CSS Test Suite Results: MS Internet Explorer For Windows 5.0b2
Last Run Through Entire Test Suite (beta 2): 1998 December 17.
Last Run Through Existing Bugs (beta 2): 1998 December 17.
Previous version of this page: bugs in beta 1.
Bugs
The list that follows is a list of bugs in the browser, that is, what is
described is in error. I have also included a few partially supported
features as bugs.
Microsoft has an email address for standards-compliance bugs:
wasp@microsoft.com. They seem responsive, at least in that they
respond to email.
- 1.1
- Miscellaneous import problems: fails to import file in quotes rather
than with
url(), but does import something that is after other
CSS.
- 1.4,
1.5
- Class/ID selectors should not be allowed to begin with digits.
The browser displays incorrect behavior, as explained in the link to the
test. It should ignore the properties, since selectors should not begin with
digits. If it accepts them, then authors using MSIE would not realize that
what they did was wrong and not supported by other browsers.
- 1.7
- MSIE Has problems handling comments with multiple stars, namely
/**/ and /* a comment **/.
- 3.2
- Line has underline and overstrike.
text-decoration should span children, but it is completely
overridden by declarations higher in the cascade on the same
property.
- 4.1.1
- Nested margins are collapsing even when separated by padding, it seems.
- 4.1.2
-
width is interpreted incorrectly to be inclusive of
padding and border. There is also extra top and bottom margin built-in
on the P element.
- 4.1.2,
4.3
- Setting margin values to
auto does not work.
- 4.1.4
- The paragraphs in the first
DIV are not high enough. Their
tops should be even with the top of the line of text that contains them.
- The misinterpretation of
width causes the third test to fail.
- The text is moved down to low in the fourth and fifth tests. It should begin at the top.
- In the fifth test, the second left-floating paragraph is mysteriously too low.
- In the sixth test, the floating paragraph should be even with the text that precedes it. (I don't like this part of the spec, but it's true.)
- The seventh test is a royal mess. The negative top margin doesn't work
properly, the bottom border of the
DIV is missing, and the left border
edge of the div is uneven.
- In the eighth test, the
DIV should not be expanded to hold
its contents. They should overflow.
- The first five
DIVs in the TABLE are too wide.
- 4.2,
5.5.01b,
5.5.02b,
5.5.03b,
5.5.04b,
5.5.05b,
5.5.06b,
5.5.07b,
5.5.08b,
5.5.09b,
5.5.10b,
5.5.11b,
5.5.12b,
5.5.13b,
5.5.14b,
5.5.15b,
5.5.16b,
5.5.17b,
5.5.18b,
5.5.19b,
5.5.20b,
5.5.21b,
5.5.22b
- Box properties are not supported at all on inline elements.
- 4.4
- Padding is not supported on the images here.
- The negative margins do not have any effect vertically.
- 4.5
- The background-color on HTML is ignored.
- 5.2.3
- No oblique fonts -- can this be done in Windows??
- 5.2.4
- The small-caps does all-caps instead.
- 5.2.5
- Font weight 500 should probably be rendered as bold, as per the example
in the CSS2 spec.
- 5.2.6
medium is not defined as the default font size, as it should be.ex units are too big.
- 5.2.7
- The
line-height percentages are interpreted relative to
normal (ca. 120%) rather than to the font-size? I'm not sure exactly how
line-height should be calculated.
- The font size does not inherit into the table.
- 5.3.4,
5.3.6,
5.3.7
- Repeat-x and repeat-y do not tile in both directions.
- 5.4.3
- Underlining does not stay the same color throughout the paragraph. This is because of the next problem.
text-decoration should span children. (It doesn't inherit.)
- 5.4.4
- Not all
aligns work. In fact, the only ones that work are
sub and super work on text. More seem to work on
images, such as top and middle.
- The image seems to be positioned oddly in some of the later tests.
- The
text-top and middle aligns should be relative to the text of the parent element.
