Open Bug 153699 Opened 22 years ago Updated 2 years ago

[meta] Check CSS implementation against tests at RichInStyle.com

Categories

(Core :: CSS Parsing and Computation, defect, P5)

defect

Tracking

()

People

(Reporter: stf, Unassigned)

References

(Depends on 1 open bug, )

Details

(Keywords: meta, Whiteboard: wontfix/invalid per comment 16?)

http://www.richinstyle.com/test/

As requested, I'll make all those tests (from frame mode) in Chimera and Mozilla.
I'll post this document and the following one on MZ and CH bug report.

MZ : v1.1a BUILD 2002061014
CH : v0.31 BUILD 20020622 16:00
iMac DV 400Mhz - 1024x768 - 32bits
OS X 10.1.5 Build 5S66 French

Explanations e.g.:
"2 MZ+CH not red"
Point 2 in that page, is not red in MoZilla and CHimera.


=== Attachment of style to pages
http://www.richinstyle.com/test/attachment/uriresolution.php
2 MZ+CH not red

http://www.richinstyle.com/test/attachment/http.php
3 4 MZ+CH red

http://www.richinstyle.com/test/attachment/preferred.php
4 MZ+CH red

http://www.richinstyle.com/test/attachment/alternate.php
1 CH alternate style sheet can't be choosen in CH - not yet applicable

http://www.richinstyle.com/test/attachment/mediacase.php
1 2 MZ+CH red

http://www.richinstyle.com/test/attachment/atimportmedia.php
5 MZ+CH red

http://www.richinstyle.com/test/attachment/styleelement.php
4 5 MZ+CH red

http://www.richinstyle.com/test/attachment/styleattribute.php
3 4 MZ+CH blue
5 MZ+CH red

=== Application of style to elements
http://www.richinstyle.com/test/application/idvalidity.php
2 MZ+CH red

http://www.richinstyle.com/test/application/classvalidity.php
21 23 MZ+CH red

http://www.richinstyle.com/test/application/case.php
ALL MZ+CH red

http://www.richinstyle.com/test/application/universal.php
1 3 7 droplist text are red in MZ and black in CH

http://www.richinstyle.com/test/application/combinations.php
1 MZ+CH blue

http://www.richinstyle.com/test/application/firstletter.php
1 2 MZ+CH spacing
8 CH I'm not sure if 8 is correct??
10 CH no red border
20 not red

OK with CH but CRASH MZ, only the points 1 to 4 are visible
if you click on the same link to reload it, crash

(console log available on request : firstletter-MZ1.1a crash.log)


http://www.richinstyle.com/test/application/firstline.php
11 & 12 are different in MZ and CH, in CH 12 is just above 11.
15 MZ+CH red

http://www.richinstyle.com/test/application/sibling2.php
6 MZ+CH green

=== Key concepts
http://www.richinstyle.com/test/keyconcepts/quotes.php
TODO LATER as CH has serif bug

http://www.richinstyle.com/test/keyconcepts/lengths.php
5 MZ+CH I'm not sure if it's ok or not???

http://www.richinstyle.com/test/keyconcepts/implied.php
???? MZ+CH seems empty ????

http://www.richinstyle.com/test/keyconcepts/grouping.php
ALL MZ+CH except 4+5 are not red

http://www.richinstyle.com/test/keyconcepts/important.php
I'm not sure of what he means "not/be different"
123 MZ show a red bg before showing wood bg only at the first load
123 CH does not show the red bg, but immediately the wood bg.

http://www.richinstyle.com/test/keyconcepts/case.php
4 MZ+CH is in Times like.

http://www.richinstyle.com/test/keyconcepts/escapes.php
6 10 MZ+CH not red
7 MZ+CH red

http://www.richinstyle.com/test/keyconcepts/comments.php
3 4 9 14 15 16 MZ+CH not red

http://www.richinstyle.com/test/keyconcepts/comments2.php
4 MZ+CH has a wood bg
8 MZ+CH red
9 10 MZ+CH not red

http://www.richinstyle.com/test/keyconcepts/comments3.php
MZ+CH I'm not sure what he means ???

http://www.richinstyle.com/test/keyconcepts/url.php
8 MZ+CH not white bg

http://www.richinstyle.com/test/keyconcepts/em.php
1 MZ ok = 20px
1 CH different 5em is hugggge
3 MZ+CH it depends of the Font Size (Cmd - or =) sometimes SAME sometimes
different/shifted/shadow-like
5 MZ+CH I'm not sure what he means ????

