Le 17/01/14 13:00, Vivien Nicolas a écrit :
What's Andrew nickname on IRC? I guess we can just spoke and figure
that out.
ahal
On 17/01/2014 03:54, Zac Campbell wrote:
Vivien, since I wrote this email I talked to Andrew Halberstadt in
person and he has a good plan for it which uses some existing logic and
knowledge that's in the mozharness package, which seems to make sense
and saves us reinventing the wheel.
I don't know which is the best solution but I will help where I can.
On 16/01/14 11:46, Vivien Nicolas wrote:
We have already a patch that reports css warning, js errors, reflows
and can report the total memory used by an app, or more precise things
like how much memory is used for the dom, per compartment, etc...
We need to clean it up and land it. As I said one of my goal is to
check-in those stuffs during tests so we won't regress anymore and the
other goal is to activate those tools by default for engineering
builds.
So if someone introduce a regression it will see it while developing
as a little colored square will pop up into his face and stays here
forever until the app is restarted or until the developer fix the
issue ;)
Zac we have not started the integration with the test framework yet
but it should not be that difficult as we probably just need to output
the right format to logcat right ?
The patch also needs to be cleaned up in order to land.
If you want to speak more about that let's do a meeting.
Vivien.
On 16/01/2014 03:28, Zac Campbell wrote:
What's the ETA on that Vivien?
I've offered to work on this which make take 3-4 weeks or so but your
solution sounds like a better long term solution.
Zac
On 15/01/14 17:52, Vivien Nicolas wrote:
For what it worth I'm working with some devtools guys to reports JS
errors, CSS warnings and reflows to the system app. The goal is to
let
us write add that as a new kind of tests to Travis so people will see
that as a failure and won't be allowed to merge when doing a PR.
On 15/01/2014 18:29, Kevin Grandon wrote:
+1 to killing all the errors. It may be nice if some day our core
apps, could run in other environments, but I think that is such a
low
priority that a single css error in the logs should trump all of
that.
-Kevin
----- Original Message -----
From: "Fabrice Desré" <fabr...@mozilla.com>
To: "Julien Wajsberg" <jwajsb...@mozilla.com>, "Hubert Figuière"
<h...@mozilla.com>, mozilla-dev-...@lists.mozilla.org
Sent: Wednesday, January 15, 2014 9:24:22 AM
Subject: Re: [b2g] Proposal: Cracking down on arbitrary Javascript
exceptions
On 01/15/2014 09:09 AM, Julien Wajsberg wrote:
CSS errors are another story, especially when we use CSS properties
that
are prefixed/unprefixed and we try that the apps somehow load in
other
browsers. But we definitely have some that needs to be fixed.
That doesn't apply to certified apps that have no chance of running
somewhere else. I cringe everytime I see a "syntax error, rule
ignored"
coming from a core app.
So yes, please kill them all.
Fabrice
_______________________________________________
dev-b2g mailing list
dev-b2g@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-b2g
_______________________________________________
dev-b2g mailing list
dev-b2g@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-b2g
_______________________________________________
dev-b2g mailing list
dev-b2g@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-b2g
_______________________________________________
dev-b2g mailing list
dev-b2g@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-b2g
_______________________________________________
dev-b2g mailing list
dev-b2g@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-b2g
_______________________________________________
dev-b2g mailing list
dev-b2g@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-b2g
