CSS1 Test Suite Results: NGLayout Engine

Status of This List

Previous Versions of this page

Bugs from Test Suite

The list that follows is a list of bugs in the browser, that is, what is described is in error. I sometimes only mention the bug the first time it occurs in the test suite. I have also included a few partially supported features as bugs because, although it would be acceptable to completely fail to support the feature, supporting it halfway can cause unexpected results.

The underlining is ugly. It does not have a stable baseline (though it does have solid color), which it should, since text-decoration is not inherited, but, rather, spans children. (See Bug 1019)
4.2, 5.5.11b, 5.5.12b, 5.5.13b, 5.5.14b, 5.5.15b, 5.5.16b, 5.5.17b, etc.
(See Bug 1606) - ADD TO THIS BUG
The resizing seems to be turning a square into a rectangle.
This page is a real mess. Cursive, fantasy, and monospace don't do what they should in the first part, but monospace somehow becomes better when there is an alternative font-family after it. Fonts in the DIV seem to be changing the size of fonts within - in the second DIV, with the monospace font. There may be some other weird things too. This is explained in bug 582. (See Bug 1038) (See Bug 581) (See Bug 582)
Judging from MSIE5, you could do a little better with the weight 900 (which also applies to the H4 with bolder. (See Bug 972)
Negative values should not be allowed. (See Bug 1039)
5.2.7, 5.4.8
Most line-heights seem to be a little big. I think they are being done relative to the default rather than to 1. However, it's hard to tell. (See Bug 585)
The BODY font size does not inherit correctly into the TABLE.
text-decoration should span children with text-decoration set to none.
Line heights in ems, percentages, and scaling factors look like they might be based on the original line-height rather than the font-size. I'm not sure.
5.5.01b, 5.5.03b, 5.5.05b
Top and bottom margins on inline elements should not do anything, because they do not increase the line spacing.
When the page is narrowed, the TABLE won't shrink past a certain point. This may be OK.
5.5.06b, 5.5.08b, 5.5.10b
Top and bottom padding on inline elements should not increase the line spacing.
The top border is 1px (?) thinner than the rest. This is especially visible on thin. (See Bug 591) (See Bug 1052)
Cosmetic bug: some of the borders (dotted, dashed especially) could use some work. (See Bug 1054)
5.5.18, 5.5.20
(Old version of suite: The thin (top & bottom) border is thinner inside the table.) Now it is gone. (This applies to tests 19 and 21 and 22 as well.) (See Bug 1055)
The vertical alignment of the floats is correct, but the paragraphs flowing around them are pushed down too far. I have a modified version of this page where it is incorrect, along with other bugs. See the CSS2 Specs.
The image being moved up and left 30 pixels is being moved right 60 pixels instead of left 30 pixels. Also, the image with a positive bottom margin is causing weird effects -- this should be fixed if the margins on inline elements bug is fixed. (See Bug 1605) (See Bug 1606)
There is a horizontal line running across the image that shouldn't be there.
The text should be higher... it should not be pushed down by the existence of the floats.
The yellow rectangle is moved down by the clear on the previous float. Elements in the normal flow should not be affected.
This test is a mess. I'm not going to describe all the problems in detail. The font isn't inheriting into the table, and some of the floats are moved down too far (although we can argue over spec wording versus intention here).
list-item doesn't work anymore. Could this have something to do with bullet placement? This should start working when inside works, because then the bullet will be in the paragraph (this isn't quite true, and actually it shouldn't be).
list-style-position: is set in ua.css on an LI, so the test doesn't work. (See Bug 1049)

Unsupported Features from Test Suite

Other Tests

Valid HTML 4.0!

(Back to CSS Testing Information, David Baron)

LDB, dbaron@dbaron.org