naive approach |
||
![]() ![]() Beta Tr flamebalrog
Sorry for the stupid questions that may (will!) follow but I want to make sure there's no doubt about...well, be prepared... On 2006-02-27 19:47:18, SAJChurchey wrote >Websites are reporting problems with Cookies in Windows XP SP2. The IE privacy settings are correct, and I've also modified the settings to be the most leniant they can be, but there are still problems. Are *all* the (client) computers runnign XP SP2 ? What's the server running ? 2k3 ? Samba as PDC ? (if you're at liberty to tell) Same version of IE everywhere ? (same "hotfix level" ?) >Roaming Profiles are currently being used, *Mandatory* roaming profiles ? >but this problem occurs inconsistently from machine-to-machine. The same account can access the site in some places but not others, By "inconsistently", do you mean that the same account on the same box will not always behave the same ? (just login/logout on the same box with the same user) Or, is it that when it starts not working for one user on one box it doesn't work again the next time(s) the same user logs in on the same box ? >and it occurs with various websites like the Microsoft Office ClipArt Gallery and the New York Times. Are there some websites that need (or just create without really needing them) cookies with which you never have any problem ? >Are there any settings that are machine-specific that might be causing these problems? Frankly ? Dunno :) (see post's subject) >I've tried all kinds of things for the past 3 days, and I haven't been able to determine the cause of the problem. Imagine spending three whole days trying to figure out why cookies won't work . . . it'll drive you to the brink of insanity. You still can write, so you're not totally lost. A common drawback when spending too much time on the same problem is your view of it often narrows to the only complicated stuff you don't need to know about (if you get what I mean). The 'trick' is to make one step back, do something else and then come back to it as a 'virgin' with the magical 'K.I.S.S.' principle in mind. (ok, this won't solve your problem, but it can help staying sane) >I've cleaned registries; I've created and reset a test profile multiple times, but I can't seem to hit it on the head. On a new box, with a new user (thus, a new profile) : are cookies created (the first time) ? Naive steps to the headache (if you've got time) : on a box where cookies don't seem to be created, set IE's home page to "about:blank", "install" (i.e. : copy) FileMon from Sysinternals, login as the user you're using for your tests, run FileMon with runas (as local admin), wait, filter out every app. you don't need to know about in FileMon (right-click, exclude process, loop), fire up IE like the user would do it, clear FileMon log (note : repeat the whole process again without this step if you don't find any failure), type the address of a site that needs cookies in the address bar and press [ENTER], login (on the site) or do whatever is supposed to create a cookie, save FileMon's log, search for failures in the log...any ? What directory ? User's profile ? ...mmmh...Other questions : - are profiles cached on (some of) the computers ? - what are the permissions on the user's profile directory when you can't taste these nice chocolate cookies web sites prepared for you ? (check with attrib too) ; - did you try with FireFox ? (note : I'm just asking if you tried, I'm not telling you to switch to FF...there's a difference) ; - ...sorry, forgot about my last question...it was about...erm...<f-word>...I may remember... (Lair) <Research-Labs> $SYS$BoycottSony.exe Replies:
|
||
| CyberArmy::Forum v0.6 Generated In 0.00662 seconds |