[Go Faster] System add-ons and safe mode
mconnor at mozilla.com
Mon Sep 14 19:26:13 UTC 2015
I don't think startup performance should matter that much in safe mode.
It's a diagnostic and debugging mode, no? That said, I'm fine with wait
and see if it's less work.
In terms of a checkpoint for this decision, are we tracking a set of things
for SUMO to watch out for? If so, we should add this... if not we should
probably start creating that list!
On 14 September 2015 at 15:07, Dave Townsend <dtownsend at mozilla.com> wrote:
> On Tue, Sep 8, 2015 at 12:47 PM, Mike Connor <mconnor at mozilla.com> wrote:
>> My mental model here is that system add-ons are indivisible
>> (conceptually) from the base app, except that we can push updates out of
>> band from core release cycles. To that end, I don't think reverting to
>> release-time versions makes the most sense, since updates may include
>> security/crash/other critical fixes that would be chemspills in a
>> monolithic model.
>> I would counter-propose that "Safe Mode" should load the latest stable
>> version. That could involve downloading the stable version (if there's an
>> update we haven't applied), which is okay, because the goal is to get users
>> to the latest known good configuration.
> I don't want to block startup every time the user runs safe mode just to
> check whether there are new system add-ons to install, particularly since
> that will probably involve building new UI to tell the user what is going
> on. We could schedule a check quickly after startup, that is planned for
> after app updates anyway (
> https://bugzilla.mozilla.org/show_bug.cgi?id=1193411) but that still
> leaves us with the question of what add-ons to run during startup. Unless
> there are strong objections I'm going to go with my original suggestion, we
> can revisit this later if we think it's the wrong choice.
>> On that note, I also wonder if Safe Mode should update to the latest
>> version as a part of the launch process. Obviously that's scope creep, but
>> in terms of "get a user to a working configuration" updating to latest
>> would seem like an improvement for users. Something to think about, at
> I don't know that that's necessarily true since some users actively avoid
> updating to keep things the way that they like them, but like you say,
> scope creep for this.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Gofaster