Source file src/fmt/doc.go

     1  // Copyright 2009 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  /*
     6  Package fmt implements formatted I/O with functions analogous
     7  to C's printf and scanf.  The format 'verbs' are derived from C's but
     8  are simpler.
     9  
    10  # Printing
    11  
    12  The verbs:
    13  
    14  General:
    15  
    16  	%v	the value in a default format
    17  		when printing structs, the plus flag (%+v) adds field names
    18  	%#v	a Go-syntax representation of the value
    19  		(floating-point infinities and NaNs print as ±Inf and NaN)
    20  	%T	a Go-syntax representation of the type of the value
    21  	%%	a literal percent sign; consumes no value
    22  
    23  Boolean:
    24  
    25  	%t	the word true or false
    26  
    27  Integer:
    28  
    29  	%b	base 2
    30  	%c	the character represented by the corresponding Unicode code point
    31  	%d	base 10
    32  	%o	base 8
    33  	%O	base 8 with 0o prefix
    34  	%q	a single-quoted character literal safely escaped with Go syntax.
    35  	%x	base 16, with lower-case letters for a-f
    36  	%X	base 16, with upper-case letters for A-F
    37  	%U	Unicode format: U+1234; same as "U+%04X"
    38  
    39  Floating-point and complex constituents:
    40  
    41  	%b	decimalless scientific notation with exponent a power of two,
    42  		in the manner of strconv.FormatFloat with the 'b' format,
    43  		e.g. -123456p-78
    44  	%e	scientific notation, e.g. -1.234456e+78
    45  	%E	scientific notation, e.g. -1.234456E+78
    46  	%f	decimal point but no exponent, e.g. 123.456
    47  	%F	synonym for %f
    48  	%g	%e for large exponents, %f otherwise. Precision is discussed below.
    49  	%G	%E for large exponents, %F otherwise
    50  	%x	hexadecimal notation (with decimal power of two exponent), e.g. -0x1.23abcp+20
    51  	%X	upper-case hexadecimal notation, e.g. -0X1.23ABCP+20
    52  
    53  String and slice of bytes (treated equivalently with these verbs):
    54  
    55  	%s	the uninterpreted bytes of the string or slice
    56  	%q	a double-quoted string safely escaped with Go syntax
    57  	%x	base 16, lower-case, two characters per byte
    58  	%X	base 16, upper-case, two characters per byte
    59  
    60  Slice:
    61  
    62  	%p	address of 0th element in base 16 notation, with leading 0x
    63  
    64  Pointer:
    65  
    66  	%p	base 16 notation, with leading 0x
    67  	The %b, %d, %o, %x and %X verbs also work with pointers,
    68  	formatting the value exactly as if it were an integer.
    69  
    70  The default format for %v is:
    71  
    72  	bool:                    %t
    73  	int, int8 etc.:          %d
    74  	uint, uint8 etc.:        %d, %#x if printed with %#v
    75  	float32, complex64, etc: %g
    76  	string:                  %s
    77  	chan:                    %p
    78  	pointer:                 %p
    79  
    80  For compound objects, the elements are printed using these rules, recursively,
    81  laid out like this:
    82  
    83  	struct:             {field0 field1 ...}
    84  	array, slice:       [elem0 elem1 ...]
    85  	maps:               map[key1:value1 key2:value2 ...]
    86  	pointer to above:   &{}, &[], &map[]
    87  
    88  Width is specified by an optional decimal number immediately preceding the verb.
    89  If absent, the width is whatever is necessary to represent the value.
    90  Precision is specified after the (optional) width by a period followed by a
    91  decimal number. If no period is present, a default precision is used.
    92  A period with no following number specifies a precision of zero.
    93  Examples:
    94  
    95  	%f     default width, default precision
    96  	%9f    width 9, default precision
    97  	%.2f   default width, precision 2
    98  	%9.2f  width 9, precision 2
    99  	%9.f   width 9, precision 0
   100  
   101  Width and precision are measured in units of Unicode code points,
   102  that is, runes. (This differs from C's printf where the
   103  units are always measured in bytes.) Either or both of the flags
   104  may be replaced with the character '*', causing their values to be
   105  obtained from the next operand (preceding the one to format),
   106  which must be of type int.
   107  
   108  For most values, width is the minimum number of runes to output,
   109  padding the formatted form with spaces if necessary.
   110  
   111  For strings, byte slices and byte arrays, however, precision
   112  limits the length of the input to be formatted (not the size of
   113  the output), truncating if necessary. Normally it is measured in
   114  runes, but for these types when formatted with the %x or %X format
   115  it is measured in bytes.
