= Сообщение: 41 из 1366 ============================================== UTF-8 = От : Konstantin Kuzov 2:5019/40 03 Jul 14 03:58:48 Кому : Michiel van der Vlist 03 Jul 14 03:58:48 Тема : Verify encoding FGHI : area://UTF-8?msgid=2:5019/40+53b49cb9 На : area://UTF-8?msgid=2:280/5555+35a3d7b0 = Кодировка сообщения определена как: CP866 ================================== ============================================================================== Konnichi wa, *Michiel-kun*! Aogu manako oyobi uketamawaru waga koe! Tomodachi _Michiel van der Vlist_ tsukuru airon _Maurice Kinal_ Nichiji - /*25 Апр 14 10:48*/, Daizai - /*Verify encoding*/:
MK>> I could have easily provided the wrong encoding in the CHRS MK>> kludge which happens all the time and for sure it would have been MK>> grunged
MvdV> Yes, I saw you attempts to throw sand in the gearbox.
MK>> just like the Russian site does with LATIN-1 since it uses iconv MK>> and there is no such thing as LATIN-1
MvdV> Yes, there is, LATIN-1 is mentioned in FTS-5003.
MK>> but instead LATIN1.
MvdV> LATIN1 is not a character encoding scheme that is "current practise" MvdV> in Fidonet.
MK>> I posted an example of this flaw in FTSC_PUBLIC to demonstrate MK>> the flaw along with evidence that it is so.
MvdV> The flaw is in the conversion that the webmaster od that site uses to MvdV> convert Fidonet messages to html. It is not the problem of Fidonet, MvdV> that he can't do it right.
I reply to this old mail as I only now read it and as the webmaster of aforementioned site I want to put my 50 cents.
What is right? And what I can't do right? My gate/bbs was implemented with idea in mind so messages fidonet->internet looks like normal internet pages and when message goes internet->fido it looks like a normal fidonet message here. So "detected" codepage which displayed on the internet side use more common to internet alias and it is a intended behaviour and there nothing wrong with that. Moreover all messages on internet side are always in UTF-8 and this message is here just for information. When message goes from fidonet tosser converts it on the fly to UTF-8 and saves to database. Why using UTF-8 internally? Mainly because of performance reasons, that way it converts only once and not on every message show. But with this decision we must make sure that message will be converted right in automatic mode or in worst case it can be "fixed" without loss later. Also as there are currently only one reliable way to determine the character set in fidonet - via CHRS kludge, we give a priority to it. So simplified gate algorithm of convertion in pseudo-code I'm using is like that:
----------------------------------------------------------------- if (CHRS_kludge is present) {