The R Project SVN R

Rev

Rev 90027 | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 90027 Rev 90200
Line 38... Line 38...
38 
   }
 38 
   }
39 
  \item{x}{the object to be written, typically a data frame.  If not, it
 39 
  \item{x}{the object to be written, typically a data frame.  If not, it
40 
    is attempted to coerce \code{x} to a data frame.}
 40 
    is attempted to coerce \code{x} to a data frame.}
41 
  \item{append}{logical.  If \code{TRUE}, the output is appended to the
 41 
  \item{append}{logical.  If \code{TRUE}, the output is appended to the
42 
    file.  If \code{FALSE}, any existing file of the name is destroyed.}
 42 
    file.  If \code{FALSE}, any existing file of the name is destroyed.}
43 
  \item{useBytes}{logical to be passed to \code{\link{writeLines}()},
 43 
  \item{useBytes}{ignored, and retained only for back-compatibility.
- 
 
 44 
    Since \code{write.dcf} now always writes UTF-8 (see \sQuote{Details}),
44 
    see there: \dQuote{for expert use}.}
 45 
    the output is no longer re-encoded to the session's native encoding.}
45 
  \item{indent}{a positive integer specifying the indentation for
 46 
  \item{indent}{a positive integer specifying the indentation for
46 
    continuation lines in output entries.}
 47 
    continuation lines in output entries.}
47 
  \item{width}{a positive integer giving the target column for wrapping
 48 
  \item{width}{a positive integer giving the target column for wrapping
48 
    lines in the output.}
 49 
    lines in the output.}
49 
}
 50 
}
Line 71... Line 72...
71 
    only) lines.
 72 
    only) lines.
72 
    \item Individual lines may not be arbitrarily long; prior to \R 3.0.2 the
 73 
    \item Individual lines may not be arbitrarily long; prior to \R 3.0.2 the
73 
    length limit was approximately 8191 bytes per line.
 74 
    length limit was approximately 8191 bytes per line.
74 
  }
 75 
  }
75 
 
 76 
 
- 
 
 77 
  DCF files are required to be encoded in UTF-8.  \code{read.dcf}
- 
 
 78 
  therefore returns values in UTF-8 (non-ASCII values being marked as
76 
  Note that \code{read.dcf(all = FALSE)} reads the file byte-by-byte.
 79 
  such), and \code{write.dcf} converts values with a declared encoding
- 
 
 80 
  (such as \code{"latin1"}) to UTF-8 and always writes UTF-8, regardless
77 
  This allows a \file{DESCRIPTION} file to be read and only its ASCII
 81 
  of the session's locale.  Any bytes which are not valid UTF-8 are not
- 
 
 82 
  dropped but preserved visually by escaping them as \samp{<xx>} (where
- 
 
 83 
  \code{xx} is the byte in hexadecimal), as done by
- 
 
 84 
  \code{\link{iconv}(sub = "byte")}.  As a consequence, the
78 
  fields used, or its \samp{Encoding} field used to re-encode the
 85 
  \samp{Encoding} field of a \file{DESCRIPTION} file is no longer used to
79 
  remaining fields.
 86 
  re-encode the other fields.
80 
 
 87 
 
81 
  \code{write.dcf} does not write \code{NA} fields.
 88 
  \code{write.dcf} does not write \code{NA} fields.
82 
}
 89 
}
83 
\value{
 90 
\value{
84 
  The default \code{read.dcf(all = FALSE)} returns a character matrix
 91 
  The default \code{read.dcf(all = FALSE)} returns a character matrix
Line 101... Line 108...
101 
 
 108 
 
102 
  For \code{write.dcf}, invisible \code{NULL}.
 109 
  For \code{write.dcf}, invisible \code{NULL}.
103 
}
 110 
}
104 
 
 111 
 
105 
\references{
 112 
\references{
106 
  \url{https://www.debian.org/doc/debian-policy/ch-controlfields.html}.
 113 
  \url{https://www.debian.org/doc/debian-policy/ch-controlfields.html},
107 
 
 - 
 
108 
  Note that \R does not require encoding in UTF-8, which is a recent
 114 
  which requires control files to be encoded in UTF-8.
109 
  Debian requirement.
 - 
 
110 
}
 115 
}
111 
 
 116 
 
112 
\note{
 117 
\note{
113 
  As from \R{} 3.4.0, \sQuote{whitespace} in all cases includes newlines.
 118 
  As from \R{} 3.4.0, \sQuote{whitespace} in all cases includes newlines.
114 
}
 119 
}