Josh Carpenter [:jcarpenter] Reporter
	Description • 11 years ago

When holding down the Delete key, we need to increase speed at which characters are deleted. The variable controlling this seems to live here on line 169 of keyboard.js (const REPEAT_RATE = 100;) Let's increase the speed by 50% and see how it feels. Right now we're too slow and it makes text deletion feel unnecessarily sluggish.

LOE and risk seem extremely low here.

	Mike Lee [:mlee]
	Comment 1 • 11 years ago

Jed, please take a look at this after bug 827811.

	Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧ Assignee
	Comment 2 • 11 years ago

It looks like we're currently spending nearly 50ms per deletion computing things (on unagi), so hopefully the repeat rate doesn't need to be less than that.

	Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧ Assignee
	Comment 3 • 11 years ago

Do we want to leave the delay before starting to autorepeat at 700ms, or should/can that be smaller?

	Josh Carpenter [:jcarpenter] Reporter
	Comment 4 • 11 years ago

Let me check that today on most recent builds and follow up with you. My phone is dead after the flight to London and I want to refresh my memory before committing one way or the other.

	Josh Carpenter [:jcarpenter] Reporter
	Comment 5 • 11 years ago

Hi Jed, some delay makes sense, but let's reduce it to 400ms.

	Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)
	Comment 7 • 11 years ago

Comment on attachment 739490 [details]
Gaia PR

Small riskless fix.

	Mike Lee [:mlee]
	Comment 8 • 11 years ago

Alex,
We have r+ and Andreas' approval to put this in 1.1 so can we get this uplifted? I'm setting tracking+ but don't see an option for requesting gaia approval.

	Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧ Assignee
	Comment 9 • 11 years ago

It's not committed to master yet — and it shouldn't be, in its current state; see https://github.com/samjoch/gaia/commit/d899b12e8ca25713db0a86b81b42fc9519d4c614#apps-keyboard-style-keyboard-css-P4

	David Flanagan [:djf]
	Comment 10 • 11 years ago

The word suggestion code has a special case for auto-repeating backspace. If we're autorepeating it does not try to compute word suggestions until the auto repeat stops. I think there is a time constant in keyboard/js/imes/latin/latin.js related to how long to wait after a key repeat backspace before computing suggestions.

So if you change the auto-repeat interval, you might also want to change that other constant too.

One issue I've seen is that auto-repeat backspace through the beginning of a sentence is funky because the latin IM is still doing automatic capitalization computations, so I've seen a pause in the backspace speed when you delete the first character of a sentence while they keyboard shifts to uppercase.  This probably doesn't affect master anymore, however.  But something to look out for.  If its an issue, we can fix it

	Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧ Assignee
	Comment 11 • 11 years ago

>  // If we get an autorepeating key is sent to us, don't offer suggestions
>  // for this long, until we're pretty certain that the autorepeat
>  // has stopped.
>  const autorepeatDelay = 250;

I'm not sure if that needs to change, given that we're just lowering the autorepeat interval from 100 to 75?

	David Flanagan [:djf]
	Comment 13 • 11 years ago

Comment on attachment 755572 [details]
Link to GitHub pull request

One line change: easy r+

	John Ford [:jhford] CET/CEST Berlin Time
	Comment 14 • 11 years ago

master: 0c0f4c24c3814ed9f6d3f9b189ea20bbefb4ddc3

	Mike Lee [:mlee]
	Comment 15 • 11 years ago

Was this uplifted for 1.1 per comment #8?