- 5.4.5
- Inline elements in the middle of words cause extra capitalization.
- 5.4.8
- Percentage values are relative to the normal amount rather than to the
font-size, I think.
- 5.5.01,
5.5.03,
5.5.05
- top-margin and bottom-margin in percentage units behave oddly (and
differently in and out of
TABLE - percent value may be
referring to height instead of width when inside of
TABLE. The behaviour inside the TABLE causes
serious display problems. Behaviour is also strange outside of
TABLE. (Note: Things work fine for padding.)
- The margin-top does not work on the list item.
- 5.5.17
border-style: dashed and border-style: dotted do not work.
- 5.5.18,
5.5.19,
5.5.20,
5.5.21,
5.5.22
- Double border set as
10px or medium has uneven halves.
- 5.5.25,
5.5.26
-
The text around the floating images is moved down too far (except for the first test in the table). It shouldn't be moved down at all.
- 5.5.25b
- The bottom edge of the background fill (white) is set by the floated image
rather than by the paragraph (throughout the test), even when the paragraph
is not yet done. When the paragraph is not yet done, the background ends
and then restarts with the next line of text.
- The negative
top-margin and bottom-margin fail, turning to zero.
- 5.5.25c
- In the sixth test, the background of the blue element should extend
behind the float.
- In the seventh test, the DIV should not be made larger because of the
float it contains.
-
- 5.5.25d
- The widths are incorrect due to MSIE's misinterpretation of
width.
- 5.5.26
- The text is pushed down too far next to floats. The top of the text
should be even with the top of the float when the text flows around the
float.
- 5.5.26b
clear does not work on floats- The shape of the
DIV.article should not be changed because
of the floats. It should be rectangular is shape and the margin should be
relative to the body, not the floats.
- 5.5.26c
- This test is a mess. The following errors stand out, although there are
a number of others:
- When box properties are overconstrained, the right-margin should be set
to
auto. This applies to the BODY.
width is interpreted incorrectly.- A box with floating children should not be enlarged to hold those children.
- 5.6.1
- The
display: list-item does nothing.
- 5.6.5
- The horizontal alignment of the list markers is wrong. See the
CSS2 specs. Since the margins should be the same, the left edge of the
text in the first test (excluding the bullet) should line up with the
left edge of the text in the second test (including the
bullet).
- 5.6.6
- The
list-style: inside only works partially for upper-alpha.
Judging from the previous test and this one, list-style: inside
is not being done by placing the marker in the text flow of the list-item, but
rather by undoing the displacement of the text and indenting the first line.
This is incorrect.
- 7.1
- Adjacent sibling selectors ("+") should either make the whole selector
be ignored, since MSIE doesn't understand them, or should be interpreted
correctly.
- Keywords cannot be quoted and should not be accepted when they are.
- Unknown units should not be interpreted as
px. For that
matter, neither should missing units, but that isn't part of the test.
- Unknown at-rules should have their contents ignored.
- Invalid keywords in a property such as
font or
text-decoration should cause the whole statement to be
ignored.
- Invalid characters (such as "!") within selectors should cause the whole
declaration block to be ignored.
// should not be treated as a comment.- CSS Strings are not correctly parsed.
\" should not
terminate a string, since it is escaped. The same is true for
\' in a string delimited by single quotes. Note: to test the rest of this test, I copied the file and removed the lines for P.twentyb and P.twentyoneb.
P.twentytwo is an error in the test suite. Ignore it. Note: to test the rest of this test, I copied the file and removed the lines for P.twentytwo.
Unsupported Features
- Alternate Stylesheets (1.1)
- :first-line
(2.3)
- :first-letter
(2.4)
- word-spacing:
(5.4.1)
- text-decoration: blink;
(5.4.3)
- white-space:
(5.6.2)
(Back to
CSS Testing Information,
David Baron)
LDB,
dbaron@dbaron.org