On Mon, Mar 09, 2009 at 11:35:57PM -0400, APseudoUtopia wrote: > On Mon, Mar 9, 2009 at 10:26 PM, Paul M Foster <paulf@xxxxxxxxxxxxxxxxx> > wrote: > > This is in two parts. First cookies. I'm a little unclear on how they > > work. From what I've read, cookies are stored by the browser. When a > > request for that cookie comes in from the server, the browser returns > > only the value, and no other data. One question: When the browser > > requests a page from a server, does it automatically search its cookies > > for that domain and send them along with the other requests? If that's > > now how it works, then how does it work? > > > > Second part is about sessions. According to the notes for the cookies > > page at php.net, it's considered bad practice to store user IDs and > > passwords in cookies. It's considered better practice to use PHP's > > native session-handling code to do this. But if a user has cookies > > turned on in the browser, then PHP will store the session information > > (possibly user ID and password) as a cookie. So what's the difference? > > > > The reference for the above is: > > http://us2.php.net/manual/en/features.cookies.php#36058 > > > > > > Paul > > > > -- > > Paul M. Foster > > > > -- > > PHP General Mailing List (http://www.php.net/) > > To unsubscribe, visit: http://www.php.net/unsub.php > > > > > > When a website sends the "Cookie:" in a HTTP header, the browser > decides what to do with it (or not to do). Generally, it saves the > cookie name and contents into a file. Although, various browsers > handle cookies differently, and some browsers ignore them completely > (or have options to). > > Within the cookie data are various things, such as the domain and > expiration. When you point the browser to a website, the browser > checks all it's cookies and see if the website matches the "domain" > field in any of the cookies. If so, it sends the name/content of the > cookie/cookies to the site in a/an HTTP header automatically. > > A very useful tool in monitoring all this and viewing what your > browser does behind the scenes is the Firefox extension "Live HTTP > Headers." > https://addons.mozilla.org/en-US/firefox/addon/3829 > > PHP sessions use cookies. When you call session_start() for the first > time, php randomly generates a unique hash ID for that session. It > sends it to the browser as a cookie with the name PHPSESSID (this is > customizable in php.ini). The server keeps a list of all the sessions > on the HDD (and expires them when needed, of course). When you store > any variable into the $_SESSION superglobal var, it stores the data ON > THE SERVER - nothing is sent to the browser. The browser only sends > the session ID cookie, which tells the server "hey, get the $_SESSION > data for this session ID." So it's up to the browser to send the > session cookie each time, else all the $_SESSION data is lost. > > Help clear it up for ya? Yes, this essentially echoes what Chis said. Thanks. It brings up another question, though. Let's say that I have a session_start() call at the beginning of a bunch of pages. So that each time one of these pages is called, the call is made to session_start(). It seems like it would screw things up royally if each call to session_start() generated a new SID. The browser sends an SID cookie to the server on page request. And if the page has a session_start() call at the beginning of it, is it reasonable to assume that the server will not generate a new SID, but instead use the one sent in the HTTP request to find its session information? (Assuming the timeout hasn't expired, the user is still using the same instance of the browser, etc.) Similarly, if an SID *isn't* sent by the browser, then a call to session_start() at the beginning of the page *will* generate a completely new SID, right? Paul -- Paul M. Foster -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
