User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0 Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0 Recently, while using Firefox and using anything related to Google, an error message saying "An unexpected message was sent. Error code: - 12263". I upgraded to Firefox 2.0 to see if the problem fixed itself. It stopped for a few days. Now it keeps popping up every night. It has recently been doing it to me on other sites while I have my Gmail open. I do not know if this is a Google problem or a Firefox problem. Is this a security flaw, such as someone trying to get my information? Reproducible: Couldn't Reproduce

what do you mean "pops up"? A little alert box that says nothing but "Error code: -12263"? Does it appear inside the web page itself? This could be a Google server issue, or a browser bug, but it doesn't look like an exploitable flaw so I'm unhiding the bug so more people can potentially see this.

A quick google search showed me that there are more people with this problem, but I see no solutions up to now. http://forums.mozillazine.org/viewtopic.php?t=484288&sid=ae3dd1b87c49fdab3c2887dd8a62a5d8 http://groups.google.com/group/alt.fan.mozilla/browse_thread/thread/7d2bfbdf57792ff8/5776413b72683a20#5776413b72683a20 Maybe a new profile could help or a reinstall in a new clean directory? http://kb.mozillazine.org/Safe_Mode_(Firefox) http://kb.mozillazine.org/Profile_Folder

SSL_ERROR_RX_RECORD_TOO_LONG -12263 This is returned from only 5 places in the code, each time because we got more data from the server than we were expecting. Are we having compatibility problems with whatever server google-analytics is using?

Confirming as several people seem to get this error. Reassigning to NSS which is the component who reports this error code.

Nelson told me, when we get this error code from NSS, it's 100% sure that the server is bad, so I'm resolving this as an invalid bug.

This bug really upsets me. It illustrates all the issues I've been trying to raise with mozilla UI czars. It takes a server error (which NSS correctly diagnosed), compounds it with a bad UI message that blames the client, and then this bug tranfers responsibility for the whole mess onto NSS. Let's review it. Let's start with the error code. According to http://www.mozilla.org/projects/security/pki/nss/ref/ssl/sslerr.html > SSL_ERROR_RX_RECORD_TOO_LONG -12263 > "SSL received a record that exceeded the maximum permissible length." > > This generally indicates that the remote peer system has a flawed > implementation of SSL, and is violating the SSL specification. Read that last sentence carefully, again. Then tell me this is an NSS bug. Yes, servers make errors, even servers from big name companies. If PSM takes that error code and displays the text "An unexpected message was sent." then PSM needs to be fixed, NOW. It needs to be shot dead. It needs to be replaced with something that can provide a CLUE about the real problem. Bug 107491 is WAY past due. I can't see us waiting for FF3 for this! I won't survive until then! The text needs to CLEARLY and uneqivocally place the blame on the server. It needs to get the user to complain to the server administrator, NOT to the NSS developers!!

*** Bug 363823 has been marked as a duplicate of this bug. ***

*** Bug 364061 has been marked as a duplicate of this bug. ***

*** Bug 365030 has been marked as a duplicate of this bug. ***

I've received the following information by email from another user who had seen this error message: "I thought I'd pass on what "fixed" the problem I had regarding the -12263 error related to bug 360052. I've read the bug reports stating that this is a server-issue and that the error could be rewritten to be more specific (although likely a proxy issue with me). BTW: one of the servers I couldn't access was mozilla's ( bugzilla.mozilla.org and addons.mozilla.org). In order to access https sites using 2.0.0.3 through our proxy, I simply had to choose SOCKS v4, instead of the default of SOCKS v5. In 1.0.7 (the default-installed version), I could connect through our proxy regardless of the SOCKS selection."

Kai, please find out what you can about that broken proxy server. If we can identify it, we can warn others about it, and create an "evangelism" bug out of it.

(In reply to comment #11) > Kai, please find out what you can about that broken proxy server. > If we can identify it, we can warn others about it, and create an > "evangelism" bug out of it. I was told they are using: Squid/2.4.STABLE6

Should we change the summary of this bug to SSL problems with Squid Socks 4 proxy ?