On Monday, March 7, 2022 at 3:47:11 PM UTC+1 Mike Taylor wrote:

> On 3/7/22 8:45 AM, Yoav Weiss wrote:
>
> On Mon, Mar 7, 2022 at 2:35 PM Arthur Sonzogni <
> [email protected]> wrote:
>
>>
>>
>> We contacted several websites. Some confirmed they will be positively 
>>>> impacted. 
>>>>
>>> Microsoft Teams, the #1 requested adding 
>>>> "allow-top-navigation-to-custom-scheme" to opt-out from this feature 
>>>> individually.
>>>>
>>>
>>> It's reassuring to hear they'd be willing to use an opt-out.
>>> Have you reached out to others on that list? It seems like tackling the 
>>> top ~3 would significantly reduce the risk.
>>>
>>
>> I confirm I sent an email for every website we see: the top 8 on Windows 
>> and a few others on Android. I asked them to confirm reception.
>>
> Thanks - looking at the list, some of them seem to have obvious "open a 
> native app" use-cases, like Microsoft Teams. Others... not so obvious.
>
> If the primary motivation is to break malicious ads, would it make sense 
> to further require a secure context for the sandboxed iframe? (Maybe this 
> question is nonsense, I'm not a security expert). Do we know what the 
> breakdown is between http/non-secure and https/secure for iframes causing 
> these kinds of external app navigations?
>
Yes, this is probably worth doing. However, this would be a different 
topic, spec & impacted users. I think this would be more of an additional 
project, and ideally we would do both if possible.
There are no metrics about security/non-secure context that I am aware of.

> Please make sure you use CountDeprecation 
>>> <https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/platform/instrumentation/use_counter.h;l=45?q=Deprecation%20UseCounter&ss=chromium>,
>>>  
>>> to get deprecation reports sent automatically. (unless there are particular 
>>> technical reasons that make this more complex than it should be. If that's 
>>> the case, let's discuss more)
>>>
>>
>> This can't only be used in blink unfortunately. This is about navigation, 
>> and happens when we are out of handlers, before calling external 
>> applications. I can't make use of CountDeprecation here. It is not 
>> impossible to create new IPCs and get it plumbed to the renderer process, 
>> but that's probably more complex than what it deserves.
>>
>
> That's probably true when considering only this particular intent, but 
> might be worthwhile to create the infra to enable this as this seems to be 
> a recurrent pattern.
>
> Agree it sounds like a useful tool to have down the road. I can't speak 
> for Yoav, but can we at least get a bug on file with enough context for 
> someone else to pick up?
>
That's the reasonable thing to do. Here is the bug:
https://bugs.chromium.org/p/chromium/issues/detail?id=1303603

-- 
You received this message because you are subscribed to the Google Groups 
"blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/a/chromium.org/d/msgid/blink-dev/fcf7bcdf-aef4-4dcb-a21b-6b5c1c5727fan%40chromium.org.

Reply via email to