Community
Participate
Working Groups
After activating the key assist popup (Shift+?, or Ctrl+Shift+L if you're in the editor), it can only be dismissed by pressing ESC (sometimes you have to give it focus first). This surprises me. On other websites that have key assist, you can press the hotkey again to toggle it off.
By comparison, in Gmail you can hide key assist in any of these ways: 1. Press the hotkey again 2. Click "Close" on the key assist panel 3. Click anywhere outside the panel 4. Press ESC (it just works, regardless of focus) We should support all of these. Right now we only have a partial implementation of #4.
I'll take this for now (since I implemented it initially)...but Mark if you are actively working in the editor and want to fix this, feel free...
Fixed (mostly) in a22312944fc6729755d2b78a06bbe9b37655c9be (In reply to comment #1) > By comparison, in Gmail you can hide key assist in any of these ways: > 1. Press the hotkey again > 2. Click "Close" on the key assist panel > 3. Click anywhere outside the panel > 4. Press ESC (it just works, regardless of focus) > > We should support all of these. > Right now we only have a partial implementation of #4. 1. fixed. 2. we don't have a close affordance but I added this to a general "beef up key assist" bug in bug 364995 3. fixed. 4. fixed. This used to work (bug 348792). I think it got broken by the content assist mode (even the order in which the modes were added could affect this). Since the content assist mode was always claiming that it was active, it swallowed esc before the key assist mode ever got a chance. I had to change the editor code to offer all modes the esc key, even if one mode claims to have processed it. This is similar to what the "escHandler" in setup.js does. The only remaining oddity in my mind is that you can still be in two modes at once in the editor. For example, press Ctrl+Shift+L to get the key panel. Press Ctrl+Space to activate content assist. now you have both modes up, and if you press cancel they both disappear. If we really want to do better, we need to look at bug 348823