506639 - cookies not saved in XML documents (e.g. in SVG)

Status: RESOLVED INVALID

(Core :: Networking: Cookies, defect)

Description

[jonathan chetwynd]

Attachment: svg-cookie broken

the code used to save a cookie works fine in html on mozilla but not svg.

open svg attachment, click on word failed, refresh browser

should read 'success'

parity opera and safari.

Comment 1

[jonathan chetwynd]

Attachment: html-cookie success

Comment 2

[jonathan chetwynd]

moving from component SVG to networking: cookies.

not sure why mozilla considers cookies as 'html'

anyone care to comment

Comment 3

[Robert Longson [:longsonr]]

Probably invalid.

https://developer.mozilla.org/en/Writing_JavaScript_for_XHTML#Problem.3a_My_Cookie_Won%27t_Be_Saved!

Comment 4

[Jonathan Watt [:jwatt]]

Bugzilla should handle this link better:

https://developer.mozilla.org/en/Writing_JavaScript_for_XHTML#Problem.3a_My_Cookie_Won't_Be_Saved%21

IMO we should implement the same for XMLDocument to remove yet another gotcha for authors using the Web platform.

Comment 5

[jonathan chetwynd]

Attachment: svg-cookie or localStorage workaround

Comment 6

[jonathan chetwynd]

Workaround: trying so far without luck to find one ie:
else if(localStorage){}
however Opera appears to balk at my attempts
and because mozilla does save a cookie, it's hard for me to imagine apart from a straight UA check... ideas gratefully appreciated

re: #4 agree, 
have filed blocking bug 506655
as mentioned mozilla does save a cookie, 
then destroys it as per this bug...

seriously getting this web app live is becoming a real trial...
~:"

Comment 7

[jonathan chetwynd]

Attachment: workaround

Comment 8

[jonathan chetwynd]

a possibly related cookie in xml bug 506692

Comment 9

[zebu.damach]

Good day,

I think I have a similar problem, therefor I did not open a new bug report.

I use Firefox 3.5.1 running under openSUSE Linux 10.3. Firefox loses its cookies from time to time. I searched the web to find a solution and came up with the following:

- delete or rename the file cookies.sqlite since it might be corrupt. I did so, but the cookies were lost again after some restarts of the browser.
- make a whole new profile to start with. I did without changing the default cookie settings. Then I saved some cookies, but after the very next browser restart, they were gone.

I also tried some cookie management extensions that did not help in any way. So I think this might be a bug.

kernel 2.6.22.19-0.3
glibc 2.6.1
KDE 3.5.7

Comment 10

[jinxmchue]

I have this problem, too, and it just popped up all of a sudden.  Cookies were working fine until a couple days ago.  I have not changed any settings for cookies and have only updated the NoScript add-on.  I don't know if that could be the problem.

Comment 11

[Robert Longson [:longsonr]]

(In reply to comment #9)
> Good day,
> 
> I think I have a similar problem, therefor I did not open a new bug report.
> 
> I use Firefox 3.5.1 running under openSUSE Linux 10.3. Firefox loses its
> cookies from time to time. I searched the web to find a solution and came up
> with the following:
> 

This bug is not about losing cookies sometimes. It is about cookies not saving ever on xml (xhtml, svg etc) pages.

Comment 12

[Nickolay_Ponomarev]

I tried to update the summary not to be so confusing, but since this bug doesn't make much sense to a casual reader, feel free to correct.

Comment 13

[Chris Lawson (gone)]

This definitely looks INVALID to me. Comment 6 et seq should be taken to a support forum if a workaround hasn't been found by this point.

Comment 14

[jonathan chetwynd]

Chris,

did you read the Description?
and try the two testcases?
and check Opera or Safari?
and review the  source in the  two testcases?

onmousedown="setCookie('user_name','chad')"  does not work in SVG

body onLoad="checkCookie()"  works in html.

Please do not close bugs you do not understand

Comment 15

[Chris Lawson (gone)]

Jonathan, just because you disagree with what I did doesn't make comment 3 any less true.

Comment 16

[jonathan chetwynd]

#15

well as per another recent bug,
either this should be taken to the working group,
or bugs need to be filed with other UAs

else the web will be 'badly' broken,
ie authors will use this as it works in Opera and Safari,
as stated.

Comment 17

[Chris Lawson (gone)]

As per *what* "other recent bug"?

Comment 18

[jonathan chetwynd]

#597974


--- Comment #15 from Boris Zbarsky (:bz) <[e-mail]> 2010-09-20 12:01:20 PDT ---

Filed opera bug (for which the followup mail is [e-mail]) and
a webkit bug at https://bugs.webkit.org/show_bug.cgi?id=46112

Comment 19

[Chris Lawson (gone)]

What the heck does bug 597974 have to do with this one?