User-Agent: Mozilla/5.0 (X11; U; Linux i686; rv:1.7.3) Gecko/20040914 Firefox/0.10 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; rv:1.7.3) Gecko/20040914 Firefox/0.10 Unlike Firefox or Seamonkey, Thunderbird doesn't have "View -> Character Encoding -> Auto-Detect -> Universal". Reproducible: Always Steps to Reproduce: N/A Actual Results: N/A Expected Results: N/A

The universal charset detector is an extremely useful feature and personally I miss it a lot (after switching from Mozilla MailNews). I would guess that it should be relatively easy to get it back; if this is the case, then IMHO it should be included in 1.0.

is this bug 161817 again ?

(In reply to comment #2) > is this bug 161817 again ? I doubt it. Bug 161817 seems to be a Windows-specific problem with how DLLs are handled on upgrade. This one happens on Linux.

Would have to look at new feature requests in 1.1 release. Might be possible for someone to develop an extension for this? Minusing for 1.0.

Am I right assuming that this is just a build config problem with actually including the universal charset detector in TBird (Basically, just adding universalchardet to MOZ_EXTENSIONS_DEFAULT in configure.in)? If so, could this be included in 1.0.2? Or is there some deeper problem with some code that still needs to be written?

if someone explains to me what it's used for and how valuable it could be in mail, I'll look at including it.

Universal Autodetector will (usually) correctly distinguish UTF-8, for a message that has no charset specified. I just ran into an example: attachment 182156 [details] (from bug 292329). (You'll need to have Display Attachments Inline to view the message correctly.) In Mozilla, if I have the Universal Autodetector turned on, that message is properly detected as UTF-8. In TB, I don't see UTF-8 being detected correctly by any of the supplied autodetectors -- and even if it did, someone who gets email primarily in English isn't going to expect to have to turn on an autodetector for Russian or Japanese or what have you. That said: the universal autodetector is a little buggy: it sometimes misdetects (bug 181344, bug 168526, bug 178495); and it doesn't support certain charsets, particularly Hebrew (bug 86999) and Arabic (bug 265030), probably others. It would be cool to have a Western autodetector that correctly distinguished between UTF-8 and ISO-8859-1 / Windows 1252. Such a detector would be relatively small and fairly easy to get right.

Mike are you in favor of adding the existing detector?

Short answer: Yes. It will be more help than hindrance to people who don't know from charsets.