53441 - Doesn't reproduce the correct ASCII character in <ALT> text while an image is loading

Status: RESOLVED WORKSFORME

(Core :: Layout, defect, P3)

Description

[Fede]

I don't mean the text in "labels" over the images, but an ASCII character 
(’) inside an image, while the image itself is loading. I see, instead, a 
square on Win 95 and a "?" on Win NT.
I'm browsing with BUILD ID: 2000091508 M18 nightly
See the big blue image (bit TOUR’s) in the page

Comment 1

[Asa Dotzler [:asa]]

over to layout.

Comment 2

[Nisheeth Ranjan]

Re-assigning 5 bugs from Clayton's list to Heikki for further triage...

Comment 3

[Heikki Toivonen (remove -bugzilla when emailing directly)]

NN 4.7 gets this right (on Linux, even), adding keywords. I was unable to test 
Mozilla on Linux at this time, but I would guess this bug is cross platform.

I think the fix should go into nsImageFrame::DisplayAltText(). I am guessing 
this is a job for [e-mail].

Comment 4

[rods (gone)]

Attachment: small testcase

Comment 5

[rods (gone)]

This looks like a parser bug to me. I dumped the content and it has some 
unprintable char.

Comment 6

[harishd]

The problem could probably be in nsString::Append code. Adding waterson to the 
CC list for input.

Comment 7

[harishd]

Probably string related. Reassigning to waterson & ccing myself.

Comment 8

[Hixie (not reading bugmail)]

Upon managerial request, adding the "testcase" keyword to 84 open layout bugs that
do not have the "testcase" keyword and yet have an attachement with the word
"test" in the description field. Apologies for any mistakes.

Comment 9

[Hixie (not reading bugmail)]

Seems like a valid HTML bug; we are indeed handling the entity differently if it
is in the attribute as opposed to in the markup proper. Taking QA.

Comment 10

[Christopher Hoess (gone)]

Read the thread spawned from the following post:

http://groups.google.com/groups?hl=en&selm=9s453v%24a13%241%40sun-cc204.lut.ac.uk

’ is some control code. The entity you want to use is ’ I suppose.

(see especially in the thread the bit about Google indexing the three-digit
control chars as such, and not as printed characters).

Should probably be invalid.

Comment 11

[Eli Friedman]

agreed

Comment 12

[Hixie (not reading bugmail)]

Woah, woah.

This is valid, because we should do the same thing in attributes as in normal
markup. We don't on Windows, last I checked (we do on Linux now, though, so
maybe that has been fixed).

Comment 13

[Eli Friedman]

Yeah, it's fixed on Windows too. (2002071408 WinME)

Should have paid more attention to what the problem was.

Comment 14

[Gérard Talbot]

References
----------

http://www.wikipedia.org/wiki/ISO_8859-1

http://www.alanwood.net/unicode/latin_1_supplement.html

http://www.unicode.org/charts/PDF/U0080.pdf

Unicode and other references indicates that code position 146 (x92) is for
Private Use 2 (PU2) according to references. 
Windows-1252 displays a right single quotation mark.

Comment 15

[Gérard Talbot]

Attachment: Character entities within the alt attribute of images

I think this bug is invalid.

Comment 16

[Christopher Hoess (gone)]

Yeah, seems to me working normally. WFM.