[tlbuild] Installing the current snapshot

[Philipp Stephani]

Am 12.10.2009 um 14:53 schrieb Angelo Graziosi:

> I think that for Latin languages, the encoding should be Latin-1 or  
> similar (in my case, Firefox is using 'Occidentale ISO-8859-1'). It  
> is hard to think that those users (or the great part of them) use  
> different encoding.

Since every relevant combination of available operating systems,  
browsers and text editors has been supporting the Unicode encodings  
for a decade or so, I see no reason not to use it. Personally I treat  
every encoding that is unable to encode the whole Unicode scalar value  
space as a legacy encoding, though I admit that some of these "legacy"  
encodings (like ASCII and the ISO-8859 family) are still quite  
relevant. HTTP servers should in any case send the correct Content- 
Type header, and BOMs might be used to identify plain text files.