hi all,

On 11/13/2015 08:58 AM, Tim Guan-tin Chien wrote:
> Michael,
> 
> Probably the way to get the best of the both worlds is:
> 
> 1. Allow user to access Gecko about: pages from System Browsers (which
> loads with chrome permissions), avoid re-implementing every page in Gaia.

You mean by granting chrome privileges to these pages? And be happy with
all the xul they use? What are "System Browsers" ?

> 2. Allow customization from System app to overwrite specific about:
> pages like netError (and have these pages loads in System app permissions).

Which is what we have now.

> I don't like the fact System/Settings becoming some kind of a special
> app where we dump things no-where-to-put into it, but I agree your use
> case on netError is valid.

The system app is special - it's part of the platform actually, as a
first class embedder.

> I just don't like the fact we are creating extra work by putting
> permission barrier and then break it with hacks.

Not sure what you mean there.

> The alternative is probably grant chrome permissions to System &
> Settings*, but that won't do us any good either. So I ask instances of
> (2), including index.html of System app (which is essentially our
> browser.xul), to be properly engineered with WebIDL-enforced APIs.
> 
> * Settings app is just like our "about:preferences" in some way.

It's about:settings actually ;)

> Have (1) and (2) co-exists is a state I can live with, again, if (2) are
> properly worked on.

Doing (2) properly (if I understand you, which I'm not sure) means:
- going through the list of about: pages we want to support.
- for each, decide on the best implementation strategy. I hear that you
prefer to create new webapis for that.

Overall it's all well and fine to discuss how to implement about: pages,
but honestly that doesn't seem like the highest priority work item. We
have much more important gaps to fill that we need people to focus on.
Our browsing experience is way behind desktop and fennec in a number of
areas: no session restore, no form autofills for instance.
So I'd like us to *not* spend time & energy on that for now. But I would
be perfectly happy to help contributors work on that if they chose to.

        Fabrice
-- 
Fabrice Desré
b2g team
Mozilla Corporation
_______________________________________________
dev-fxos mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-fxos

Reply via email to