Can handle NetBeans all textmate features?

[Chris2011]

It came into my mind, when I wanted to add vue into NetBeans completely over textmate for syntax highlighting and LSP for vue features, etc. The problem here is, it seems that NetBeans don't handle includes or sub languages from textmate.

Vue can handle html (default), pug, haml, ts, js, css, scss, less, stylus, postcss. And if some pattern is matching for example lang="pug" there is this block in the textmate file:

"patterns": [
  {
      "include": "#tag-stuff"
  },
  {
      "contentName": "text.pug",
      "begin": "(>)",
      "beginCaptures": {
          "1": {
              "name": "punctuation.definition.tag.end.html"
          }
      },
      "end": "(?=</template>)",
      "patterns": [
          {
              "include": "text.pug"
          }
      ]
  }
]

and when I understand it correctly, it searches for a registered pug textmate module, but how is this done in NetBeans? Is this already implemented and if yes, how this is working? Maybe @jlahoda or @matthiasblaesing can help here. Thx.

[matthiasblaesing]

My understanding of this:

netbeans/ide/textmate.lexer/src/org/netbeans/modules/textmate/lexer/TextmateTokenId.java

Lines 119 to 154 in 70f6941

	public static void refreshGrammars() {
	Map<String, FileObject> scope2File = new HashMap<>();
	Map<String, Collection<String>> scope2Injections = new HashMap<>();
	Map<String, String> mimeType2Scope = new HashMap<>();
	FileObject editors = FileUtil.getSystemConfigRoot().getFileObject("Editors");
	if (editors != null) {
	Enumeration<? extends FileObject> en = editors.getChildren(true);
	while (en.hasMoreElements()) {
	FileObject candidate = en.nextElement();
	Object attr = candidate.getAttribute(GRAMMAR_MARK);
	if (attr instanceof String) {
	String scope = (String) attr;
	scope2File.put(scope, candidate);
	attr = candidate.getAttribute(INJECTION_MARK);
	if (attr instanceof String) {
	for (String s : ((String)attr).split(",")) {
	scope2Injections.computeIfAbsent(s, str -> new ArrayList<>()).add(scope);
	}
	} else {
	mimeType2Scope.put(FileUtil.getRelativePath(editors, candidate.getParent()), scope);
	}
	}
	}
	}
	synchronized (LanguageHierarchyImpl.class) {
	if (!Objects.equals(LanguageHierarchyImpl.scope2File, scope2File) ||
	!Objects.equals(LanguageHierarchyImpl.scope2Injections, scope2Injections) ||
	!Objects.equals(LanguageHierarchyImpl.mimeType2Scope, mimeType2Scope)) {
	LanguageHierarchyImpl.scope2File = scope2File;
	LanguageHierarchyImpl.scope2Injections = scope2Injections;
	LanguageHierarchyImpl.mimeType2Scope = mimeType2Scope;
	MimeDataProviderImpl.updateAllMimeTypes();
	}
	}
	}

runs through all "config" file/objects under the Editors branch in the system file system. It will pick up all files, that have an textmate-grammar or inject-to attribute. For your usecase you are looking for the former. For C this looks like this:

In the same file you find:

netbeans/ide/textmate.lexer/src/org/netbeans/modules/textmate/lexer/TextmateTokenId.java

Lines 161 to 183 in 70f6941

	IRegistryOptions opts = new IRegistryOptions() {
	@Override
	public String getFilePath(String scopeName) {
	synchronized (LanguageHierarchyImpl.class) {
	FileObject file = scope2File.get(scopeName);
	return file != null ? file.getNameExt() : null;
	}
	}
	@Override
	public InputStream getInputStream(String scopeName) throws IOException {
	synchronized (LanguageHierarchyImpl.class) {
	FileObject file = scope2File.get(scopeName);
	return file != null ? file.getInputStream(): null;
	}
	}
	@Override
	public Collection<String> getInjections(String scopeName) {
	synchronized (LanguageHierarchyImpl.class) {
	return scope2Injections.get(scopeName);
	}
	}
	};
	this.grammar = new Registry(opts).loadGrammar(scope);

This is the code that makes the textmake grammar/lexer definitions available to NetBeans and as you can see it uses the scope2File map created by the code mentioned above to find the grammar files for the provided scope name.

The entry from the graphic above is created by this:

netbeans/cpplite/cpplite.editor/src/org/netbeans/modules/cpplite/editor/file/CDataObject.java

Line 117 in 70f6941

@GrammarRegistration(grammar="resources/c.tmLanguage.json", mimeType=MIMETypes.C)

The annotation processor in the textmate.lexer module compiles that into a layer.xml and that in turn results in the tree entry you see.

So register the grammar and you should be done.

[Chris2011]

Thanks for your info. So I need to register all supported embedded languages for the vue mimetype that I will create? So HTML, PUG, TS, SASS, LESS, etc.? They need to be registered before, right?

[matthiasblaesing]

Yes that would be my understanding.

[Chris2011]

Oki, will check this. Thx.