http://www.richinstyle.com/test/keyconcepts/ex.php
2 MZ+CH no X has the same size as the image


=== Color and backgrounds
http://www.richinstyle.com/test/color/rgb3.php
3 MZ+CH not blue

http://www.richinstyle.com/test/color/rgb6.php
2 MZ+CH not black

http://www.richinstyle.com/test/color/percent.php
5 7 8 MZ+CH not black

http://www.richinstyle.com/test/color/decimal.php
4 MZ+CH not black

http://www.richinstyle.com/test/color/backgroundcolor.php
2 MZ seems to always have white gap between each text line of that paragraph
2 CH has not always white gap between each text line, it depends of the font size.

http://www.richinstyle.com/test/color/backgroundrepeat.php
8 MZ+CH bg not repeated line by line

http://www.richinstyle.com/test/color/forms.php
Lots of differences between MZ and CH (no color support)

=== Fonts
I'll continue the rest later as CH as Serif problem for the moment.
Stephane, this testing is great. It's a bit broad, and many of the issues are
probably already filed as bugs, but anything that isn't can be filed as a new
independent bug blocking this one.

Marking this as meta.
Keywords: meta
Summary: Chimera vs Mozilla CSS Tests → Chimera vs. Mozilla CSS Tests
I'll respond to your observations on a numbered basis.

Attachment of style to pages:

1. Their "absolute" style sheet is an invalid URL; not found.

2. Is '<meta http-equiv="link"' a valid way to link a stylesheet? Needs proof.

3. Their CSS is ".preferred {color: red}", and their HTML is '<p
class="preferred">'. This should be red. Why do they think it shouldn't?

4. Filed as bug 153750.

5. Something's wrong with their test. My own much simpler test works correctly
regardless of case.

6. Filed as bug 153755

7. This does seem to be a genuine Mozilla bug. Filed as bug 153753.

8. I'm not familiar with specificity. Let me get back to you on this one.
Depends on: 153750, 153753, 153755
Since this bug is about making sure Chimera's style system is on par with
Mozilla's, let's put this on the Chimera product.
Assignee: dbaron → saari
Component: Style System → General
Product: Browser → Chimera
QA Contact: ian → winnie
Version: other → unspecified
Also note that, last I checked, a significant number of Matthew Brealey's (the
author of http://www.richinstyle.com/) tests are incorrect.
Application of style to elements:

1. Correct. Filed as bug 153777.

2. For item 21, there's nothing in the spec that says class attribute values
(cdata-list) can't start with a hyphen. Item 23 says it should be red.

3. I believe he's wrong here. CSS doesn't say IDs and classes are
case-sensitive. Bug 109918, comment #11 has some pertinent quotes on this. Maybe
Baron can enlighten us? Is this problem INVALID?

4. Since Chimera uses native widgets, it probably can't apply styles to text in
them. This is probably WONTFIX.

5. He's wrong. Nowhere does he define a CSS statement with the selector,
".class1:link".

6. a. (1) He appears to be correct. letter-spacing should not apply to
first-letter. Filed as bug 153781.
   b. (2) His testcase has whitespace between the first and second letters.
Removing it cures the problem. word-spacing is not applied.
   c. (8) Already filed by bug 60289.
   d. (10) Already filed as bug 103189.
   e. (20) The specification on handling parsing errors
[http://www.w3.org/TR/REC-CSS2/syndata.html#parsing-errors] isn't quite clear on
 this matter.

7. a. (11) His HTML is invalid which is why things don't look right (though they
look similarly wonky to me in Moz and Chimera). When the HTML is corrected, the
problems go away.
   b. (12) He's right. Margin properties should not apply to first-line
pseudo-elements. This has already been reported in bug 83479.
   c. (15) The marker shouldn't be red, and it isn't. The text should be red.

