25230 - [TEXT] Square symbol, weird character translation?

Status: VERIFIED INVALID

(Core :: DOM: HTML Parser, defect, P3)

Description

[chris]

The  symbol is appearing on the search results page of this restaurant
search engine when browsed by Mozilla. The behavior does not occur in any
other browser.  Above is an example search that reveals this problem.

Comment 1

[laurel]

*** Bug 25286 has been marked as a duplicate of this bug. ***

Comment 2

[leger]

Assigning all open "[e-mail]" bugs to "[e-mail]" to weed 
thru.

Comment 3

[Andrew Brobston]

I see this also in Win98 build 2000-02-17-15, but not in Netscape 4.7. Marking 
platform as "All" and keyword 4xp.

Viewing the page source in Mozilla yields many of the same square symbols, which 
do not appear in Netscape 4.7's view source. As a result, I do not think this is 
an internationalization bug. For lack of anywhere better to send this, I'll send 
it to Parser.

Comment 4

[Richard Zach]

I don't see this on Linux build 2000.02.21.08.  Do you still see it?

Comment 5

[Andrew Brobston]

Yes, on Win98 2000-02-22-08. Attaching screenshot shortly.

Comment 6

[Andrew Brobston]

Attachment: Screenshot of box character bug

Comment 7

[chris]

If this helps, when opening a file in OS 9, and typing in the file name, I typed 
the DEL key and got this symbol..maybe that will be of help to someone..

CJF

Comment 8

[leger]

Updating QA contact info, assigning to parser eng.

Comment 9

[rickg]

Petersen: Can you verify this on mac for me? 
Leger: parsing has nothing to do with font rendering on the mac.
(But I'll help track this down).

Comment 10

[Chris Petersen]

Yes, this bug is occuring on the Mac March 20th build (2000032006).

Comment 11

[rickg]

Harish -- I think you recently addressed a similar issue?

Comment 12

[harishd]

Looks like we are not handling vertical-tab ('\v').  Could be layout related.
CCing troy and nisheeth.

Comment 13

[harishd]

Attachment: Reduced case

Comment 14

[harishd]

Looks like buster's. Reassigning.

Comment 15

[Erik van der Poel]

The vertical tab character (0x0B) is being displayed as a box. Not sure how to
solve this. Nav4 and IE seem to display a space, instead of a box. Perhaps we
need to convert 0B to a space, but maybe not in the content model, since we may
want to write the file back out as it was (in the editor). Perhaps we need to
map 0B to space down in the font engine. There are a number of other real
Unicode space characters that also need such treatment, so perhaps we can fix
all of these the same way.

Comment 16

[troy]

The text transformer already checks for "discarded characters" and it just 
ignores them, i.e., they're skipped and not added to the transformed text

You could do that with the vertical tab character as well, assuming you don't 
want it displayed

Comment 17

[buster]

Here is the initial batch of layout [TEXT] bugs.  I'll search for more today, 
but this should keep you busy for a while!

Please review these for possible dogfood or nsbeta2 candidates, and assign your 
own priority and milestone for each.  The current settings were relative to my 
bug list, and they might not make sense any more.

Comment 18

[brian]

on linux, urls and much text in dialogs shows up as squares

Comment 19

[Erik van der Poel]

Vertical tab is an old, obscure control character that is supposed to have a
vertical spacing effect. It shouldn't be in a Web page. Instead of modifying
Mozilla, it might be better to fix the Web page(s). Marking INVALID.

Comment 20

[Jan Carpenter]

verified