thomas_linder_puls:
> But even in C++ and C# there is also a set of fixed/statuc keywords (dark-blue), like class, public, private, static, return, for, if, etc. Such lists will of course vary when languages develop over time, but a huge core of them typically continue to exist.
Scintilla lexers may cover more than one language with the cpp lexer handling C, C++, Java, C#, JavaScript, and others. There would have to be a tighter selection (possibly called something like language or lexer mode) for keywords.
Language communities often have different ideas of just what is a keyword with various literals and predefined identifiers treated as keywords.
> Unfortunately, I do not have such lists for the languages that Scintilla lexers supports. I doubt that you do either (did you have the lists for Visual Prolog, before I listed them above?).
Some projects use keyword lists from SciTE’s .properties files. No one has provided a SciTE .properties file for Visual Prolog.
> It is true that Scintilla itself will have an extra maintenaince overhead from this, but the alternative it that either all the direct users (e.g. the Notepad++ programmers) each have an equally large overhead, or even worse that each of the users of their programs (e.g. the Notepad++ users) all have such an overhead or is completely restrained from using a certain language (as it is the case for Notepad++ users).
Effort can be moved from other projects to Scintilla by expanding the scope of Scintilla. That isn’t going to work unless there is a corresponding increase in labour provided to Scintilla to handle, discuss, integrate, and test this additional scope.
Neil