Very true, 4KB is a large amount of data to send on each request.
Until something like FF storage stuff is more widely implemented then
I think devs have little options.

I was thinking of putting some kind of calculation on how much space
is being used, and some way to inform the developer on reaching the
limit, but I don't know :-/

On Aug 13, 2:02 pm, "Dan G. Switzer, II" <[EMAIL PROTECTED]>
wrote:
> LetsSurf,
>
> >I've just finished writing a cookieJar plugin for jQuery, as I
> >couldn't find anything like it for jQuery.
>
> >http://www.jdempster.com/category/code/jquery/cookiejar/
>
> >Thought I would post it here in the hope of some feedback.
> >Please let me know what you think. Is this a good approach to the
> >problem. If you found any problems or have any suggestions.
>
> Looks interesting, but one suggestion I would make is to put a reminder
> prominently displayed on the page that reminds developers that:
>
> 1) Cookies are sent with each request header.
> 2) The total Cookie size per domain is generally 4K (which is per the 
> spechttp://www.ietf.org/rfc/rfc2109.txt.) This is not a per Cookie limit, but 
> a
> total name-value pair.
>
> Here's what Microsoft says about IE:
>
> http://support.microsoft.com/kb/306070
>
> "Each cookie begins with a name-value pair. This pair is followed by zero or
> by more attribute-value pairs that are separated by semicolons. For one
> domain name, each cookie is limited to 4,096 bytes. This total can exist as
> one name-value pair of 4 kilobytes (KB) or as up to 20 name-value pairs that
> total 4 KB. If the computer does not have sufficient space to store the
> cookie, the cookie is discarded. It is not truncated. Applications should
> use as few cookies as possible and as small a cookie as possible.
> Additionally, applications should be able to handle the loss of a cookie."
>
> I only point these out as important as it could be very tempting for a
> developer to want to stick a large JSON packet in a cookie, but they must
> know the repercussions of doing this.
>
> -Dan

Reply via email to