I find myself constantly doing a "Find All" for a certain identifier, searching through the hits for the definition, opening the definition, and going back and closing the Find Results window. It would be super handy if this workflow were shortened to just a "Find Definition" command.

Granted this would need some language level integration in order to determine what hit was most likely the definition, but it could also be done via some popup list of the hits ordered by likelihood of being a definition which the user could click and it would pop open that find result without the intermediate step of having a Find Results window to deal with.

Furthermore, having a shortcut for this command such as command- or control-clicking (or double-clicking) a word would be super rad. I used a shortcut like that in Visual Studio for years and loved it.

Editing files to set preferences is nice and flexible, but it's a bit of a pain to have to select the default file first (to see what's available) and then select the user file (to make changes). That also subtly encourages editing that default file. Here's a simple solution:

- For each kind of preference in the Preferences menu, display a single item: the name of the preference.

- Selecting that item displays both the default file and the user file in a split view, e.g. defaults above, user below. The user can then read the desired info from the top pane and copy the appropriate lines into the other.

- The Preferences menu item could then be much simpler: a single entry for each preference would display the pair of files (as just described).

To reiterate, I see these advantages:

- Simpler Preferences menu: one item per kind of preference.

- It is easier to for the user to edit preferences: one selection does it all.

- There is less temptation to edit the default file.

Hitting F12 (Jump to Definition) should work if hovering over an import statement in a given language.  Language examples:

LESS:

@import "other_file.less";

PHP

include "other_file.php";

I understand this may be more practical for some languages than others, I certainly wouldn't expect it to work in all languages in all cases.

Simple operations on each line in big file (50k lines) are not very fast. And I see only one CPU core is used at this time. I think performance can be improved if you run those multi line operations in parallel.