Re: [dev] regarding surf and cookie handling

From: Calvin Morrison <>
Date: Tue, 21 Feb 2012 16:15:55 -0500

On 21 February 2012 15:13, Nick <> wrote:
> Quoth Peter Hartman:
>> The other reason (besides hatred of libsoup) that surf never bothered
>> implementing anything other than what it currently has is that (to my
>> knowledge) there has yet to be a bona fide web-site that the current
>> system fails on.  There's a lot of FUD, but none (to my knowledge)
>> that is bona fide.
> Indeed. I remember chatting about possible fail states of the
> current cookie mechanism, but it all got a bit boring and so I
> forgot the outcome. I think basically it boiled down to our system
> could fail potentially if sites using multiple windows expected to
> interact in particularly stupid ways using cookies.
> But even if our system works my vote would be to go the libsoup way.
> We shouldn't pretend that surf is anything but a shim atop a
> monster, and using their cookie management means it's less likely to
> fail in unexpected ways in the future.
>> and if Nick approves the patch, I don't see there being
>> anyone out there who would dis-approve it.
> The power that comes of being vocal in a userbase of < 5 ;)

Okay so it looks my implementation was a bit half-assed. I didn't
realize that SoupCookieJarTxt would only load at startup, and not
reread the cookies into the jar, only write them. Luckily there is a
changed singal [1] that is emitted. Using some libsoup fo we can
intercept the cookie(s) and add/delete them to our current CookieJar.

This is stil is favorable than the current approach as far as I can tell.

Also Integrating further into libsoup could prove productive. We use
the SoupCookieJar cookie policy, which would allow the user to accept
all, deny all or deny 3rd party cookies automagically [2]

Just looking into libsoup a little bit now and spouting off some ideas.

Calvin Morrison

Received on Tue Feb 21 2012 - 22:15:55 CET

This archive was generated by hypermail 2.3.0 : Tue Feb 21 2012 - 22:24:05 CET