Currently, by pressing ctrl+R one gets funciton lists only for the current file. 

i'd like to see RPM spec file sintax higthligted in the editor

Clicking a line number and dragging down over other lines should select all lines within that range including the final line the mouse is over (currently the last line is not included in selection, which differs from standard behavior in other apps).

As a longtime, die-hard Emacs user I found Sublime Text 2 is simply awesome; however the only thing that's keeping me from purchasing it is it's lack of a plain text mode because easily I spend half of my editing time on a tmuxed console.

Having Sublime Text 2 available for console -with the same functionality graphical version have- would be simple amazing.

Btw: an Emacs-like key-binding feature would be greatly welcome also!

Regexes cannot handle complex language constructs. Parser would be a much better match in many situations. 

When selecting a character the status bar says "1 characters selected". It would be nice if it also displayed the unicode value for the characters that were selected. I do a lot of localization work.

When you middle click (or Shift+click) an icon of a program that's already running in the taskbar in Windows 7/8 the standard behavior is to open a new instance of the program. Sublime doesn't seem to comply with that and instead focuses on the currently open window.

Since we can remove files/folders it would be extremely useful for us to then be able to order by extension types.

Currently when closing a cloned view of a file that is currently open and has changes, you will get a warning box. Since the file is still open, this is unnecessary.

The ability of specifying word separators is great.
However there are many cases where their treatment is a bit counter-intuitive (see below).

Proposal in brief: three separate settings (word breaks indicated by | )

1. special_chars: they get treated as the current word_separators. i.e. word boundaries precede and follow them as in "bla |bla|:| bla"
   
2. pre_word_separators: they introduce a word boundary only before themselves. e.g.: "bla|\bla| bla" but not "bla|\|bla| bla"
3. post_word_separators: they introduce a word boundary only after themselves. e.g.: "bla |bla'|bla" (think of prime in math mode, not string quotes.)
   

Example: backslash in LaTeX.
If you consider backslash as a word separator in LaTeX, you end up not being able to select a command as a single word, by double clicking for example. But if you exclude it from the word separators two commands in a row would be considered as a single word e.g.: "\cmd\cmd".
My solution would make this work as expected and is general and clean enough to find applications in other use cases.