User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0) Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8a2) Gecko/20040526 My home page is set to <http://my.yahoo.com/>. I have a cookie set to log me in automatically. When roaming is enabled, it does not read the cookie; I get the login screen. When I attempt to log in, it tells me that I am rejecting cookies. It reads the cookie correctly every time when roaming is disabled. Please note that this happens whether or not I have "cookies" as one of the files set to upload to the roaming server. Reproducible: Always Steps to Reproduce:

Do did you once roam "Cookie Permissions"? Probably dup of bug 244558. Please reopen, if wrong. *** This bug has been marked as a duplicate of 244558 ***

Nope, that's not it. Just to be sure, I downloaded the latest build, created a new profile locally and on the server, and *only* selected bookmarks, address book, and user.js. Problem still persists as long as I have "enable roaming access" checked.

What's in your user.js? Anything relevant?

I was seeing this as well. I'm testing this on my home box for now, if you need any more data points I'd be happy to gather them when I get home tonight. I'm running the nightlies on WinXP using Modern chrome. I can try the same thing as Kim - new install and profile. For me it happens when I first start Moz. If I try to login to groups.yahoo.com it won't take the cookie. If I blow away any existing cookies (for example - manage cookes and delete the Yahoo cookies) it'll work from then on until I stop Moz and restart it.

Do you see the same on other sites? Can you check the cookies.txt and hostperm.1 and cookperm.txt (if existant) - before you close Mozilla - after you closed Mozilla - after you restarted Mozilla if it looks sane? I wonder if I accidently overwrite it or anything. Apart from bug 244558, I can't think of a cause for this bug.

Okay, I created a brand new user profile in Mozilla and didn't mess with any of the defaults, except I set it to ask me for each cookie so I could limit it to just Yahoo for testing purposes. When roaming is not enabled, Yahoo sets a bunch of cookies and there's no problems. It modifies a couple whenever I exit Mozilla, restart, and re-enter My Yahoo, but it doesn't add any new ones. However, when roaming is enabled and I try to enter My Yahoo, it seems to duplicate cookies, only with slightly different values. For instance, I always had one cookie like this: .my.yahoo.com TRUE / FALSE 1271361600 V T=626375&G=75497 (the numbers have been changed) The value of "T" stays consistent but "G" changes. When roaming is enabled, there is a second cookie added with the same value of "T" but a new value for "G". Similar behavior is being seen on another cookie. So it looks to me like enabling roaming makes Mozilla add new cookies rather than modify existing ones. Oddly enough, I am eating cookies while writing this.

> Oddly enough, I am eating cookies while writing this. Yeah, maybe Mozilla roams so much that it wants to eat cookies, too. Mozilla is a big beast, so it eats all it can get. Nice cats, BTW. I hope Mozilla won't eat *them*.

Ben, any news on this? Not to be a pest, but this is a major loss of functionality. I can't use roaming until this is fixed. Let me know if there's any other information I can provide.

No, and I'm currently working on the other major roaming bugs. Your description didn't ring a bell. > Let me know if there's any other information I can provide. Would be interesting to compare with other sites, esp. those that you do not regularily visit, to exclude problems with cookperm / hostperm. > except I set it to ask me for each cookie so I could limit it to > just Yahoo for testing purposes. Leave that alone (in a fresh profile), please, with that you set host permissions.

> So it looks to me like enabling roaming makes Mozilla add new cookies rather > than modify existing ones. I see that as well. When roaming the cookies.txt file, it file shows a bunch of identical cookies from gmail, google, slashdot, lwn, etc.

qawanted: Need somebody to figure out what's going on.

I would be interested in helping test/investigate this, but I use Thunderbird. Any ETA for getting roaming ported to Thunderbird to get more eyeballs on it?

No, if you want to help, getting a copy of the suite should be the least of the problem. Besides, Thunderbird is not a browser, this is about cookies, and may be application-specific, too.

My bad. I could test with Firefox, which I am also using. But since a port to Aviary seems unlikely soon, I'll defer this to other testers for now.

What's needed here is dedicated testing: creating a new profile, cookies for a few sites including my yahoo and others, observing the files (permissions etc.) and their contents at various moments (before browser close, after close, right before download, right after download / before cookie manager reads (if possible)) and in the cookie manager UI before and after visiting the site and writing down any changes. You can't do that as fly-by testing.

Oh, and then we need an idea why the change of cookies happens when it shouldn't. I don't have any currently.

(In reply to comment #16) > What's needed here is dedicated testing: [...] > You can't do that as fly-by testing. I knew that, I'm just not interested in testing the Suite right now. ("fly-by testing", LOL.) You're doing a great job, Ben. Good luck!

Kim, I'm unable to reproduce the problem with Mozilla 1.8a6. Could you test and make sure that you are still experiencing the problem with that build?

No word form Kim so I'm marking this worksforme. Reopen if you are able to reproduce this with a newer build.