> > > Vitalije, are you thinking of a strict or lenient parser? > > EKR > I must say that I thought that the writing full lexer/parser for the jungle of current javascript dialects was out of question. It seemed to me way too much work for just a single language. That is why I did abandon this approach and proposed some kind of combination of human intelligence and dummy but fast computer text operations, so that user can very easily shape code the way he/she likes.
When rust language appeared on this forum, I've got interested in it especially when I learned that it is quite easy to make python extensions in rust. Rust also allows pattern matching which is wonderful feature for writing parsers. There is a rust library that parses modern javascript. When I saw all these, I had an impulse to try it for Leo import when I got some time to experiment with this. There is also python module ply <https://github.com/dabeaz/ply> for writing lexer/parser functions in pure python. AFAR this module was requiring latest Python3.6, but as I write this and check the previous link, it seems that they made ply compatible with both python2 and python3 versions (which is good news for Leo if this module is to be used). I can try to: - write lexer/parser using ply in pure python - write python extension for using rust library for creating lexers/parsers Any of this will build AST of imported script. With AST at ones disposal, one can easily check all function definitions, object definitions, property definitions for their start line and end line. In case they contain more lines than some user preference this peace of code can be extracted into separate node, inserting @others if it is not already inserted. Also types of AST nodes can give very good suggestion for what the headline might be. Javascript is important language nowadays and it might be worth building lexer/parser for it. But it is open-ended project. Javascript syntax is permanently changing and even native javascript tools have a hard time to keep up. On the other hand for the effective use of AST we still would need those text manipulating functions. It seems to me that these functions should be our first target, and then if they would not suffice, build lexer/parser that would generate AST which can be used to automatically split the source file into nodes. Importers are extremely important for using Leo like an ordinary text editor. Suppose some user wants to try Leo and issues command leo somesource.js Leo would respond by opening new empty outline and importing the given source file. If it takes too much time (more than 500ms) it would make very poor impression on the user. In case the user is forgiving and willing to wait that long, at least what he/she sees after waiting must look good enough. Otherwise, most of the users would regret for even trying Leo. Vitalije -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/leo-editor. For more options, visit https://groups.google.com/d/optout.
