, p *pather) { + root := e + for root.parent != nil { + root = root.parent + } + p.candidates = append(p.candidates, root) +} + +// selectParent selects the element's parent into the candidate list. +type selectParent struct{} + +func (s *selectParent) apply(e *Element, p *pather) { + if e.parent != nil { + p.candidates = append(p.candidates, e.parent) + } +} + +// selectChildren selects the element's child elements into the +// candidate list. +type selectChildren struct{} + +func (s *selectChildren) apply(e *Element, p *pather) { + for _, c := range e.Child { + if c, ok := c.(*Element); ok { + p.candidates = append(p.candidates, c) + } + } +} + +// selectDescendants selects all descendant child elements +// of the element into the candidate list. +type selectDescendants struct{} + +func (s *selectDescendants) apply(e *Element, p *pather) { + var queue queue[*Element] + for queue.add(e); queue.len() > 0; { + e := queue.remove() + p.candidates = append(p.candidates, e) + for _, c := range e.Child { + if c, ok := c.(*Element); ok { + queue.add(c) + } + } + } +} + +// selectChildrenByTag selects into the candidate list all child +// elements of the element having the specified tag. +type selectChildrenByTag struct { + space, tag string +} + +func newSelectChildrenByTag(path string) *selectChildrenByTag { + s, l := spaceDecompose(path) + return &selectChildrenByTag{s, l} +} + +func (s *selectChildrenByTag) apply(e *Element, p *pather) { + for _, c := range e.Child { + if c, ok := c.(*Element); ok && spaceMatch(s.space, c.Space) && s.tag == c.Tag { + p.candidates = append(p.candidates, c) + } + } +} + +// filterPos filters the candidate list, keeping only the +// candidate at the specified index. +type filterPos struct { + index int +} + +func newFilterPos(pos int) *filterPos { + return &filterPos{pos} +} + +func (f *filterPos) apply(p *pather) { + if f.index >= 0 { + if f.index < len(p.candidates) { + p.scratch = append(p.scratch, p.candidates[f.index]) + } + } else { + if -f.index <= len(p.candidates) { + p.scratch = append(p.scratch, p.candidates[len(p.candidates)+f.index]) + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} + +// filterAttr filters the candidate list for elements having +// the specified attribute. +type filterAttr struct { + space, key string +} + +func newFilterAttr(str string) *filterAttr { + s, l := spaceDecompose(str) + return &filterAttr{s, l} +} + +func (f *filterAttr) apply(p *pather) { + for _, c := range p.candidates { + for _, a := range c.Attr { + if spaceMatch(f.space, a.Space) && f.key == a.Key { + p.scratch = append(p.scratch, c) + break + } + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} + +// filterAttrVal filters the candidate list for elements having +// the specified attribute with the specified value. +type filterAttrVal struct { + space, key, val string +} + +func newFilterAttrVal(str, value string) *filterAttrVal { + s, l := spaceDecompose(str) + return &filterAttrVal{s, l, value} +} + +func (f *filterAttrVal) apply(p *pather) { + for _, c := range p.candidates { + for _, a := range c.Attr { + if spaceMatch(f.space, a.Space) && f.key == a.Key && f.val == a.Value { + p.scratch = append(p.scratch, c) + break + } + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} + +// filterFunc filters the candidate list for elements satisfying a custom +// boolean function. +type filterFunc struct { + fn func(e *Element) string +} + +func newFilterFunc(fn func(e *Element) string) *filterFunc { + return &filterFunc{fn} +} + +func (f *filterFunc) apply(p *pather) { + for _, c := range p.candidates { + if f.fn(c) != "" { + p.scratch = append(p.scratch, c) + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} + +// filterFuncVal filters the candidate list for elements containing a value +// matching the result of a custom function. +type filterFuncVal struct { + fn func(e *Element) string + val string +} + +func newFilterFuncVal(fn func(e *Element) string, value string) *filterFuncVal { + return &filterFuncVal{fn, value} +} + +func (f *filterFuncVal) apply(p *pather) { + for _, c := range p.candidates { + if f.fn(c) == f.val { + p.scratch = append(p.scratch, c) + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} + +// filterChild filters the candidate list for elements having +// a child element with the specified tag. +type filterChild struct { + space, tag string +} + +func newFilterChild(str string) *filterChild { + s, l := spaceDecompose(str) + return &filterChild{s, l} +} + +func (f *filterChild) apply(p *pather) { + for _, c := range p.candidates { + for _, cc := range c.Child { + if cc, ok := cc.(*Element); ok && + spaceMatch(f.space, cc.Space) && + f.tag == cc.Tag { + p.scratch = append(p.scratch, c) + } + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} + +// filterChildText filters the candidate list for elements having +// a child element with the specified tag and text. +type filterChildText struct { + space, tag, text string +} + +func newFilterChildText(str, text string) *filterChildText { + s, l := spaceDecompose(str) + return &filterChildText{s, l, text} +} + +func (f *filterChildText) apply(p *pather) { + for _, c := range p.candidates { + for _, cc := range c.Child { + if cc, ok := cc.(*Element); ok && + spaceMatch(f.space, cc.Space) && + f.tag == cc.Tag && + f.text == cc.Text() { + p.scratch = append(p.scratch, c) + } + } + } + p.candidates, p.scratch = p.scratch, p.candidates[0:0] +} |