Why would you need to preserve the technical details of the encoding, which is entirely platform-dependent? If you nevertheless want to upload the binary representation of text data, just read the file as an octet stream. Burdening the text interface with support for such exotic demands is not a good idea. Cheers, Chris
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Maciej Stachowiak Sent: Sunday, September 23, 2007 3:20 AM To: Křištof Želechovski Cc: [EMAIL PROTECTED]; 'Aaron Boodman'; 'Ian Hickson'; [EMAIL PROTECTED]; 'Dave Camp' Subject: Re: [whatwg] Offline Web Apps On Sep 22, 2007, at 8:48 AM, Křištof Želechovski wrote: > An arbitrary file is a sequence of bytes and it is up to the > application how > these bytes are interpreted as characters. > Moreover, a text file conceptually does not contain a string; it > contains, > by convention, lines of text. The result of reading a file as text > should > be a sequence of lines, not a string, with the line breaking > characters > removed. Disagree. If the goal is to later upload the text file, then you want to preserve line endings and other special characters as-is. Regards, Maciej
