108173 - extra blank line appears

Status: RESOLVED INVALID

(Core :: Layout, defect)

Description

[Henrik Gemal]

I'me attaching a testcase where a extra blank line appears.
Mozilla shows the testcase as:
----------------
test 1
test 2

test 3

test 1
test 2
test 3

while IE shows the testcase correctly as:
----------------
test 1
test 2
test 3

test 1
test 2
test 3

tested on build 20011101

Comment 1

[Henrik Gemal]

Attachment: testcase

Comment 2

[Christopher Hoess (gone)]

Don't know what's going on here, will construct a valid testcase shortly. My
recent CVS build shows the extra line under test 1.

Comment 3

[RC]

Mozilla inserts a blank line.
Opera inserts a blank line.
IE inserts a blank line too, if you move the </p> tag up:

<p>
test 1<br></p><!-- close paragraph tag moved here --><div style="display:none">
- test 1a<br>
</div>

test 2<br><div style="display:none">
- test 2a<br>
</div>
test 3<br><div style="display:none">
- test 3a<br>
</div>

Comment 4

[Henrik Gemal]

I know it has something to do with the <p>
not sure what but it's weird....

Comment 5

[RC]

Reduced testcase, with borders so you can see what's happening:

<p style="border: thin dashed">
test 1<div style="border: thin dashed"></div>
test 2<br>
test 3<br>

Comment 6

[RC]

Attachment: Screenshot comparing IE/OP/NS rendering of reduced testcase

Comment 7

[Christopher Hoess (gone)]

I think the cause of this line is that the <div> implicitly terminates the <p>.
Rendering of the testcase immediately above (with dotted borders) is the same if
I insert </p> just before the <div>; the thickish-looking dotted line is, of
course, the empty <div>. Are we actually doing anything wrong here?

Comment 8

[Mats Palmgren (inactive)]

Attachment: Testcase #2

Comment 9

[Mats Palmgren (inactive)]

I think the bug is in IE. In attachment 63437 [details] there are 3 examples. The first
is the reduced testcase of the problem: <P>text<DIV></DIV>
It could be parsed as (example 2) <P>text</P><DIV></DIV> 
or as (example 3) <P>text<DIV></DIV></P>
Mozilla and Opera does 2. IE seems to parse it as 2 but renders it with a quirk
of nuking the margin that normally is there when a proper end-tag is present.
I suggest resolving this as INVALID - I think we are doing the right thing.

Comment 10

[Mats Palmgren (inactive)]

Attachment: Testcase #3

Comment 11

[Mats Palmgren (inactive)]

I think we are doing the right thing here. (2002-06-20-08 trunk win98)
-> INVALID.