Stefano Mazzocchi wrote:
[EMAIL PROTECTED] wrote:
antonio 2004/04/05 05:25:36
Log: removing encoding="ISO-8859-1"
Why?
Yes, why? UTF-8 is the XML default, but very few general purpose editors (i.e. non-XML editors) handle it correctly. So I'm wondering the value of this change : can't every dict file use the encoding that's most suited for its target language.
I agree with Sylvain, IMHO this clearly falls into the "if it ain't broke don't fix it" category of changes.
Also, it's fairly hard to check that the change of encoding declaration, and the related binary changes to adapt the files to the new encoding, have not caused any subtle problems.
Antonio, did you test all the impacted components after making your encoding changes? Do they work as before?
We've had problems with encodings, CVS and editors not handling all this properly before
(http://nagoya.apache.org/bugzilla/show_bug.cgi?id=19619), so we must be careful when making such changes. And not make them without a good reason IMHO.
...Also, how does CVS handle UTF-8 files? I mean what happens if a byte in a 2-bytes character is a "\n"? When checked out on Window, does it become "\r\n"? This can lead to very weird questions...
Yes. Hopefully, setting the "-kb" binary flag on these files avoids the problem, but we might then encounter the whitespace/extra end-of-lines problems again.
-Bertrand
