IMAP4 ID extension - On or off?

[AnthonyL]

Following the recent issues that Thunderbird users have had with UK's BT (British Telecom) due to a change in server configuration I've read the now locked:

https://bugzilla.mozilla.org/show_bug.cgi?id=1571655

and the RFC at https://tools.ietf.org/html/rfc2971

I carried out the change to
mail.server.default.send_client_info from True to False which worked. BT now say they have "implemented a workaround " and so reverting to True now works again.

However I have several email accounts in my Thunderbird setup and wonder if I have read the RFC correctly in which case is there any downside in leaving the value at False which will mitigate any further risk due to server misconfigurations with any of the providers?

[tanstaafl]

The recommendation in https://bugzilla.mozilla.org/show_bug.cgi?id=741374#c18 was to change it per (badly misconfigured server) server, not the default setting. My impression is that modifying a default setting has no effect on existing accounts, its only used when you create a new account. Perhaps your fix was actually due to BT making a change well before they announced it.

16adc000:mail.messagingengine.com:A:SendData: 5 ID ("name" "Thunderbird" "version" "52.8.0") is an excerpt from a old IMAP log showing the use of the ID command. The RFC claims that command is only meant for bug reports and usage statistics. The problem appears to be that the BT falsely claimed it supported the ID command and then when it received one it behaved badly. However, the same problem could occur if they lied about supporting a different IMAP command (since the problem was supposedly caused by a admin misconfiguring the server, not a bug in the server software).

I suggest you stick with the default value of true since Thunderbird has a small market share and providing this debugging information might be useful in the future (when Thunderbrid specific problems occur).

[AnthonyL]

Thanks.

I suggest you stick with the default value of true since Thunderbird has a small market share and providing this debugging information might be useful in the future (when Thunderbrid specific problems occur).

Conversely Thunderbird has a small market share and (careless) providers might not ensure testing or compliance. Two other clients I use, one on Linux and another on Android, were unaffected so presumably they don't bother with send this ID (can't see where it is set per server, and as the instruction is to exit and reload it would appear to be global).

Of those complaining there were only two clients mentioned, though the sample size is small.

Certainly in the BT case I could not save nor send mail which differs somewhat from the earlier example you mention.

Anyhow, I'll leave it be for the moment and try and remember this incident should/when something similar recurs.