188634 - Table moves when page is reloaded

Status: RESOLVED WORKSFORME

(Core :: Layout: Tables, defect, P3)

Description

[Aaron Leventhal]

Go to http://www.jsonline.com/sports/buck/
Notice the positioning of the sidebar "Bucks, NBA Game Tracker"

Click on the "Go to Story" link, and then hit Alt+Left arrow to go back in
histroy when that page loads.
Notice the new positioning of the sidebar, it's different now.

Comment 1

[Aaron Leventhal]

More info:
1. I am using January 6, 2003's build
2. Sometimes reload helps, sometimes it doesn't
3. Clicking on a link outside the site and coming back doesn't seem to create
the problem. You have to click on one of the Bucks story links and then come
back to see the problem. Even then it doesn't always seem to happen on every
link. Perhaps it depends on the positioning of the "Game tracker" sidebar in the
new page.
4. Turning of the memory and disk cache doesn't help fix the problem

Comment 2

[Aaron Leventhal]

Bernd on IRC said he saw the problem too.

Comment 3

[Aaron Leventhal]

Also note, sometimes force-reload fixes the page, sometimes it doesn't

Comment 4

[Aaron Leventhal]

It looks like the sidebar is a table, nested inside another table.

Comment 5

[Aaron Leventhal]

Attachment: One kind of corruption, sidebar overwrites other content

Comment 6

[Aaron Leventhal]

Attachment: Screenshot #2. Another type of corruption, with sidebar above rest of page content

Comment 7

[Aaron Leventhal]

Aha, I've found that just reloading the page can make that sidebar move around.

Comment 8

[Andrew Schultz]

Attachment: testcase

reproduced with linux trunk build 20030111.  images from the URL are <div>s.

the table with the green <div> extends into the second column of the parent
table.

Comment 9

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

But isn't the original bug report an incremental reflow bug, where a page
displays differently when loaded originally?

We have some quirks for placement of tables near floats.  This testcase might
just be demonstrating one, although I'm not sure.

Comment 10

[Andrew Schultz]

indeed.  with the URL, changing font size fixes layout (reloading the page never
fixed it for me), but that doesn't work with the testcase.  I made the testcase
based on the initial layout.

I'll see if I can come up with a testcase that also exhibits the incremental
reflow problem.

Comment 11

[Andrew Schultz]

Attachment: testcase with reflow problem

this is similar to the first testcase but fixes itself if you increase font
size and stays fixed even if you then decrease the font size back to normal. 
However, if you then decrease it again, the bug shows up again and stays even
if you increase font size back to normal.

Comment 12

[Andrew Schultz]

Comment on attachment 111313 [details]
testcase with reflow problem

wrong testcase, argh

Comment 13

[Andrew Schultz]

Attachment: testcase with a reflow problem

the real thing

Comment 14

[Bernd]

this looks like wfm to me, please reopen but describe what the problem is in attachment 111315 [details], which we render like webkit and opera