Re: Log and special characters

On 02.08.2017 11:25, Ben RUBSON wrote:
> We would then be able to correctly log 'André'

Actually, this is how I most often get it, on the web and in scam emails :

"Hi André,"


To the savvy and experienced multilingual-application-programming expert, this of course 
is entirely transparent :
- the letters "é" are in reality the (bad) interpretation as ISO-8859-1, of the UTF-8 
2-bytes sequence \xc3\xa9, which as you well know now, represents the Unicode character 
with codepoint 233 (decimal) (or E9 (hexadecimal)), which is the printable latin letter "é".
- the misinterpretation is due to some program in the chain leading to this HTML page or 
email, which does not, or incorrectly, support UTF-8, and which has interpreted this as 2 
bytes, instead of 1 character.

(It gets even funnier when some other program in the chain tries to do "the right thing" 
and re-encodes these 2 characters as UTF-8, thus yelding 4 bytes which have nothing to do 
anymore with the original, no matter how encoded.  And I am sure than anyone dealing with 
the Chinese language has better stories to tell.)

Knowing how it happens does nothing to alleviate the frustration though.

It even happens with organisations which by all means /should/ really know better.
The following is extracted from emails received from the *Apache httpd* Developpers 
mailing list :

[...]
Weird behaviour with mod_ssl and SSLCryptoDevice
	85066 by: jean-frederic clere
	85067 by: Stefan Eissing
	85068 by: jean-frederic clere
	85069 by: jean-frederic clere
	85075 by: Jan Kaluža  <------------


[...]
scoreboard and http2
	85003 by: Stefan Eissing
	85004 by: Graham Leggett
	85005 by: Plüm, Rüdiger, Vodafone Group  <---

Poor Rüdiger, who systematically sees his name mangled each time he contributes.
And who knows what Jan's name really looks like ?

Mind you, this is about Apache httpd, the webserver which powers more than 50% of 
worldwide websites.  So I believe that other character-set confused programmers have some 
excuse.
:-)