Skip to content

Commit

Permalink
Parser/scanner for Doctype
Browse files Browse the repository at this point in the history 
Fixes part of #221
Doctype parsing setup. Internal DTD not complete, it instead stores as string

Signed-off-by: Nikolas Komonen <nikolaskomonen@gmail.com>
  • Loading branch information
@NikolasKomonen
NikolasKomonen committed Nov 15, 2018
1 parent ae60f52 commit d550ef1
Show file tree
Hide file tree
Showing 11 changed files with 455 additions and 71 deletions.
80 changes: 77 additions & 3 deletions org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/DocumentType.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,17 +18,42 @@
*/
public class DocumentType extends Node implements org.w3c.dom.DocumentType {

public enum DocumentTypeKind{
PUBLIC,
SYSTEM
}

/** Document type name. */
String name;

private DocumentTypeKind kind;
private String publicId;
private String systemId;
private String internalDTD; //TODO: THIS IS TEMPORARY. Implement actual parsing.

private String content;
int startContent;
int endContent;



public DocumentType(int start, int end, XMLDocument ownerDocument) {
super(start, end, ownerDocument);
}

/**
* @return the internalDTD
*/
public String getInternalDTD() {
return internalDTD;
}

/**
* @param internalDTD the internalDTD to set
*/
public void setInternalDTD(String internalDTD) {
this.internalDTD = internalDTD;
}

public String getContent() {
if (content == null) {
content = getOwnerDocument().getText().substring(getStartContent(), getEndContent());
Expand All @@ -52,6 +77,24 @@ public String getName() {
return name;
}

public void setName(String name) {
this.name = name;
}

/**
* @return the DocumentTypeKind
*/
public DocumentTypeKind getKind() {
return kind;
}

/**
* @param kind the DocumentTypeKind to set
*/
public void setKind(DocumentTypeKind kind) {
this.kind = kind;
}

/*
* (non-Javadoc)
*
Expand Down Expand Up @@ -109,7 +152,14 @@ public NamedNodeMap getNotations() {
*/
@Override
public String getPublicId() {
throw new UnsupportedOperationException();
return this.publicId;
}

/**
* @param publicId the publicId to set
*/
public void setPublicId(String publicId) {
this.publicId = cleanURL(publicId);
}

/*
Expand All @@ -119,6 +169,30 @@ public String getPublicId() {
*/
@Override
public String getSystemId() {
throw new UnsupportedOperationException();
return this.systemId;
}

/**
* @param systemId the systemId to set
*/
public void setSystemId(String systemId) {
this.systemId = cleanURL(systemId);
}

private static String cleanURL(String url) {
if (url == null) {
return null;
}
if (url.isEmpty()) {
return url;
}
int start = url.charAt(0) == '\"' ? 1 : 0;
int end = url.charAt(url.length() - 1) == '\"' ? url.length() - 1 : url.length();
return url.substring(start, end);
}





}
4 changes: 2 additions & 2 deletions org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/Node.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -334,11 +334,11 @@ public List<Node> getChildren() {
}

/**
* Add node child
* Add node child and set child.parent to this.
*
* @param child the node child to add.
*/
public void addChild(Node child) {
public void addChildAndSetItsParent(Node child) {
child.parent = this;
if (children == null) {
children = new XMLNodeList<Node>();
Expand Down
54 changes: 42 additions & 12 deletions org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/XMLParser.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@

import org.eclipse.lsp4xml.commons.BadLocationException;
import org.eclipse.lsp4xml.commons.TextDocument;
import org.eclipse.lsp4xml.dom.DocumentType.DocumentTypeKind;
import org.eclipse.lsp4xml.dom.parser.Scanner;
import org.eclipse.lsp4xml.dom.parser.TokenType;
import org.eclipse.lsp4xml.dom.parser.XMLScanner;
Expand Down Expand Up @@ -59,7 +60,7 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso
case StartTagOpen: {
Element child = xmlDocument.createElement(scanner.getTokenOffset(), text.length());
child.startTagOpenOffset = scanner.getTokenOffset();
curr.addChild(child);
curr.addChildAndSetItsParent(child);
curr = child;
break;
}
Expand Down Expand Up @@ -116,7 +117,7 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso
Element element = xmlDocument.createElement(scanner.getTokenOffset() - 2, text.length());
element.endTagOpenOffset = endTagOpenOffset;
element.tag = closeTag;
current.addChild(element);
current.addChildAndSetItsParent(element);
curr = element;
}
break;
Expand Down Expand Up @@ -159,7 +160,7 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso

case CDATATagOpen: {
CDataSection cdataNode = xmlDocument.createCDataSection(scanner.getTokenOffset(), text.length());
curr.addChild(cdataNode);
curr.addChildAndSetItsParent(cdataNode);
curr = cdataNode;
break;
}
Expand All @@ -181,7 +182,7 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso
case StartPrologOrPI: {
ProcessingInstruction prologOrPINode = xmlDocument.createProcessingInstruction(scanner.getTokenOffset(),
text.length());
curr.addChild(prologOrPINode);
curr.addChildAndSetItsParent(prologOrPINode);
curr = prologOrPINode;
break;
}
Expand Down Expand Up @@ -217,7 +218,7 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso

case StartCommentTag: {
Comment comment = xmlDocument.createComment(scanner.getTokenOffset(), text.length());
curr.addChild(comment);
curr.addChildAndSetItsParent(comment);
curr = comment;
try {
int endLine = document.positionAt(lastClosed.end).getLine();
Expand All @@ -228,7 +229,6 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso
} catch (BadLocationException e) {
LOGGER.log(Level.SEVERE, "XMLParser StartCommentTag bad offset in document", e);
}
// }
break;
}

