----- Original Message ----- From: "Jose Alberto Fernandez" <[EMAIL PROTECTED]> To: "Ant Developers List" <[EMAIL PROTECTED]> Sent: Monday, December 10, 2001 2:34 PM Subject: Re: [PATCH] srcFile to echo
>>From: "Steve Loughran" <[EMAIL PROTECTED]> >> >> We could have an encoding parameter; default to UTF8 for the sake of >> consistency with build files themselves. Or default to local? hmmm. My >> preference is UTF8. If the user has an alternative, it could be stated >> >How many files in your file system are *not* in your local encoding? well, my OS locale alternates between UK (home, laptop keyboards) and US (desk and laptop config), so although encoding is constant, locale and representation of time varies wildly (ex, I have to switch date from YYYY-MM-DD (ideal) to MM/DD/YY to import my US bank account data, and DD/MM/YY to import UK stuff. >Most of us in "latin-1" world do not really care, but I bet guys in Japan >and such may care. Are those people's files usualy in UTF8? I suspect that UTF16 may be quite common in the higher entropy languages; UTF8 is only an efficient coding for unicode where the %age of high byte content is low. > >The default should be the most natural or common one. In particular it would >be nice if the default encoding would match those used by other existing >tasks that look at the content of files in the file system Interesting issue. I quite like the fact that default of XML files is UTF8; stops surprises creeping in as stuff gets sent around. Unless otherwise stated, a file is going to parse consistently on their box to your box. But for importing other stuff, there is some value in having local input being the basic option: even though UTF8 would be the encoding I would tend to use. -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
