64817 - Mozilla frame rendering is wrong when having more than 50 frames?

Status: RESOLVED INCOMPLETE

(Core :: Layout: Images, Video, and HTML Frames, defect)

Description

[Xuân Baldauf]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; m18) Gecko/20010106
BuildID:    2001010620

When you see the two choices at
http://justchat4.medium.net:4000/test/netscapeCommunicatorFrameTest.html ,
display differences should be very less and subtile, as it is in Internet
Explorer 5.5. There should be only one black area more. But instead of rendering
correctly, mozilla does not take the col #25 (and any subsequent cols) into
account for rendering, it scales all the frame column widths up to window width,
even if the frame column width is only 5 pixels and the window width is 1024 pixels.


Reproducible: Always
Steps to Reproduce:
Create a nested frameset of 25+1 columns, where every column occupies a width of
5 pixels and the last column occupies the remaining width of the window. Every
of the 25 columns is actually a frameset consisting of two rows, where the first
row is black and the second row is white. The second row always occupies the
height the first row does not use. The first row of the first column occupies 0
pixels height, the first row of the second column occupies 1 pixel height, the
first row of the third column occupies 2 pixels height, and so on...

Actual Results:  Your browser should render a stair-like figure, where every
stair occupies 5 pixel in width.

Expected Results:  Mozilla renders a stair-like figure, where every stair
occupies 1/25th of the width of the window, so the effective width of one stair
can be 40 pixels (if the window width is 1000 pixels).

This test case was actually developed to find frame height values which will not
be altered by the famous Netscape Communicator 4.7 frame alignment bug. To
understand what I mean, simply use Netscape Communicator to view the test case
pages.

JavaScript has to be turned on at the viewing browser.

Comment 1

[Boris Zbarsky [:bzbarsky]]

I see this on linux build 2001-01-05-08

Comment 2

[Amarendra Hanumanula]

QA Contact update

Comment 3

[Christopher Hoess (gone)]

This URL is 404.  Removing.

Comment 4

[Xuân Baldauf]

That URL should contain the same resource and be more persistent.

Comment 5

[Xuân Baldauf]

When horizontally resizing the window where all frames scale to the whole
horizontal size of the display area of that window (
http://mozilla.bugs.test.medium.net/bugzilla.mozilla.org/bug64817/netscapeCommunicatorFrameTest.25.html
), you will very probably see different rendering artefacts consisting of
vertical lines which are neither black or white when using Mozilla 2001090603 on
Win98SE. These rendering artefacts are not visible at all widths, but
at many ones, so just try 3 or 4 times if you do not see this. The count of the
vertical artefact lines seems to depend on a modulo-function of the current
window width (it changes with every change of the window width, most of the time
it increases with every increase of the window width).

These rendering artefacts also seem to apply to the "good" example for this bug
(
http://mozilla.bugs.test.medium.net/bugzilla.mozilla.org/bug64817/netscapeCommunicatorFrameTest.24.html
)

when lowering the width of the window below the prefered width of all frames
together (i.e. below 120 pixels in this case).

Comment 6

[Kevin McCluskey (gone)]

Bulk reassignin HTML FRAME/IFRAME bugs to Eric.

Comment 7

[Kevin McCluskey (gone)]

Bulk re-assigning all of Eric's HTMLFrame bugs to John.

Comment 8

[Wayne Mery (:wsmwk)]

http://mozilla.bugs.test.medium.net/bugzilla.mozilla.org/bug64817/netscapeCommunicatorFrameTest.24.html
is gone
=> incomplete