! What precise workflow do I need for me to be viable to use this?
! From a user (novice) point of view, how does it look like?

- Buffer edit history separate from caret history (tagging history blocks with carets???)
- We need regex for: [FormatCode matching, IsCode matching,
- Macros
- Window position: vscode opens in fullscreen and then remembers what position it was close in (could be a config option)
- On Linux: Try to implement is_debugger_present()
- Maybe IPC for te.exe when it's already open and file arguments are passed it should perhaps open a buffer in current window??
- Add <<File>> <<ProjectDirectory>> template strings to Open (Then remove SEtWorkdirhere)
- :Set Filename to name current buffer ??? :O and others like that!!
- Make a fuzzy command !> grep and fuzzy over it??? (doesn't seem very useful for grep)
- Make the equivalent of SearchOpenBuffers but for cmds like !@git grep -n "@>"
- Add Bool variable
- Guide on the first page for new users with links to configs, tutorials

Debug session:
- Report errorf - use coroutine dialogs
    - Replace in render layer also
- BlockAllocator something is not working there which only showed after executing OpenCode on many files
- Some bad allocating happening in Clipboard for sure OR MAYBE BLOCK ALLOCATOR ACTION

New UI Session
- Uneditable buffers ?
- Maybe marked allocations??? So that we can associate allocations with a buffer or view and then dealloc all at the same time

- Open with seek string (open at pattern) filename:32 filename:/^Window$/
- Show what process/coroutines are running and allow to kill (active process buffer?)
- Database idea: use special buffers to store information
    - Editing the buffer doesn't seem to be the slow part rather, accessing the data and putting it into the buffer (potentially hitting many different memory locations) I have a crazy idea to use buffers in order to store the names in a serialized format
    - non editable buffers (raw ops ok, non-raw no op)
    - DBBuffer

- Try to add Tracking Allocator and rewrite the app, free all memory at the end of the app and check all is well

FEATURE Some decl/function indexing in fuzzy format
FEATURE dump text editor state to file, restore state
- word complete
- escapeing multiple cursor after ctrl + d should put the cursor where it was (probably will need to swap secondary and primary cursor for new cursor
- draw indentation levels like in sublime (those lines) - we render chars one by one so seems relatively easy to figure out if whitespace belongs to beginning of line (make sure to add max value like 40 because of big files)
- code sections, visual demarkation if beginning of line has a very specific text + goto next / goto prev section hotkey!
- combine glyph and selection rendering

- Test stdin writing code
- Implement shell interaction (the valid cmd lines should start with '>' or '$', user can add more lines like this to expand the command size maybe?, if we have a case where we have a line with '>' but the last line doesn't have (just a space) then it should execute?)
- drop text into window
- I think the way sublime text and we display line highlights is confusing with multiple cursors (line highlight can be confused with selection)
- make the editor replayable, store events and then replay, be careful about globals
- maybe open should return multiple options if there are many more? (like in sublime if many symbols you get a window and you choose and it automatically jumps you to the symbol in the background)
- I want a way to assign flags to buffers/views/windows from user perspective so that console window concept can be created from user space
- font cache and on demand unicode loads
- color parens, braces
- gap buffer
- optimize rendering - command buffer, and vertice buffer instead of vertice buffer with scissor










WORD COMPLETE
1. Tokenize the file (or maybe couple last viewed files)
2. Match the tokens (identifiers) with prefix string and produce a list
3. Remove duplicates
4. Sort by proximity to cursor

We save the iterator, it's enough for it to be global with unique for it arena
If prefix ask is the same - reuse the iterator go to next word, if there are none go to next buffer ...
    else - dump old, create new