After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 130072 - Option to remove cookies after session
Option to remove cookies after session
Status: RESOLVED WONTFIX
Product: epiphany
Classification: Core
Component: Backend
git master
Other All
: Normal enhancement
: Ephy/XR
Assigned To: Epiphany Maintainers
Epiphany Maintainers
: 145755 (view as bug list)
Depends on:
Blocks: 755292
 
 
Reported: 2003-12-27 12:37 UTC by Christian Persch
Modified: 2015-09-29 04:18 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
proposed fix (608 bytes, patch)
2003-12-27 14:58 UTC, Christian Persch
none Details | Review

Description Christian Persch 2003-12-27 12:37:43 UTC
Many cookies have ridiculous lifetimes, like somewhere in 2038 :)

We should pick a sensible upper lifetime (in days), and set
network.cookie.lifetime.days and
network.cookie.lifetime.enabled accordingly.
Comment 1 Christian Persch 2003-12-27 14:58:54 UTC
Created attachment 22734 [details] [review]
proposed fix
Comment 2 Marco Pesenti Gritti 2003-12-27 15:19:01 UTC
Could you explain the reason of having a limited lifetime ?
Does the lifetime "countdown" start on the last time the cookie is 
accessed or the time it's created ?
Comment 3 Christian Persch 2003-12-27 15:53:10 UTC
Without lifetime limit, the cookie expires when the site has said to,
and many sites have unacceptable values there, like 2038 [i.e. the
absolute maximum unix time]. Therefore, when you visit some site which
sets a cookie, it lives forever, long after you've last accessed the
site (you may even have been there only once), and long after all
other traces of you having been there, like the history entries, have
expired.

The "countdown" starts when the cookie is set or modified; since most
sites update the cookie when you visit them again, the "countdown"
starts anew then.
Comment 4 Marco Pesenti Gritti 2003-12-27 16:02:05 UTC
Sounds good to me.
Comment 5 Marco Pesenti Gritti 2004-01-26 18:41:43 UTC
what's up with this ?
Comment 6 Christian Persch 2004-01-26 19:27:15 UTC
You ok'd to commit it, but spark raised an objection: what about sites
which don't update their login cookies (which would lead to you having
to re-login). [Don't know if those sites exist :)]
Comment 7 spark 2004-01-27 14:11:01 UTC
I think we should try this early on after we branch for 1.2 so we can
see if it causes any problems.
Comment 8 Luis Villa 2004-02-26 22:30:31 UTC
Couple things:
* please don't put this in as is; the default can't be 'eat my data
after a month' or you'll get a lot of 'I entered my password at this
site, and for some reason it isn't remembered when I go back.' Make it
an option, sure, but have it default to false.
* I'd suggest making this something like 'eat all personal data after
30 unused days' or something like that- tie together cookies,
passwords, history, etc. Just a thought.
Comment 9 Christian Persch 2004-02-27 15:18:15 UTC
The password is remembered in the password manager; if the web site
sets a cookie this just makes the next login a bit faster.

We already delete history info after 10 days, so I don't see why other
personal data like cookies shouldn't be forgotten after a certain
time, too.
Comment 10 Luis Villa 2004-02-27 15:30:28 UTC
I'm not particularly sure that deleting the history after ten days is
sane either. It certainly bugs me when I'm trying to look for older
sites. IMNSHO, one should never delete user data by default. Like I
said, I think tying them into one preference ('delete personal data
after X days') with a default to keep forever is the right one. 
Comment 11 Marco Pesenti Gritti 2004-02-27 19:21:26 UTC
Luis, never deleting history would cause serious performances 
issues, loading pages, using history dialog and using completion.
I dont think we have the same problem with cookies though.
Comment 12 Marco Pesenti Gritti 2004-02-28 00:36:36 UTC
My opinion on the cookie lifetime is that it's not worth risking to
introduce problems. What we really gain by expiring these cookies ?
They are not visible to most users anyway ...
Comment 13 Christian Persch 2004-10-13 10:51:20 UTC
Mass reassigning of Epiphany bugs to epiphany-maint@b.g.o
Comment 14 Keywan Najafi Tonekaboni 2004-10-16 09:25:41 UTC
an option to forget cookies after the session would be nice. session means
deleting all cookies, after the last epiphany window is closed.
this is in my opinion the best compromise between trouble-free visiting website
and privacy. 

Of course this shouldnt be an default option.

I am an administrator for an internet cafe and we dont use epiphany, besause it
isnt possible to save cookies just temporally.
Comment 15 spark 2005-01-03 22:54:16 UTC
Modifying summary to reflect refocus of this bug.
Comment 16 Chris Rebert 2005-06-13 23:01:03 UTC
This feature would be really handy from a privacy standpoint. Firefox already
has this feature, why not epiphany? I hope this is implemented soon.
Comment 17 Christian Persch 2005-07-30 21:17:16 UTC
Target: 1.8 -> 1.10 due to feature and UI freeze.
Comment 18 Keywan Najafi Tonekaboni 2006-03-15 09:03:30 UTC
I want ask what the status of this cookie problem is. I think this should have a higher priority and isn't "just" an enhancement. Privacy issues are really important and when I check from time to time my saved cookies I am shocked which garbage their is.

I would be really happy, if this could be realized in the next version (2.16). If this feature would be avaible, epiphany would be perfect for me. I like this browser really.
Comment 19 Christian Persch 2006-08-17 14:04:26 UTC
Mass changing target 2.16 -> 2.18
Comment 20 Chris Rebert 2007-11-18 01:00:42 UTC
Is there any progress on this? The last comment is from a year ago and this ticket is *very* important from a privacy standpoint. Since epiphany can be built against Firefox, isn't there a way to access the Firefox backend for this feature and build off that?
Comment 21 Chris Rebert 2007-11-18 01:06:37 UTC
Ah, it turns out there's a setting for this:  http://kb.mozillazine.org/Network.cookie.lifetimePolicy

Just set it to '2' in about:config to have all cookies be session cookies.

So, all that's really needed is a simple GUI to change the value of the setting.
Comment 22 Michael Monreal 2008-04-01 23:12:46 UTC
+1 for this feature from me. Cookies can be dangerous. A browser should not force a user to keep cookies longer than really required.

The workaround mentioned in comment #21 is only valid for the gecko backend, which is bad. If WebKit is going to be used more in the future, this is one of the most critical features IMHO.

Bug #145755 is a dupe of this btw.
Comment 23 Reinout van Schouwen 2008-04-01 23:15:44 UTC
*** Bug 145755 has been marked as a duplicate of this bug. ***
Comment 24 Diogo Campos 2015-09-29 03:05:43 UTC
Fixed by the Incognito Mode?
Or still a wanted option to the Normal Mode?
Comment 25 Michael Catanzaro 2015-09-29 03:11:04 UTC
Yup, this is what Incognito Mode is for. Well, you could make a case for wanting to save history but not cookies... I don't think we really need that level of configurability, so I will close this bug, but if someone wants to submit a patch, I think that'd be fine.