8. I find it hard to believe empty elements are supposed to be ignored. Where's
the specification citation for this? Not to mention, his P isn't empty; it
contains whitespace.
Depends on: 60289, 83479, 103189, 153777, 153781
Actually, this does belong back on Browser.
Assignee: saari → dbaron
Component: General → Style System
Product: Chimera → Browser
QA Contact: winnie → ian
Summary: Chimera vs. Mozilla CSS Tests → Check CSS Compliance against tests at RichInStyle.com
Version: unspecified → other
...and confirming.
Status: UNCONFIRMED → NEW
Ever confirmed: true
http://www.richinstyle.com/test/  (part 2)

Same as part 1 except:
CH : v0.31 BUILD 20020623 16:00 (instead of 0622)

I skip for the moment the chapters FONTS and TEXT waiting for CH to support
Serif, fantasy and cursive. http://bugzilla.mozilla.org/show_bug.cgi?id=153283#c3

=== LISTS
http://www.richinstyle.com/test/lists/displaylistitem.php
MZ+CH none is working, list items are from 1 to 21.

http://www.richinstyle.com/test/lists/liststyletype.php
14 15 21 MZ+CH show me "?." (CH known bug for Unicode?)

http://www.richinstyle.com/test/lists/liststyletypes.php
5 MZ+CH has a dot "5." instead of "5"

http://www.richinstyle.com/test/lists/styling.php
5 MZ+CH not aligned with top of image

http://www.richinstyle.com/test/lists/styling2.php
4 6=black-circle MZ+CH incorrect

http://www.richinstyle.com/test/lists/styling3.php
MZ+CH I'm not sure ??? First 2. is over blue-frame text.

=== BOX MODEL
http://www.richinstyle.com/test/box/margincollapsing.php
2 MZ+CH The 2 Ms not single, shifted/shadowed
MZ are less shifted in several font sizes
CH almost always shifted except in 1 font size

http://www.richinstyle.com/test/box/bordersidecolor.php
All Author comments incorrect, must be Top,B,L,Right.

http://www.richinstyle.com/test/box/bordercolor.php
3 Author CSS error bottom -> blue
7 MZ+CH not black

http://www.richinstyle.com/test/box/hr.php
11 MZ+CH  HR is not blue
17 MZ+CH should it be solid or just the Left/right ends ???

http://www.richinstyle.com/test/box/borderwidths.php
4 MZ+CH not aligned not spec ???

http://www.richinstyle.com/test/box/inlinenonreplaced2.php
MZ+CH if 6 is almost correct, the others 3 4 5 are not then.
Check the rest too???


=== POSITIONING
http://www.richinstyle.com/test/positioning/inlinetext.php
I don't understand what he means ???

http://www.richinstyle.com/test/positioning/images.php
2. MZ+CH is it incorrect ???

http://www.richinstyle.com/test/positioning/dropcaps.php
1 2 MZ+CH not aligned

http://www.richinstyle.com/test/positioning/floatnonfloat.php
2 MZ+CH part of main text is covering top of right-float
3 MZ+CH draw sometimes, change/drag window width, the right border of top-right
float outside the yellow div, +1 pixel.
4 I don't understand what he means ???
5 MZ green/red text not aligned OK !
5 CH green/red text aligned Error ?

http://www.richinstyle.com/test/positioning/height.php
2 MZ+CH is the right-vertical green line correct ???
3 MZ+CH results (visible or not) are different depending of Font Size.
3 Seems better in MZ than CH.
6 MZ OK
6 CH the green is not perfectly aligned on the right (1 blue pixel remaining) !

http://www.richinstyle.com/test/positioning/clear.php
Seems page is incorrect, author error.

http://www.richinstyle.com/test/positioning/positionabsolute.php
3 MZ seems to left the red div with more space on the right, more blue than CH.
4 5 and top blinking blue line ??? I don't understand ???

http://www.richinstyle.com/test/positioning/positionrelative.php
9 MZ+CH T seems to float

http://www.richinstyle.com/test/positioning/positionfixed1.php
Results are very different !!!
MZ Horizontal Scroll bar covered by green div
CH Horizontal Scroll bar over the green div

http://www.richinstyle.com/test/positioning/positionfixed4.php
MZ left image not animated, in CH it's animated.


http://www.richinstyle.com/test/positioning/absolutecontainer.php
MZ+CH I don't understand the author, not sure it's correct ???
3 MZ the horizontal scrolling does not change the text 3.
3 CH strange, if your window is not width enough to have an horizontal scroll
bar, dragging the thumb to right makes a change in text 3, one more line.
To reproduce click on frame-menu to see that page, scroll thumb to the right and
watch text 3 redrawed differently. It happens only once, must click again on
frame-menu link to reproduce each time.

http://www.richinstyle.com/test/positioning/relativecontainer.php
1 2 MZ+CH incorrect, both special lines invisible.


