Option for CODEX to ignore certain comments (or code)

[clotodex]

I took a stab at implementing this.

How it works:

• # codex-ignore will ignore this and the next line
• # codex-on/off will toggle and ignore everything inbetween
• if enabled, comments starting with an item from skip_list, currently todo and fixme, will be skipped

A few open questions:

• [ ] creation of vim commands (not my expertise here, don't know if optional arguments etc could be handled, would follow your lead)
• [ ] how to best make the skip_list adjustable, should it be a vim variable or part of the function, or do we need it at all?
• [ ] should #codex-ignore [n] take an optional n parameter to skip not just the next, but the next n lines
• [ ] how to handle whole e.g. todo comment blocks, i.e. when the comment spans multiple lines, currently I am just removing the one containing the todo
• [ ] multilingual support, currently I am just parsing based on #, is there a way to get comments language-independant or is there a smarter way to parse this

Feel free to check it out, add your comments and thoughts and then let's go from there.

[clotodex]

After a discussion with @tom-doerr I will take an opinionated approach to this implementation

• vim commands: whatever I can figure out, since they are on top and just wrap python functions this could be easily changed later
• skiplist - would rather make this project (.skipcodex) or file specific (# CODEX_SKIP_PREFIX todo fixme ...) => for now most likely just a on/off config with the default
• [n] - implementation later
• todo blocks => would shelve this for now and rather handle it with codex on and off (if it is a long todo explanation, adding this should not be a problem, lateron this behavior could be a flag)
• multilingual support => is kind of essential from the get go => will introduce an optional dependency to tree-sitter

[clotodex]

Thoughts to tree-sitter (might give performance hit - dont know)

def search_comments():
    # Get the buffer's content as a string
    buffer_content = vim.current.buffer[:]

    # Get the tree-sitter parser for the buffer's language
    language = vim.eval("tree_sitter_language#get_parser().language")
    parser = vim.eval("tree_sitter_language#get_parser().parser")

    # Parse the buffer's content
    tree = parser.parse(buffer_content)

    # Query the tree for comment nodes
    comments = tree.root.query("//comment")

    # Print the line number and text content of each comment node
    for comment in comments:
        start_byte = comment.start_byte
        end_byte = comment.end_byte
        comment_text = buffer_content[start_byte:end_byte]
        line_number = buffer_content.count('\n', 0, start_byte) + 1
        print(f"Line {line_number}: {comment_text}")

Checking if installed:

if vim.eval("exists('g:loaded_tree_sitter')") == '1':
    # Tree-sitter is installed
else:
    # Tree-sitter is not installed

[tom-doerr]

Can the existing code already be merged to the project? It's still displayed for me as a draft

[clotodex]

Can the existing code already be merged to the project? It's still displayed for me as a draft

In theory it is ready to merge, however I did not even test it yet - If you test it, feel free to merge. Sadly I don't have much time right now and only will get to it in the upcoming weeks.