1497518 - Some keyboard shortcuts broken in GMail, GSuite, Remember The Milk, etc. if keyCode of keypress event is same as its charCode

Status: RESOLVED FIXED

(Web Compatibility :: Site Reports, defect, P1)

Description

[Andrew Overholt [:overholt]]

This just started with today's nightly AFAICT.

Comment 1

[Andrew Overholt [:overholt]]

Masayuki's been working on some related stuff here so perhaps he knows what's going on.

Comment 2

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Hmm, must be regression of bug 1479964. Is it fixed with disabling "dom.keyboardevent.keypress.set_keycode_and_charcode_to_same_value"?

Comment 3

[Andrew Overholt [:overholt]]

(In reply to Masayuki Nakano [:masayuki] (JST, +0900) from comment #2)
> Hmm, must be regression of bug 1479964. Is it fixed with disabling
> "dom.keyboardevent.keypress.set_keycode_and_charcode_to_same_value"?

Yes, thanks.

Comment 4

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Sigh, that means that the change breaks Closure...

Comment 6

[Marco Bonardo [:mak]]

This also causes Google Docs to misbehave, not just on keyboard shortcuts, I couldn't type anymore a lot of chars like R, T, down, left, backspace in a document, it made it pretty much unusable. Reverting the pref fixed it.

Comment 7

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Sigh. There are some UA name check around keyCode referrers in the source code of rememberthemilk <https://www.rememberthemilk.com/app/_/2256/compiled.2256.ja.js>. Anyway, we need to contact Closure team before making bug 1479964 ride the train.

Comment 8

[Steve Fink [:sfink] [:s:]]

Reminiscent of (but different than) bug 1474476.

Comment 9

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

I'll investigate what actually caused breaking the apps today.

Comment 11

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Oh, sorry. I should've posted here rather than bug 1479964.

As far as I've checked roughly, Google Closure needs to update by themselves around here (at least):
https://github.com/google/closure-library/blob/719529feafc9d60eecd2f620756d3e5cc84a9250/closure/goog/events/keyhandler.js#L400-L413
https://github.com/google/closure-library/blob/719529feafc9d60eecd2f620756d3e5cc84a9250/closure/goog/events/keyhandler.js#L439-L449
https://github.com/google/closure-library/blob/719529feafc9d60eecd2f620756d3e5cc84a9250/closure/goog/events/keycodes.js#L413-L415
because they check both UA string and keyCode. So, if UA is Gecko, they assumes that one of keyCode or charCode of keypress event is 0, but we need to change it for compatibility with the other browsers.

Can somebody contact the developers?

It might be better to release bug 968056.  If so, they don't need to check if keypress event is caused by non-printable key with checking whether keyCode value is non-zero. (Although if they need to support ESR60, they need more checks!!)

On the other hand, I'm not sure changing our behavior is safer than making broken web sites fix by themselves.

Comment 13

[Mike Taylor [:miketaylr]]

> Can somebody contact the developers?

I'll carve out time for this tomorrow.

Comment 14

[Mike Taylor [:miketaylr]]

I've filed https://github.com/google/closure-library/issues/932, Adam, can you contact Google on our ML to raise there as well? Thanks.

Comment 15

[Adam Stevenson [:adamopenweb]]

Contacted via partner mailing list.

Comment 17

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Well, looks like that Google Docs works fine even without the blacklist. Could you verify that more? Because I don't usually use Google Docs. If they have already fixed, I'll file a bug to remove Google Docs from the blacklist.

Comment 18

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

I emailed to rememberthemilk.com.

Comment 19

[Hsin-Yi Tsai (she/her) [:hsinyi]]

Hi Florin,
Could your team help to verify this issue that has been no longer valid on these websites?
- Google docs
- rememberthemilk.com
- medium.com

Thank you.

Comment 20

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

I got a reply from rememberthemilk.com. They fixed their bug on Beta channel today. I confirmed it. And they said that it'll be shipped this week in Stable channel. So, we can remove rememberthemilk.com from the blacklist this weekend.

On the other hand, could you somebody to check whether Google Docs has already fixed their bug with removing "docs.google.com" from "dom.keyboardevent.keypress.hack.use_legacy_keycode_and_charcode" on your Nightly build? After somebody verified that, I'll remove docs.google.com from the blacklist by default.

Comment 21

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

(In reply to Hsin-Yi Tsai [:hsinyi] from comment #19)
> Hi Florin,
> Could your team help to verify this issue that has been no longer valid on
> these websites?
> - medium.com

Ah, no, medium.com didn't block this bug. They blocked bug 1483553 which blocked bug 1440189.

Comment 22

[Vlad Lucaci, Thunderbird QA (:vlucaci)]

Hello guys,

We no longer encountered the issues described above on Google Docs,rememberthemilk.com,Google Email.

Testing was done on the following platforms:
- Windows 10x64
- macOS 10.12.6
- Ubuntu 16.04x64

And on the following builds:
- 65.0a1 (2018-11-25)
- 64.0b12

Comment 23

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Vlad Lucaci (:vlucaci) from comment #22)
> Hello guys,
> 
> We no longer encountered the issues described above on Google
> Docs,rememberthemilk.com,Google Email.
> 
> Testing was done on the following platforms:
> - Windows 10x64
> - macOS 10.12.6
> - Ubuntu 16.04x64
> 
> And on the following builds:
> - 65.0a1 (2018-11-25)
> - 64.0b12

Thanks :vlucaci :)

Comment 24

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Both Google and Remember The Milk fixed the issues!