That's all for the moment, the rest later...
Summary: Check CSS Compliance against tests at RichInStyle.com → Check CSS implementation against tests at RichInStyle.com
Stephane, if it's possible, you should probably slow down and do the kind of
analysis I've been doing. A number of his tests are flawed, and so you should
only report real Mozilla or Chimera problems here.

If that's not possible, then you can continue the way you've been doing it. It
will just take longer for us to go through them.
greg, thanks for your kind answer.
As far as I can I try to see if there is a problem in the author code when it does not 
become too difficult for me, I'm not a CSS specialist, neither HTML pushed to such levels.
I'll try to improve my knowledge on both subjects.
Anyway if you have time consuming tests to do, finding an exact path to trigger a naughty 
bug, etc, you can send them to me.
In the mean time, I'll try to find some more beta testers via some of my mailing lists.
Thanks for such a great product.
Keep up the good job !
Key concepts:

1. Chimera/20020621 WorksForMe.

2. Only occurs in quirks mode to support bad IE coders.

3. It is empty.

4. The CSS isn't valid, W3 CSS validator reports errors.

5. You're right, his meaning is unclear. It all looks right to me.

6. I don't actually see any mention in the CSS specification that quoted font
names must be matched case-independently. I think he might be wrong here.

7. (6 & 10) Something's wrong with his pages. When you take the examples and put
it in their own pages, they works fine (or click his "e" for example button).
   (7) He's in error. Comment contents don't allow tokenization (thanks, Baron);
see bug 153996.

8. These all appear to be covered by bug 60290.

9. (4, 8, & 9) Probably more examples of problems covered by bug 60290.

10. Another apparently empty page. Oh, boy.

11. I think he's wrong. It should be yellow.

12. (1) They look the same to me.
    (3) At Text Zoom=100%, it looks as he expects. Mozilla's text zoom does
tricks of its' own to enlarge things, which conflict with his tricks. This isn't
a bug, but bug 131236 addresses the issue.
    (5) They're the same size; that's enough.

13. Appears to be a genuine bug, filed as bug 154007.
Depends on: 60290, 131236, 153996, 154007
*** Bug 153698 has been marked as a duplicate of this bug. ***
Applicaton of style to elements, test #3 is actually correct.  HTML defines
class and ID as case-sensitive.  The page is in quirks mode, however... testing
CSS compliance with a testcase that has no doctype is just asking for trouble.  ;)
Depends on: 154755
Color and backgrounds:

1. It shouldn't be blue. His style rule is "color: 900" and that's not valid CSS.

2. Only happens in Quirks mode.

3. 5. He's wrong. Real numbers are allowed to have more than one digit after the
decimal point in CSS.
   7 & 8. I think he may be right. Filed as bug 154755, but for Strict mode only.

4. I think he may be right. Added as a comment to the aforementioned bug 154755.

5. This draws correctly in both Mozilla and Chimera.

6. It is not clear to me in the spec on what it means to restrict background
image tiling in an inline element that generates more than one line box.
Mozilla's interpretation makes as much sense to me as anything else.

7. Chimera won't support styles in form elements that require widgets. All
others are stylable.
http://www.richinstyle.com/test/frames.php (part 3)

BE SURE to try several font sizes (Cmd-= and Cmd--) before thinking it's ok or
not several results are changing according to font size.

I'm continuing the tests and thanks to Greg Kolanek for his follow up of what I
don't understand. I'll add a paragraph number/letters to ease a little bit your
work.
I've tryed to contact the author of richinstyle to correct his errors, no answer!

MZ : v1.1a BUILD 2002061014
CH : v0.31 BUILD 20020628 16:03 << changed from previous tests
IE : M$IE 5.2
iMac DV 400Mhz - 1024x768 - 32bits
OS X 10.1.5 Build 5S66 French

=== Key concepts (was skipped in previous tests)
A. http://www.richinstyle.com/test/keyconcepts/quotes.php
5. CH+MZ not in New York (font not in OS X) seems to be in Times (Serif) ok ?

6.7.8. CH+MZ not in Courier New, it's in Sans Serif (default font) and not
monospaced.

9.10.11. CH+MZ not in New York (not Times) (font not in OS X) and not a Serif
(it's the default font).


=== Fonts (was skipped in previous tests)
A. http://www.richinstyle.com/test/fonts/generic.php
2. CH+MZ not in Cursive, ok in IE.

