21193 - [FLOAT] Half-size Box Acid Test Layout broken due to rounding

Status: RESOLVED FIXED

(Core :: Layout: Floats, defect, P2)

Description

[Jikes Incorporated]

Oh dear.  Here comes another **** bug report.

Problem: http://style.verso.com/boxacidtest/half/  does not lay out correctly.
Standard 100% Box Acid Test does fine, 200% Box Acid Test does fine, 50% Acid
Test barfs, throwing boxes all over the place.

Platform: Win 95
Build: Daily, 12-8-99
Bug Reporter: Clueless, lazy, and at work without source.

Whether this is due to icky Win32 widget barfing, or actual layout problems, I
do not know, but it won't automagically fix itself no matter what I toggle or
manipulate, and it IS supposed to lay out correctly.

When I get home I will whack on it with a CVS build for Linux.  Please please
please fix CSS1!  There is a special place in Heaven for all you saints who are
working on Mozilla.  Thank you!

Comment 1

[troy]

Looks like a problem in the block code

Comment 2

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

My memory is that this is actually not a bug, and the half-size boxacidtest is
invalid, because of rounding issues.

Marking as duplicate of bug 4098.  If you want this dealt with, reopen that bug
** and reassign it to the default owner of the component **.

*** This bug has been marked as a duplicate of 4098 ***

Comment 3

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

Actually, I'm reopening this, because it works fine on Linux, and is therefore a
platform parity issue.  Marking as such.

Comment 4

[leger]

Clearing DUPLICATE resolution due to reopen.

Comment 5

[leger]

Updating to default Layout Assignee...kipp no longer with us :-(

Comment 6

[troy]

Why are you re-reassing layout bugs? Do NOT touch layout bugs.

The bugs are assigned to Kipp so they can stay neatly organized until we have a
new owner for the block/inline code.

Comment 7

[buster]

mass moving all Kipp's pre-beta bugs to M15.  Nisheeth and I will
prioritize these and selectively move high-priority bugs into M13 and M14.

Comment 8

[buster]

mine! mine mine mine!  all mine!  whoo-hoo!

Comment 9

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

This is currently a problem on Linux.  Removing pp keyword, adding css1 keyword,
marking All/All, adding dependency to boxacidtest tracking bug, and marking
[FLOAT].

I think the correct solution to this bug is in the float positioning code:  If a
float doesn't fit horizontally in its parent by less than half a pixel (or
perhaps it should be done by whether the edge would be rounded to outside the
parent / over another float when displayed, if that's possible), then I think it
should not be carried to the next "line" of floats.

I think this is a real problem, since people do use floater-based structures to
make columns using CSS.

Comment 10

[rickg]

moving all buster m15 bugs to m16.

Comment 11

[buster]

pushing off until after beta2

Comment 12

[buster]

redistributing bugs across future milestones, sorry for the spam

Comment 13

[ekrock's old account (dead)]

Nom nsbeta3, recc nsbeta3+. This test suite is highly influential and closely 
watched in standards compliance circles. css1 compliance issue.

Comment 14

[Hixie (not reading bugmail)]

As per meeting with ChrisD today, taking QA.

Comment 15

[Jason Eager]

Blah! Anyone actively working on this?

Comment 16

[Marc Attinasi]

Taking a stab at prioritizing buster's nsbeta3 bugs...

According to IanH, this is a rounding problem caused by half-pixel values - 
deemed edge-case.

Comment 17

[buster]

marking future:  not enough time to get to this edge case before RTM.  if 
someone else wanted to look into adding pixel rounding, I would be happy to work 
with them on it.

Comment 18

[karnaze (gone)]

Moving to m1.0.

Comment 19

[Hixie (not reading bugmail)]

anyone remember what Todd's current server is called?

Comment 20

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

We need to change nsBlockReflowState::CanPlaceFloater (near the beginning) to
allow a half pixel of overflow.

Comment 21

[Kevin McCluskey (gone)]

Build reassigning Buster's bugs to Marc.

Comment 22

[karnaze (gone)]

Changing url and moving to m1.1.

Comment 23

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

Changing URL back to the one that works.

Comment 24

[unapersson]

The URL given appears fine in both 1.0 and the latest nightly (on Win NT4).
Should this bug be marked fixed?

Comment 25

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

I think it would be easy to construct testcases that are still broken.

Comment 26

[Arthur]

The rounding issue is still there on linux 1.2.1 with text zoom 100%. Try
playing with the text zoom and it'll display sometimes just fine and sometimes
messed up.

Comment 27

[Andrew Schultz]

==> floats

Comment 28

[louis bennett]

perhaps of interest to some - the layout switches between proper and improper
rendering when you increase or decrease font size. (1.3a / os x)

Comment 29

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

*** Bug 195505 has been marked as a duplicate of this bug. ***

Comment 30

[Ronald Tilby]

Attachment: XHTML Testcase Floating Tables Width sums to 100%

I think this testcase is a good demonstration of the problem.

Comment 31

[Vedran Miletic]

retargeting

Comment 32

[Travis Chase]

Is this bug still a problem? the URL for the test case is long gone,

Comment 33

[Ronald Tilby]

Although the url in comment 0 is dead, the URL in the url field of the bug
(http://style.cleverchimp.com/boxacidtest/half/) is still alive, and still fails
to render properly using Mozilla 1.7Alpha.  The XHTML Testcase in the
attachments also still exhibits the problem.

Comment 34

[Julien "_FrnchFrgg_" RIVAUD]

I agree there is a problem with rounding here. But I am not sure the testcase
description is good: It expects the each whole div to fit on a line.

But, IMHO, the sum of table's widths is 100%; so calculating I get:

Total width of n tables stacked one after the other on one line =

sum for each table(width + borders + padding)

=

100% div's width + n * 1px > parent div's width

(for what I reckon, either you decide that the half border of collapsed borders
table bleeds and as such can overwrite the other border; then sum = 100% OK, or
you decide to enlarge/inflate table's margin to prevent overlapping, and then
the previous calculus is good)

So the bug would be solved by (I think -- yes I do... sometimes)
a) Solving the "borders of collapsed table" bug
(https://bugzilla.mozilla.org/show_bug.cgi?id=155955)
b) Solving the "rounding errors" bug (do not recall the number)

_FrnchFrgg_

Comment 35

[Ronald Tilby]

The testcase is now WFM, but the URL still fails using Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20061208 Minefield/3.0a1 ID:2006120804 [cairo]

Comment 36

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Attachment: screenshot of URL link

The testcase now WFM on Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9) Gecko/2008053008 Firefox/3.0 and the URL shows a correct half-size image.

Should this be resolved WFM? (This is the last bug blocking tracking bug 8914)

Comment 37

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

Might this have been fixed by bug 287624 or (more likely) bug 417178?

Comment 38

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

Fixed between Linux nightlies 2008-03-24-04-trunk and 2008-03-25-04-trunk, thus by bug 417178.