1213604 - Left-floated content is shifted to far right, at jmp.com support page

Status: RESOLVED WORKSFORME

(Core :: Layout: Floats, defect)

Description

[2ljpswpih]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:44.0) Gecko/20100101 Firefox/44.0
Build ID: 20151010030234

Steps to reproduce:

Go to http://www.jmp.com/support/help/Join_Data_Tables.shtml


Actual results:

The page is not shown properly.


Expected results:

The page should be shown properly.

Comment 1

[Andre Klapper]

Site is shown properly here in Firefox 41. Cannot reproduce the problem.

For future reference, please read https://developer.mozilla.org/en-US/docs/Mozilla/QA/Bug_writing_guidelines and add a more useful description to this report. Thanks!

Comment 2

[Andre Klapper]

If "The page is not shown properly", please describe what that means. Or provide a screenshot.
Please also test with a fresh new profile and with any extensions/add-ons disabled.

Comment 3

[2ljpswpih]

Attachment: Here is the screenshot of a problem

Comment 4

[2ljpswpih]

"Site is shown properly here in Firefox 41." I wrote version of Firefox earlier - Firefox Nightly.

Comment 5

[2ljpswpih]

It is also bad displayed in latest Firefox Beta, also in safe mode, also with fresh profile. 
For future references, please carefully read information contained in the bug report.

Comment 6

[2ljpswpih]

Please don't so quickly make decisions about "cancelling" the bug, because this bug exists.

Comment 7

[2ljpswpih]

Today update didn't fix the problem.

Comment 8

[2ljpswpih]

Today update didn't fix the problem.

Comment 9

[2ljpswpih]

In the latest beta build (42.0b6) bug persists.

Comment 10

[2ljpswpih]

This bug exists all the time.

Comment 11

[Simona Badau, Desktop QA, PTO 22-30June]

Mozilla/5.0 (Windows NT 6.1; WOW64; rv:45.0) Gecko/20100101 Firefox/45.0
Build ID: 20151108030417

I can reproduce the issue on the latest Nightly. Setting the status of this bug to NEW.

Comment 12

[Boris Zbarsky [:bzbarsky]]

Simona B, can you hunt down a regression range for this with mozregression, since this is clearly a regression between 41 and 42?

Comment 13

[Daniel Holbert [:dholbert]]

I've already got mozregression going; I'll have a range shortly.

Comment 14

[Daniel Holbert [:dholbert]]

mozilla-central regression range is:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=32712cd01159&tochange=5b54831761b1

Strongly suspect bug 538194, which had to do with float handling. (And the right-shifted content here is inside of a "float: left;" element, with id="left".

Comment 15

[Daniel Holbert [:dholbert]]

Inbound regression range confirms that this is a regression from bug 538194:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=838a7d98c1d41c2bf67497d0a80cca98a6001899&tochange=5dcb38c7f1b84d7fc4d0255a2b10165dfaceb941

(The other patches in that push are about transform/perspective & seem unlikely to be involved here.)

Tagging dbaron as needinfo. Also, a reduced testcase would likely help to isolate the problem, if anyone is up for trying to create a smaller testcase by hacking away at the jmp.com page. (Ideally the testcase should "work" in Chrome & old Firefox versions, but "break" in new Firefox versions).

Comment 16

[Ryan VanderMeulen [:RyanVM]]

Haven't reduced it down, but deleting the menu bar at the top ("JMP Software", "Using JMP",...) using the Inspector makes the main page content shift back to the center.

Comment 17

[Ryan VanderMeulen [:RyanVM]]

Attachment: more reduced testcase

Not all the way reduced yet, but it's a start. The offsetting appears tied to the list elements.

Comment 18

[Ryan VanderMeulen [:RyanVM]]

...except the mixed content blocker is blocking the external style sheet. *sigh*

Comment 19

[Ryan VanderMeulen [:RyanVM]]

Attachment: reduced testcase 1, for Chrome/Firefox difference (though not for Firefox 41/42 difference)

This is basically the most-reduced I can get the page where there's still a rendering difference between Chrome and Firefox. That said, it renders the same in Firefox 41 and 42, so I'm not entirely sure it's the "right" testcase. Hopefully it helps.

Comment 20

[Ryan VanderMeulen [:RyanVM]]

Attachment: reduced testcase 2, for Chrome/Firefox 41/Firefox 42+ differences

This testcase shows the difference between 41 and 42+. Interestingly, it also renders differently in Chrome. It might be possible to further-reduce this, but this is already pretty well-minimized and I don't intend to go any further with it.

Comment 21

[2ljpswpih]

It looks like the bug is fixed.

Comment 22

[Daniel Holbert [:dholbert]]

Perhaps the site has been fixed? I don't think Firefox's behavior changed here.

In particular, I can't reproduce the bug in old nightly 20151108030417, which was previously reported as reproducing the bug (comment 11).  And the attached reduced testcases render the same in current Nightly as they do in that old Nightly 20151108030417.

Comment 23

[Ryan VanderMeulen [:RyanVM]]

Yes, looks like the site changed to work around whatever issue we had. The attached testcases still show the behavior differences as described.

Comment 24

[Daniel Holbert [:dholbert]]

Looks like our behavior remains the same as it was the day this bug was filed, and Chrome has changed behavior to match us.

I tested a few old versions and determined:

• Chrome 44 renders both attached testcases differently from us.
• Chrome 54 renders testcase 1 the same as we do, but testcase 2 differently from us.
• Current Chrome (v102 dev) renders the attached testcases the same as we do.

So our behavior is interoperable (now that Chrome has changed), and the site was probably depending on a now-fixed Chrome bug.

Comment 25

[Daniel Holbert [:dholbert]]

(I also confirmed that the original site still renders properly, as observed in comment 22 - comment 23.)