W3C

- DRAFT -

Browser Extension CG teleconf

16 Feb 2017

Agenda

See also: IRC log

Attendees

Present
mikepie, John-Gal2, Florian, scottlow
Regrets
Chair
Florian
Scribe
Florian

Contents


<scribe> ScribeNick: Florian

* Status update and issue resolution on https://browserext.github.io/browserext/

mikepie: I haven't had a chance to go through the IDs questions yet
... But other than what we have on the call today that's the only thing left, so we'll look into it
... on issue 15, we're past the 10 days after resolution, so we can close
... Next issue, we had already specified that the object returned by webdriver's getBrowserExtension method had an id, and I also specified that it should have the nae

name

kmag: This needs to be localized
... Not sure which locale to use though

Florian: This is a setting that is often wrong

kmag: if we want to match the browser's UI locale, we need to expose that

Florian: In some cases you may want to match the site rather than the UI

scottlow: since the ID is an inpredictable number, you need a name to be able to know which extension your talking to

Florian: If that's what its for, it doesn't need localization, it needs predictability

mike: so we can use the default locale

RESOLUTION: Add the name property, and define it to be the default locale

^ that was issue 31

mikepie: Issue 46: This is about interacting with permission prompts
... I made 3 additions
... the API lets you get the localized text of the permission prompt, and based on it accept it or dismiss it

kmag: I'm worried about text matching.

florian: what options do we have

kmag: we have an permission name string, and return that in json
... these names are already standard

<mikepie> requestingUrl

mikepie: I'll replace that string with a JSON object, including "requestingUrl"...
... Also "permissions"

Florian: Why plural?

kmag: I think there can be multiple permissions in one dialog

mikepie: Yes, at install time, an extension can ask for multiple permission

scottlow: are these optional permissions?

kmag: in our case, geolocation is a bit special, but otherwise yes

scottlow: are these about security permissions as well?

kmag: no

Florian: We will deal with run-time permissions with this as well?
... what if there's multiple ones queued up

kmag: you would have to accept or dismiss them in the order they are being displayed at. You cannot get the list of all queued permissions at once

Florian: Don't we have a risk to make author depends on browser specific ordering of the permission prompts

kmag: you do, but that's the sanest thing we can do anyway.
... this can only happen if requesting multiple permissions at the same time

Florian: Yeah, this isn't great, as we may get authors to depend on the ordering even though they shouldn't, but I cannot think of a better alternative

mikepie: I'll add a note to call it out
... and the third property is the prompt text

RESOLUTION: add requestingURL, permissions and promptText
... the name of the method is getBrowserPermission (without Text)

kmag: we should also return a list of possible actions, such as "accept for session" in addition to just "accept" and "dismiss"

mikepie: Yes, I can add that. I like it.

RESOLUTION: add "actions", which must include "accept" and "dismiss", and and possibly other actions as available

Florian: we need to come of with names for these

mikepie: This is also something you'd want to handled declaratively

kmag: that worries me
... I think the default one should not be "dismiss" temporarily, it should be "deny"

mike: this should be for the session, right?

mikepie: with webdriver, the install are just for the session

<mikepie> mandatory: accept; deny (for session? permanent?)

mikepie: session and permanent is the same, since the extension is only installed for the session

Florian: right, but deny and dismiss are different

kmag: and we should go with deny rather than dismiss as the mandatory one

RESOLUTION: replace dismiss with deny

Florian: do we need to come up with names?

mikepie: since there's not difference between permanent and for this session only, allow, deny and dismiss are all we need

Florian: so we only have 2 possible cases at install time, and 3 at run time

kmag: at install time, the only choise is really install extension or not

mikepie: I'll add a session to explain what kind of thing can appear there

Florian: if there can be other names, do we need to standardize?

all: yes, but not now.

Florian: mikepie, you can merge the PR once the changes are in

Next meeting

mikepie: How about the ides of March

all: agreed.

Summary of Action Items

Summary of Resolutions

  1. Add the name property, and define it to be the default locale
  2. add requestingURL, permissions and promptText
  3. add "actions", which must include "accept" and "dismiss", and and possibly other actions as available
  4. replace dismiss with deny
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.148 (CVS log)
$Date: 2017/02/16 02:20:37 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.148  of Date: 2016/10/11 12:55:14  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/name/names/
Found ScribeNick: Florian
Inferring Scribes: Florian
Present: mikepie John-Gal2 Florian scottlow
Agenda: https://lists.w3.org/Archives/Public/public-browserext/2017Feb/0003.html
Got date from IRC log name: 16 Feb 2017
Guessing minutes URL: http://www.w3.org/2017/02/16-browserext-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


[End of scribe.perl diagnostic output]