Personal attacks again, hideheader. I told you once to stop it. If you can't provide a technical argument (percentage of those who use those rare keymaps) then just don't say anything.
You want numbers? Nearly 100% of Chrome users are using the "default" keymap with its Linux-like assignments, approximately 0% are using the "sublime" keymap. "vim" and "emacs" are two of the three keymaps provided by Firefox. If you wanna say that they're rarely used, then let's see your evidence.
You've yet to proffer evidence to support any of your opinions.
The primary goal of the numbers topic was to make the discussion less off-topic. You've already confirmed my opinion:
Nearly 100% of Chrome users are using the "default" keymap
Further discussion is moot, so I think it's not necessary, at least here and now.
No, not when it's just as easy to do it right.
We'll never be able to do it completely right, we can only intercept all known browser hotkeys (which is on my to do list anyway), but we won't be able to intercept the find toolbar when it's invoked from a menu. BTW, hardcoding a bunch of platform-specific hotkeys doesn't seem like doing it 'right', but unfortunately there's no other way. Also there could be a better solution for the advanced users like using another editor which already handles all these issues, so maybe we won't have to reinvent the wheel.
The primary goal of the numbers topic was to make the discussion less off-topic.
The primary goal of that tired dodge was to deflect attention from your own entirely unsupported opinion with a peremptory demand for ridiculously detailed support of my response. The secondary goal was to divert attention from a number of problems that you'd rather bury than fix.
You only need to understand how the keymaps work in CodeMirror to do it
Do you need me to remind you why we must have our own low-level keyboard handlers for the browser hotkeys also used for search/gotoline inside the code boxes?
I've fixed 1 and 2. I want to get this release out the door shortly, so I'm going to leave the keymap stuff alone, but if you want to PR it up, I'll take a look after the release.
So you won't be fixing the screwy Find behavior, either?
I'll fix it once I encounter it, certainly. The thing is, it was you who added the keymap option whereas I don't have any issues its default behavior added by yours truly...
And now we have the truth. The "I'm 40 years old and have been with a woman" bit never did ring true, kid.
I'm successfully learning from kids to be young, lol, and all my friends whom I respect are cheerful and have young souls. Would suffocate probably if I were as sour as you.
why we must have our own low-level keyboard handlers
Because CodeMirror doesn't handle anything outside of its boxes. So we have to handle all the different built-in search/navigation hotkeys used in browsers on different platforms but only when those are defined in currently selected CodeMirror keymap, a dozen, probably, or more. Anyway the number doesn't matter, the difficult part is avoiding hardcoded values as much as possible and having to read currently assigned CM hotkeys in their mnemonic form to decide if the pressed hotkey is one of them and reroute it to the corresponding CM command.
P.S. An alternative solution: maybe we can put everything (the header block with options and applies-to stuff) inside CodeMirror boxes as inner panels (AFAIK it's possible) maintaining current appearance more or less, remove gaps between the boxes and window bounds and use padding instead. Thus it might become impossible for a user to focus anything except CodeMirror, and we'll be able to use its keymap mechanism exclusively.
The thing is, it was you who added the keymap option...
What do keymaps have to do with it? "findNext" and "findPrev" should always begin from where the last search operation left off. They shouldn't skip occurrences and jump into a different section, they shouldn't repeat the same search from a different starting point, they shouldn't open the "Find" bar to begin a new search.
An alternative solution: maybe we can put everything (the header block with options and applies-to stuff) inside CodeMirror boxes as inner panels ... maintaining current appearance more or less, remove gaps between the boxes and window bounds and use padding instead.
Comments
Most people use Windows, but not everyone uses Windows.
You want numbers? Nearly 100% of Chrome users are using the "default" keymap with its Linux-like assignments, approximately 0% are using the "sublime" keymap. "vim" and "emacs" are two of the three keymaps provided by Firefox. If you wanna say that they're rarely used, then let's see your evidence.
The only one who's whining is you, almost constantly, although obviously you think otherwise, hideheader.
As for the default keymap (not sublime) being the only choice, I don't mind. The simpler the better.
And since you edited your post:
Apparently there's no need for my evidence anymore :-D
A diller, a dollar,
A second-rate scholar. No, not when it's just as easy to do it right.
The primary goal of the numbers topic was to make the discussion less off-topic. You've already confirmed my opinion:
Further discussion is moot, so I think it's not necessary, at least here and now.
We'll never be able to do it completely right, we can only intercept all known browser hotkeys (which is on my to do list anyway), but we won't be able to intercept the find toolbar when it's invoked from a menu. BTW, hardcoding a bunch of platform-specific hotkeys doesn't seem like doing it 'right', but unfortunately there's no other way. Also there could be a better solution for the advanced users like using another editor which already handles all these issues, so maybe we won't have to reinvent the wheel.
Being old is hard for you :-)
Do you need me to remind you why we must have our own low-level keyboard handlers for the browser hotkeys also used for search/gotoline inside the code boxes?
So you won't be fixing the screwy Find behavior, either? And now we have the truth. The "I'm 40 years old and have been with a woman" bit never did ring true, kid. Sure. Let's see if you can.
I'll fix it once I encounter it, certainly. The thing is, it was you who added the keymap option whereas I don't have any issues its default behavior added by yours truly...
I'm successfully learning from kids to be young, lol, and all my friends whom I respect are cheerful and have young souls. Would suffocate probably if I were as sour as you.
Because CodeMirror doesn't handle anything outside of its boxes. So we have to handle all the different built-in search/navigation hotkeys used in browsers on different platforms but only when those are defined in currently selected CodeMirror keymap, a dozen, probably, or more. Anyway the number doesn't matter, the difficult part is avoiding hardcoded values as much as possible and having to read currently assigned CM hotkeys in their mnemonic form to decide if the pressed hotkey is one of them and reroute it to the corresponding CM command.
P.S. An alternative solution: maybe we can put everything (the header block with options and applies-to stuff) inside CodeMirror boxes as inner panels (AFAIK it's possible) maintaining current appearance more or less, remove gaps between the boxes and window bounds and use padding instead. Thus it might become impossible for a user to focus anything except CodeMirror, and we'll be able to use its keymap mechanism exclusively.
Anyway I'm still inclined to remove the keymap option. But hopefully you're right and it's easy to fix. I'll look into it someday.