Closed
Bug 162589
Opened 23 years ago
Closed 23 years ago
about:cache-entry should not be sent as a HTTP Referer header.
Categories
(Core :: Networking: Cache, defect, P1)
Core
Networking: Cache
Tracking
()
VERIFIED
FIXED
mozilla1.2alpha
People
(Reporter: darin.moz, Assigned: darin.moz)
Details
Attachments
(1 file, 1 obsolete file)
2.83 KB,
patch
|
bbaetz
:
review+
hjtoi-bugzilla
:
superreview+
|
Details | Diff | Splinter Review |
1) load about:cache
2) click on a http page that is cached
3) notice that you're looking at an about:cache-entry URL that has the original
http or https URL encoded in it.
4) click on the link, and the about:cache-entry URL will be sent as the HTTP
Referer header.
this is a security concern when the cached URL is a HTTPS redirect to a HTTP
site. normally, we don't expose HTTPS URLS to HTTP sites (or other HTTPS sites
for that matter). however, because the HTTPS URL is encoded in a
about:cache-entry URL, our checks to prevent sending a HTTPS URL as a HTTP
Referer would fail.
Comment 1•23 years ago
|
||
So we should just add 'about' to the list in nsHTTPChannel? Maybe we're better
off with a white list (http, https (sometimes), ftp, gopher) instead of the
blacklist....
Comment 2•23 years ago
|
||
I agree that white list is much better than black list.
Assignee | ||
Comment 3•23 years ago
|
||
yeah, a whitelist would make a lot more sense.
ultimately, however, i was thinking that each nsIProtocolHandler should be able
to tell us if it would allow URLs of its type to be sent as a HTTP referrer, but
maybe it really should be the HTTP implementation that decides. hmm. anyways,
a whitelist would be really easy to implement and it'd keep all of the decisions
about HTTP referrers in one place.
Status: NEW → ASSIGNED
Comment 4•23 years ago
|
||
Well, we could use another one of the flag bits, although last time I suggested
adding a flag to a frozen interface even if 'things' without the flag behaved as
before, I got a somewhat mixed reaction... We could try again now that we have a
contcrete example to discuss on npm.xpcom, I suppose.
Assignee | ||
Comment 5•23 years ago
|
||
i'm going to go with the white-list approach for now, and let such a discussion
go on in parallel... i think it's important to have a quick fix to this bug.
Assignee | ||
Updated•23 years ago
|
Severity: normal → critical
Keywords: mozilla1.2
Priority: -- → P1
Target Milestone: --- → mozilla1.2alpha
Assignee | ||
Comment 6•23 years ago
|
||
simple whitelist for http referrer: only allows http, https, and ftp. https
still follows the same referrer rules. tested against about:cache-entry, and
verified that it is no longer sent as a referrer.
Assignee | ||
Comment 7•23 years ago
|
||
Attachment #96781 -
Attachment is obsolete: true
Comment 8•23 years ago
|
||
Comment on attachment 96782 [details] [diff] [review]
v1.1 patch - forgot about gopher
gopher forever!
r=bbaetz
Attachment #96782 -
Flags: review+
Comment on attachment 96782 [details] [diff] [review]
v1.1 patch - forgot about gopher
sr=heikki
Attachment #96782 -
Flags: superreview+
Assignee | ||
Comment 10•23 years ago
|
||
fixed-on-trunk
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
Comment 11•23 years ago
|
||
verified trunk 09/05/02 - about:cache is not being sent as referer - winNT,
linux, mac
Status: RESOLVED → VERIFIED
Comment 12•22 years ago
|
||
adding benc@netscape.com to CC list
Updated•22 years ago
|
Group: security
You need to log in
before you can comment on or make changes to this bug.
Description
•