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

Issue 727821 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner:
Closed: Jun 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows , Mac
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Middle Click Navigation is Broken

Reported by asmoor...@gmail.com, May 30 2017

Issue description

UserAgent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36

Steps to reproduce the problem:
1. Find any webpage with a mixture of regular links and javascript event links.
2. Middle-clicking the regular links still behaves normally (opens link in new tab in the background).
3. Middle-clicking javascript event links does nothing.

What is the expected behavior?
Consistency!

Some folks somewhere decided (after 15+ years) that a raw middle-click shouldn't throw a click event. Well that's all well and good and technically true. But the problem is that there is an expected behavior for middle-click in all modern browsers and for all practical purposes there is no such thing as a true "raw" middle-click action. If the browser itself is deciding to have a feature where for all regular links the middle-click should be exactly as a regular click but open a new tab then this should remain consistent with the javascript click event. If the end user wishes to turn off the "middle-click to open new tab" feature then by all means consider it a "raw" middle-click and throw no click event.

As it stands right now my browser is completely broken as I look across a sea of links and have to guesstimate which ones will middle-click and which ones won't *even on the same page*.

What went wrong?
Inconsistency.

It's one thing to ask what event should a middle-click throw?

It's an entirely different conversation to ask how does our browser itself interpret the middle-click before the event is even handed off to javascript?

Yes, I am aware that control+click does the action that I prefer. But I shouldn't need to resort to a 2 handed keyboard  and mouse operation to simply do what *still* works on plain links! But now follow that control+click logic through. If the original "problem" that this change solves is that some javascript somewhere was mis-interpreting the middle-click, how long until some javascript somewhere is mis-interpreting the control+click? Has any real problem been solved here?

Did this work before? Yes 54?

Chrome version: 58.0.3029.110  Channel: stable
OS Version: Ubuntu 14.04
Flash Version: 

Firefox has done the same thing with the same insane results - links *within the same page* behave differently depending on regular or javascript. To an end user links are all the same!
 
Cc: ligim...@chromium.org
Labels: Needs-Triage-M58 Needs-Bisect

Comment 2 by asmoor...@gmail.com, May 31 2017

Another noteworthy fact about these javascript links that you might want to direct to a new tab is that you cannot right-click them and open in a new tab from there. Maybe a possible feature enhancement could be to detect things that have a click event handler and add "open in a new tab" to their context menu.
Cc: kavvaru@chromium.org
Labels: Needs-Feedback
asmoore82@ Thanks for the issue.

Could you please provide us any sample test file or URL to triage the issue from test team end.

Thanks,

Comment 4 by asmoor...@gmail.com, May 31 2017

I made this to show the simplest case of inconsistency: https://jsfiddle.net/rm6tbv93/7/
Project Member

Comment 5 by sheriffbot@chromium.org, May 31 2017

Labels: -Needs-Feedback
Thank you for providing more feedback. Adding requester "kavvaru@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
Components: Blink>Input
Labels: -Pri-2 -Needs-Bisect -Needs-Triage-M58 hasbisect-per-revision M-61 OS-Mac OS-Windows Pri-1
Owner: nzolghadr@chromium.org
Status: Assigned (was: Unconfirmed)
Able to reproduce the issue on Windows 10, mac 10.12.4 and Ubuntu 14.04 using chrome reported version #58.0.3029.110 and latest canary #61.0.3116.0.

Bisect Information:
=====================
Good build: 55.0.2853.0    Revision(416812)
Bad Build : 55.0.2855.0    Revision(417475)

Change Log URL: 
https://chromium.googlesource.com/chromium/src/+log/3eda8caca348d8c2c2f72d28af48e0757f83d1f2..7be81ca9b22fe6abb5a96d200f77167cae987b5f

From the above change log suspecting below change
Review url: https://codereview.chromium.org/2302513002

nzolghadr @ - Could you please check whether this is caused with respect to your change, if not please help us in assigning it to the right owner.

Thanks...!!
Cc: dtapu...@chromium.org
Labels: auxclick Hotlist-Input-Dev
Status: WontFix (was: Assigned)
I understand your frustration and the troubles that this change might have caused. However we also had a lot of people complaining about the other way (which was sending click event for middle button  issue 255 ). Either way some set of developers were unhappy and some websites were broken. But from our analysis at the time we realized the portion of people that don't expect click for the middle button was much more the other group. So we had to make a decision that makes the life of most developers/users/websites easier at the cost of a smaller portion of codes on the web needing to fix to use auxclick.

We tried our best to reach the developers before the feature launch including github issues, blog posts (https://developers.google.com/web/updates/2016/10/auxclick) and comments on the reported bugs. Also we had this feature behind the flag for quite sometime before 55. But we are hoping to get better at our outreach in the future.
The core of this issue doesn't really have anything to do with external web developers, their complaints, or outreach. At its core the browser's internal behavior is now simply inconsistent.

Is it intended for control+click and middle-click *links* behavior to be the same or no?

If you are set on removing/breaking common features that have been in web browsers for 15+ years, at least be consistent.
Regarding ctrl+click maybe this issue is related:
https://github.com/WICG/auxclick/issues/13

and we do have that in mind to send auxclick for any click action that is accompanied by something else (like control in this case).

Sign in to add a comment