Audit Ctrl/Accel for keyboard shortcuts

[ian-r-rose]

The default shortcuts schema file has about 10 shortcuts that use Ctrl rather than Accel. Some may be deliberate, but I think many are not. We should go through and determine which ones should be changed to Accel.

[ian-r-rose]

cc @NoahStapp @jennalandy

[saulshanabrook]

Notes from meeting:

Ctrl maps to command on a mac more naturally. So we should switch the shortcut so that it is more natural on a mac.

We don't have a way to specify platform specific keyboard shortcuts declaratively. Maybe it would make sense to add platform specific keyboard shortcuts, but this would not be 1.0.

[ian-r-rose]

Here are the shortcuts that use Ctrl, as well as a proposal for how to handle them:

Command	Old	New	Reason
application:activate-next-tab	Ctrl Shift ]	Unchanged	Avoid collision with browser shortcuts on Mac+Chrome (#1681)
application:activate-previous-tab	Ctrl Shift [	Unchanged	Avoid collision with browser shortcuts on Mac+Chrome (#1681)
editmenu:redo	Ctrl Shift Z	Accel Shift Z	Better for Mac users
editmenu:undo	Ctrl Z	Accel Z	Better for Mac users
filemenu:close-and-cleanup	Ctrl Shift Q	Accel Shift Q	Better for Mac users ?
helpmenu:toggle	Ctrl Shift H	Removed	Unused shortcut
console:linebreak	Ctrl Enter	Accel Enter	Better for mac users, notebook uses Ctrl Enter to do something different
notebook:enter-command-mode	Ctrl M (??)	Unchanged	Feature parity with classic
notebook:run-cell	Ctrl Enter	Unchanged	Feature parity with classic
notebook:split-cell-at-cursor	Ctrl Shift -	Unchanged	Feature parity with classic

If somebody with a Mac is able to test whether Cmd Shift [/] for activating next/previous tab conflicts with them (as suggested in #1681), I'd appreciate knowing about it. I'd also be interested in the behavior of Cmd Shift Q on Macs.

[jasongrout]

On my mac, Cmd Shift [ and Cmd Shift ] switch firefox tabs.

[ian-r-rose]

@jasongrout So that would mean that keeping our shortcut unchanged is probably a good idea. Do you know if those shortcuts are overridable?

[jasongrout]

filemenu:close-and-cleanup Ctrl Shift Q Accel Shift Q Better for Mac users ?

On a mac/firefox, Cmd W closes the tab, Shift Cmd W closes the window, and Cmd Q closes the application. Accel Shift Q seems open, though it probably is more in line with closing all of JupyterLab than closing a single tab. But I'm not sure there is a better choice.

By the way, filemenu:close-and-cleanup is disabled in the launcher tab, so if you're just used to using one shortcut, it messes you up. You can use Ctrl W (JLab close tab) to close the tab.

[jasongrout]

Do you know if those shortcuts are overridable?

I tried, and it doesn't seem I can override them in firefox.

[ian-r-rose]

On a mac/firefox, Cmd W closes the tab, Shift Cmd W closes the window, and Cmd Q closes the application. Accel Shift Q seems open, though it probably is more in line with closing all of JupyterLab than closing a single tab. But I'm not sure there is a better choice.

Sounds good. I'd prefer to keep the churn here on the minimal side, since keyboard shortcuts can be pretty ingrained.

By the way, filemenu:close-and-cleanup is disabled in the launcher tab, so if you're just used to using one shortcut, it messes you up. You can use Ctrl W (JLab close tab) to close the tab.

Interesting, I suppose we could have the close-and-cleanup delegator itself call close if there is nothing that has registered itself for close-and-cleanup.

Do you know if those shortcuts are overridable?

I tried, and it doesn't seem I can override them in firefox.

Okay, let's keep those as is, then. Thanks for checking.

[jasongrout]

+1 to everything else in your list.

[ian-r-rose]

@jasongrout is there anything you disagree with?

[jasongrout]

Nope, all looks good to me. I verified some of your statements about how things work in classic notebook on mac as well.

The only note I have at this point is about the filemenu:close-and-cleanup command not working for things like launchers.

[ian-r-rose]

I'll see if I can come up with something clever for that case.

[ian-r-rose]

@jasongrout #6447 is ready, with some cleanup of close-and-cleanup.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related discussion.