5. CH+MZ not in Fantasy, ok in IE.

B. http://www.richinstyle.com/test/fonts/specific.php
1. I don't know what's going here ???
Line is in Zapf Dingbats, the second one in the style ???
style='font-family: WingDings, "Zapf Dingbats"'
Those 2 fonts are in my Mac, works in AppleWorks, not in TextEdit.

C. http://www.richinstyle.com/test/fonts/ui.php
1. to 6. and 10. CH+MZ similar, I don't know how to check this ???
9. IE error, CH+MZ ok

D. http://www.richinstyle.com/test/fonts/keywords.php
All. CH+MZ all are incorrect and results depend of font sizes Cmd-- and Cmd-=.
     CH seems to works almost perfectly in one font size.

E. http://www.richinstyle.com/test/fonts/percentages.php
3. CH+MZ incorrect, 2 lines. ok in IE.

F. http://www.richinstyle.com/test/fonts/fontsizeadjust.php
1. CH+MZ Is this correct ???

G. http://www.richinstyle.com/test/fonts/fontstretch.php
All. CH+MZ no difference, all same condensed

H. http://www.richinstyle.com/test/fonts/fontstyle.php
2. CH+MZ don't understand what he means ???
Times or Times New Roman not available in Oblique in OS X

I. http://www.richinstyle.com/test/fonts/fontweight.php
3. CH+MZ two lines, again depends of font size

J. http://www.richinstyle.com/test/fonts/fontweights.php
1. to 9. CH+MZ difference, more bold starts at 700. IE starts at 500.
12. CH+MZ Only works in one font size due to 'double line bug'

K. http://www.richinstyle.com/test/fonts/font.php
2. 6. CH+MZ not 25px high, ok in IE.
8. 17. 18. 19. 20. CH+MZ the popupmenu has a huge height, not in IE.

L. http://www.richinstyle.com/test/fonts/lineheight.php
All. CH+MZ again huge buttons, not in IE.
5. CH+MZ seems to produce different font style here ???

M. http://www.richinstyle.com/test/fonts/lineheight2.php
2. MZ ok same height. But CH does not!!!
3. CH+MZ do NOT have the same height!!!
4. MZ does not show the THIS only the SHOULD is visible under certain window
width, enlarge/reduce to see it appear/disappear!!
4. CH ok I think??? THIS SHOULD is always visible.
5. MZ ok unreadable.
5. CH it's not unreadable !!

N. http://www.richinstyle.com/test/fonts/linebox.php
2. 3. 5. CH+MZ again two lines depends of font size.
4. CH+MZ in smaller font size, the beginning of sentence is covered by the green
square image. Depending also on font size, the blue text is above and not always
below!
6. 7. CH+MZ results are different and incorrect? depending of font size. IE has
color and seems ok, 6 red bg and 7 black frame border.

O. http://www.richinstyle.com/test/fonts/linebox2.php
1. 2. CH+MZ incorrect! ok in IE.

P. http://www.richinstyle.com/test/fonts/linebox3.php
All. CH+MZ+IE seems ok, same results for 3 browsers.


Q. http://www.richinstyle.com/test/text/forms.php
...later... too huge and WAITING the lineheight BUG to be corrected for CH+MZ,
if it's a bug (huge buttons, see K.8. etc.)

That's all for today.
I'll test the last Q. link and the other chapters, later.
I would recommend ignoring the richinstyle.com tests, they are infamous for
being badly written and often incorrect tests.
*** Bug 160053 has been marked as a duplicate of this bug. ***
Depends on: 230703
Depends on: 230704
Depends on: 230706
OS: MacOS X → All
Hardware: Macintosh → All
Whiteboard: wontfix/invalid per comment 16?
Not Invalid or Wontfix, Anne. This bug tracks the good RIS tests Mozilla fails,
not the flawed ones.
Note that I've gone through every single test on that site and the ones
depending on this bug are all the good ones we fail.
Assignee: dbaron → nobody
QA Contact: ian → style-system
Severity: normal → S3
Summary: Check CSS implementation against tests at RichInStyle.com → [meta] Check CSS implementation against tests at RichInStyle.com
You need to log in before you can comment on or make changes to this bug.