Expand All @@ -241,15 +241,45 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso

case StartDoctypeTag: {
DocumentType doctype = xmlDocument.createDocumentType(scanner.getTokenOffset(), text.length());
curr.addChild(doctype);
curr.addChildAndSetItsParent(doctype);
doctype.parent = curr;
curr = doctype;
break;
}

case Doctype: {
case DoctypeName: {
DocumentType doctype = (DocumentType) curr;
doctype.startContent = scanner.getTokenOffset();
doctype.endContent = scanner.getTokenEnd();
doctype.setName(scanner.getTokenText());
break;
}

case DocTypeKindPUBLIC: {
DocumentType doctype = (DocumentType) curr;
doctype.setKind(DocumentTypeKind.PUBLIC);
break;
}

case DocTypeKindSYSTEM: {
DocumentType doctype = (DocumentType) curr;
doctype.setKind(DocumentTypeKind.SYSTEM);
break;
}

case DoctypePublicId: {
DocumentType doctype = (DocumentType) curr;
doctype.setPublicId(scanner.getTokenText());
break;
}

case DoctypeSystemId: {
DocumentType doctype = (DocumentType) curr;
doctype.setSystemId(scanner.getTokenText());
break;
}

case InternalDTDContent: {
DocumentType doctype = (DocumentType) curr;
doctype.setInternalDTD(scanner.getTokenText());
break;
}

Expand All @@ -271,14 +301,14 @@ public XMLDocument parse(TextDocument document, URIResolverExtensionManager reso
// FIXME: don't use getTokenText (substring) to know if the content is only
// spaces or line feed (scanner should know that).
String content = scanner.getTokenText();
if (content.trim().length() == 0) {
if (content.trim().length() == 0) { // if string is only whitespaces
break;
}
int start = scanner.getTokenOffset();
int end = scanner.getTokenEnd();
Text textNode = xmlDocument.createText(start, end);
textNode.closed = true;
curr.addChild(textNode);
curr.addChildAndSetItsParent(textNode);
break;
}

Expand Down
12 changes: 8 additions & 4 deletions org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/parser/Constants.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@

public class Constants {

public final static int _BNG = "!".codePointAt(0);
public final static int _EXL = "!".codePointAt(0);
public final static int _MIN = "-".codePointAt(0);
public final static int _LAN = "<".codePointAt(0);
public final static int _RAN = ">".codePointAt(0);
Expand Down Expand Up @@ -49,10 +49,14 @@ public class Constants {

public static final Pattern ATTRIBUTE_VALUE_REGEX = Pattern.compile("^[^\\s\"'`=<>\\/]+");

public static final Pattern PROLOG_NAME_OPTIONS = Pattern.compile("^(xml|xml-stylesheet)$");
public static final Pattern URL_VALUE_REGEX = Pattern.compile("^\"[^<>\"]*\"");

public static final Pattern PROLOG_NAME_OPTIONS = Pattern.compile("^(xml|xml-stylesheet)");

public static final Pattern DOCTYPE_KIND_OPTIONS = Pattern.compile("^(PUBLIC|SYSTEM)");

public static final Pattern PI_TAG_NAME = Pattern.compile("^[a-zA-Z0-9]+");

// public static final Pattern DOCTYPE_NAME =
// Pattern.compile("^[_:\\w][_:\\w-.\\d]*");
public static final Pattern DOCTYPE_NAME =
Pattern.compile("^[_:\\w][_:\\w-.\\d]*");
}
5 changes: 5 additions & 0 deletions org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/parser/MultiLineStream.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -87,6 +87,11 @@ public int peekChar() {
return peekChar(0);
}

/**
* Peeks at next char at position + n. peekChar() == peekChar(0)
* @param n
* @return
*/
public int peekChar(int n) {
int pos = this.position + n;
if (pos >= len) {
Expand Down
2 changes: 1 addition & 1 deletion org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/parser/ScannerState.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,6 @@
public enum ScannerState {
WithinContent, AfterOpeningStartTag, AfterOpeningEndTag, WithinProlog, WithinDoctype, WithinTag, WithinEndTag,
WithinComment, AfterAttributeName, BeforeAttributeValue, WithinCDATA, AfterClosingCDATATag, StartCDATATag, AfterPrologOpen, PrologOrPI,
WithinPI
WithinPI, AfterDoctypeName, AfterDoctypePUBLIC, AfterDoctypeSYSTEM, AfterDoctypePublicId, AfterDoctypeSystemId, AfterInternalDTDStartBracket, WithinDTD, WithinInternalDTD

}
13 changes: 9 additions & 4 deletions org.eclipse.lsp4xml/src/main/java/org/eclipse/lsp4xml/dom/parser/TokenType.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -33,16 +33,21 @@ public enum TokenType {
AttributeName,
AttributeValue,
StartPrologOrPI,
StartDoctypeTag,
PrologName,
PIName,
PIContent,
PIEnd,
PrologEnd,
Doctype,
EndDoctypeTag,
Content,
Whitespace,
Unknown,
EOS
EOS,
StartDoctypeTag,
DoctypeName,
DocTypeKindPUBLIC,
DocTypeKindSYSTEM,
DoctypePublicId,
DoctypeSystemId,
InternalDTDStart,
EndDoctypeTag, InternalDTDEnd, InternalDTDContent
}
Loading

0 comments on commit d550ef1

Please sign in to comment.