12401 - {css1} weaker rule in linked stylesheet is winning

Status: VERIFIED FIXED

(Core :: CSS Parsing and Computation, defect, P2)

Description

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

DESCRIPTION:  The (red) left margin of the paragraph in the above URL disappears
after the page finishes loading.  It flashes its correct appearance first.

STEPS TO REPRODUCE:  Load above URL.

ACTUAL RESULTS:  Paragraph is black text on gray background, surrounded by gray
padding,  black border, and then surrounded by red on top, bottom, and right
(top and bottom thicker than right), BUT NOT LEFT.  The expected results do
appear quickly while the page is loading.

EXPECTED RESULTS:  Paragraph should be black text on gray background, surrounded
by gray padding, black border, and red on *ALL 4* sides.  (The red should be
thicker on the top and bottom than the right and left.)

DOES NOT WORK CORRECTLY ON:
 * Linux, apprunner, 1999-08-20-13-M10

ADDITIONAL INFORMATION:
Are you trying to do carried out side-margins or something?

Comment 1

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

I'm seeing the same problem in the second test paragraph in
http://www.editions-eyrolles.com/livres/glazman/Tests/boites/boites2.htm .

Comment 2

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

From the look of this, it could actually be style system, not layout.  Something
to do with margin-left vs. margin???  I'm also seeing it on
http://www.editions-eyrolles.com/livres/glazman/Tests/boites/boites3.htm

Comment 3

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Another one:
http://www.editions-eyrolles.com/livres/glazman/Tests/boites/boites15.htm

Comment 4

[troy]

Changing to style system since that's what it looks like

Comment 5

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

What's happening here is that a rule in the linked stylesheet with lower
specificity is winning.  This shouldn't happen.  Perhaps the embedded and linked
stylesheets are somehow being treated as different levels of the cascade?

Retitling bug.  Marking [TESTCASE].  In the attached testcase, the text should
be green, but it's red.

Comment 6

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Attachment: linked CSS file used by testcase

Comment 7

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Attachment: test case

Comment 8

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

ALSO DOES NOT WORK CORRECTLY ON:
 * Windows, apprunner, 1999-08-24-09-M10

Marking All/All.

Comment 9

[Hixie (not reading bugmail)]

[Just adding two radars...]

Comment 10

[Peter Linss]

Moving non-beta 1 items to M15

Comment 11

[Hixie (not reading bugmail)]

This is also tested in part 4 of the folowing EvilTests test:
   http://www.bath.ac.uk/%7Epy8ieh/internet/eviltests/cascade/easy.html

This seems like quite a fundamental problem. Is it a recent regression?

Comment 12

[Peter Linss]

Not a regression. It was implemented this way because that's the way I
interpreted the spec.

Comment 13

[Chris Waterson]

Hyatt says this bug has cost too many man hours; he and cmanske just spent two
each chasing false leads before realizing that someone had checked in some bad
editor css, which was having unintended effects on more specific rules.

This shouldn't be M15 -- can it be fixed ASAP, even before the content sink
high colonic?  Can someone else help fix it?

/be

Comment 14

[leger]

Moving to M11...M10 is out the door.

Comment 15

[Christine Hoffman]

Testing using 10/21 app on Win, Mac and Linux with original test case, verified
bug fixed. Red margin displays on page load.

Comment 16

[Hixie (not reading bugmail)]

Migrating from {css1} to css1 keyword. The {css1}, {css2}, {css3} and {css-moz}
radars should now be considered deprecated in favour of keywords.
I am *really* sorry about the spam...