Created attachment 488789 [details] [diff] [review] rm I was actually going to start on bug 496429, but I ran into too many of these instead of app-specific prefs. I removed the last use of mail.migration.copyMailFiles in bug 608638, which I call a good thing. news.wrap_long_lines was unused as far back as Classic - maybe it was used in Netscape 4.x, and got the axe after that, dunno. mail.server_type reflected whether you had a single POP account in Netscape 4.x, or an IMAP account (or accounts, since the pref with the server name was plural for IMAP) mail.default_drafts and mail.default_templates might have been just what they sound like they were - a way to change the default name for the Drafts and Templates folders - though the lack of any users of them in Classic is suspicious mail.imap.server_sub_directory seems unlikely to have ever worked, in any world that had more than a single IMAP account mail.imap.max_cached_connections is not the number we use by default (that's 5), and we do not use it to initialize the actual mail.server.serverN.max_cached_connections pref (we don't initialize it at all, displaying a rather inaccurate "0" when you first open advanced server prefs for an IMAP server) mail.imap.new_mail_get_headers is more convenient than most, because nsIMsgIncomingServer.(idl|cpp) have comments explaining that it was the 4.x global that's now the per-server downloadOnBiff mail.imap.cleanup_inbox_on_exit could have been anything, or possibly nothing mail.leave_on_server appears to have been the global version of the current per-server pref as does mail.check_new_mail (both times) mail.pop3_gets_new_mail, as the comments in nsIMsgIncomingServer.(idl|cpp) say, was the 4.x version of the per-server downloadOnBiff for POP mail.check_time is another global now per-server mail.pop_name looks like it was the username for your single POP account mail.pop_password was used in Classic's cmd/winfe/femess.cpp. Classic. mail.identity.* are now per-identity, and while extensions/pref/autoconfig does think it might be able to use mail.identity.useremail, that's just one of its many many problems mail.use_fcc is per-identity, as in mail.cc_self, as is mail.attach_vcard, as is mail.fcc_folder mail.request.return_receipt became unused between 0.9.4 and 1.0. If you want to know what replaced it, I recommend jabbing yourself in the eye with a fork until all the tines are bent, and then asking yourself if you really want some more of the same. Despite the comment in all-thunderbird.js, mail.showPreviewText does nothing, and "enables preview text in mail alerts and folder tooltips" is actually mail.biff.alert.show_preview news.use_fcc is per-server, as is cc_self and fcc_folder news.notify.on is the oddly-named pref for "ask me before downloading more than n message", both it and its friends .max_articles and .mark_old_read are per-server The volume of ldap_2.autoComplete removals worries me, but near as I can tell they were just added by copying all the ldap_1.autoComplete prefs, adding some, and never using any of them (though from 0.9.4 to 1.9.1, skipDirectoryIfLocalMatchFound had UI to set it, which is nice, it just didn't _ever_ have anything that actually consumed it) Setting mail.signature_file for *nix, and then unsetting it for all in all-thunderbird.js and browser-prefs.js would be sort of silly, except that it's unused since it's per-identity (so it's still sort of silly). as are the unused mail.directory and news.directory prefs mailnews.reply_with_extra_lines apparently reflects how it used to be important for *nix users to jam their replies right up against the quote, while everyone else wanted some space between them Among the rest of the totally unused ifdef *nix prefs, the most amusing are the sash_geometry ones, which if you dig up Netscape 4 pref documentation, were unused even then
Comment on attachment 488789 [details] [diff] [review] rm (In reply to comment #0) > mail.imap.new_mail_get_headers is more convenient than most, because > nsIMsgIncomingServer.(idl|cpp) have comments explaining that it was the 4.x > global that's now the per-server downloadOnBiff I think we should probably remove those comments (including the bit relating to mail.pop3_gets_new_mail). I think removing the old pref and not having any migration from the old system, means that we just don't need to keep a record of it in the code. r=Standard8 with those comments removed.
And http://hg.mozilla.org/comm-central/rev/582f934b7f0b because it's so much more fun to address review comments after you've landed than before.