531158 - add support for IMAP ID extension rfc 2971

Status: RESOLVED FIXED

(MailNews Core :: Networking: IMAP, defect, P3)

Description

[David :Bienvenu]

The IMAP ID extension gives the client a way to identify itself to the server, mainly for statistical and bug-tracking reasons.  I think we should support this, probably with an opt-out pref. We already put the same information in the user-agent header.

Comment 1

[David :Bienvenu]

Attachment: add support for ID command

this sends the app name and version number to the server. We parse but ignore the response, which may contain information about the server. At some point, we may want to gather anonymous statistics about the hosts our users use.

Comment 2

[David :Bienvenu]

Attachment: store id response as a pref

this adds code to store the id response as a per-server pref. This could be useful for support issues, if we added the value of this pref to the about support output. The string pref is mail.server.serverX.serverIDResponse

Comment 3

[David :Bienvenu]

I'll need to add the ID command to the fake server before I can write a test for this...

Comment 4

[David :Bienvenu]

Attachment: fix with unit test

This adds a unit test, after adding support for the ID command to fakeserver. I had to move the getting of the xulappinfo to the UI thread in order for the xpcshell test to work, unfortunately.

Comment 5

[Mark Banner (:standard8)]

Comment on attachment 519221 [details] [diff] [review]
fix with unit test

+static nsCString gAppName;
+static nsCString gAppVersion;

Static nsCStrings aren't allowed (memory leaks on shutdown).

In comment 0 you said about an opt-out pref, I don't see that in the code.

Did you also say to me at some stage that we basically send this information already to IMAP servers via a different method?

Comment 6

[David :Bienvenu]

I can add an opt-out pref, but I can't imagine adding UI for it. Essentially it would be useful for the edge cases where the server advertises ID but doesn't handle it correctly.

I said that an IMAP server can essentially tell that it's talking to Thunderbird (or a derivative thereof) by the protocol pattern - it's essentially a fingerprint, and not really a secret, though it would require a bit of work.

Comment 7

[David :Bienvenu]

Attachment: fix addressing comments

added a hidden pref for turning this off, addressed other review comments.

Comment 8

[Mark Banner (:standard8)]

Comment on attachment 530382 [details] [diff] [review]
fix addressing comments

Review of attachment 530382 [details] [diff] [review]:
-----------------------------------------------------------------

This looks fine, we should probably get Neil to do a quick sr because of the minor api changes.

Comment 9

[David :Bienvenu]

fixed on trunk - http://hg.mozilla.org/comm-central/rev/a1c6b295010b