New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 682260 link

Starred by 1 user
Status: Fixed
Owner:
█ bmcquade@chromium.org
Last visit > 30 days ago
Closed: Jan 2017
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug



Sign in to add a comment

doc.write blocking for effective 2g shouldn't trigger when on 2g but not effective 2g

Project Member Reported by bmcquade@chromium.org, Jan 18 2017 
The current impl for doc write blocking on effective 2g will trigger if the user is on 2g or on effective 2g.

This can make experiment result analysis confusing as there are multiple different triggering cases.

We should simplify to only trigger if actually on effective 2g as determined by NQE.

We can always get the prior behavior by enabling both 2g and effective 2g triggering in the same experiment.
Project Member

Comment 1 by bugdroid1@chromium.org, Jan 18 2017 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/0b460c475f6f6fba74d9ba731398a1ce5328d4b0

commit 0b460c475f6f6fba74d9ba731398a1ce5328d4b0
Author: bmcquade <bmcquade@chromium.org>
Date: Wed Jan 18 22:40:29 2017

Restrict effective 2g blocking to cases where NQE says we are 2G-like.

The current impl triggers the effective 2g blocking if we're on either
2g or effective 2g.

This change simplifies things so we only trigger effective 2g blocking if
on effective 2g.

This makes experiments cleaner, as we can evaluate the impact of this
feature when NQE determines we're on 2G.

BUG= 682260 

Review-Url: https://codereview.chromium.org/2643773002
Cr-Commit-Position: refs/heads/master@{#444520}

[modify] https://crrev.com/0b460c475f6f6fba74d9ba731398a1ce5328d4b0/third_party/WebKit/Source/core/loader/FrameFetchContext.cpp

Comment 2 by bmcquade@chromium.org, Jan 27 2017

Status: Fixed (was: Started)
►

Sign in to add a comment