   116  
   117  For floating-point values, width sets the minimum width of the field and
   118  precision sets the number of places after the decimal, if appropriate,
   119  except that for %g/%G precision sets the maximum number of significant
   120  digits (trailing zeros are removed). For example, given 12.345 the format
   121  %6.3f prints 12.345 while %.3g prints 12.3. The default precision for %e, %f
   122  and %#g is 6; for %g it is the smallest number of digits necessary to identify
   123  the value uniquely.
   124  
   125  For complex numbers, the width and precision apply to the two
   126  components independently and the result is parenthesized, so %f applied
   127  to 1.2+3.4i produces (1.200000+3.400000i).
   128  
   129  When formatting a single integer code point or a rune string (type []rune)
   130  with %q, invalid Unicode code points are changed to the Unicode replacement
   131  character, U+FFFD, as in [strconv.QuoteRune].
   132  
   133  Other flags:
   134  
   135  	'+'	always print a sign for numeric values;
   136  		guarantee ASCII-only output for %q (%+q)
   137  	'-'	pad with spaces on the right rather than the left (left-justify the field)
   138  	'#'	alternate format: add leading 0b for binary (%#b), 0 for octal (%#o),
   139  		0x or 0X for hex (%#x or %#X); suppress 0x for %p (%#p);
   140  		for %q, print a raw (backquoted) string if [strconv.CanBackquote]
   141  		returns true;
   142  		always print a decimal point for %e, %E, %f, %F, %g and %G;
   143  		do not remove trailing zeros for %g and %G;
   144  		write e.g. U+0078 'x' if the character is printable for %U (%#U)
   145  	' '	(space) leave a space for elided sign in numbers (% d);
   146  		put spaces between bytes printing strings or slices in hex (% x, % X)
   147  	'0'	pad with leading zeros rather than spaces;
   148  		for numbers, this moves the padding after the sign
   149  
   150  Flags are ignored by verbs that do not expect them.
   151  For example there is no alternate decimal format, so %#d and %d
   152  behave identically.
   153  
   154  For each Printf-like function, there is also a Print function
   155  that takes no format and is equivalent to saying %v for every
   156  operand.  Another variant Println inserts blanks between
   157  operands and appends a newline.
   158  
   159  Regardless of the verb, if an operand is an interface value,
   160  the internal concrete value is used, not the interface itself.
   161  Thus:
   162  
   163  	var i interface{} = 23
   164  	fmt.Printf("%v\n", i)
   165  
   166  will print 23.
   167  
   168  Except when printed using the verbs %T and %p, special
   169  formatting considerations apply for operands that implement
   170  certain interfaces. In order of application:
   171  
   172  1. If the operand is a [reflect.Value], the operand is replaced by the
   173  concrete value that it holds, and printing continues with the next rule.
   174  
   175  2. If an operand implements the [Formatter] interface, it will
   176  be invoked. In this case the interpretation of verbs and flags is
   177  controlled by that implementation.
   178  
   179  3. If the %v verb is used with the # flag (%#v) and the operand
   180  implements the [GoStringer] interface, that will be invoked.
   181  
   182  If the format (which is implicitly %v for [Println] etc.) is valid
   183  for a string (%s %q %x %X), or is %v but not %#v,
   184  the following two rules apply:
   185  
   186  4. If an operand implements the error interface, the Error method
   187  will be invoked to convert the object to a string, which will then
   188  be formatted as required by the verb (if any).
   189  
   190  5. If an operand implements method String() string, that method
   191  will be invoked to convert the object to a string, which will then
   192  be formatted as required by the verb (if any).
   193  
   194  For compound operands such as slices and structs, the format
   195  applies to the elements of each operand, recursively, not to the
   196  operand as a whole. Thus %q will quote each element of a slice
   197  of strings, and %6.2f will control formatting for each element
   198  of a floating-point array.
   199  
   200  However, when printing a byte slice with a string-like verb
   201  (%s %q %x %X), it is treated identically to a string, as a single item.
   202  
   203  To avoid recursion in cases such as
   204  
   205  	type X string
   206  	func (x X) String() string { return Sprintf("<%s>", x) }
   207  
   208  convert the value before recurring:
   209  
   210  	func (x X) String() string { return Sprintf("<%s>", string(x)) }
   211  
   212  Infinite recursion can also be triggered by self-referential data
   213  structures, such as a slice that contains itself as an element, if
   214  that type has a String method. Such pathologies are rare, however,
   215  and the package does not protect against them.
   216  
   217  When printing a struct, fmt cannot and therefore does not invoke
   218  formatting methods such as Error or String on unexported fields.
   219  
   220  # Explicit argument indexes
   221  
   222  In [Printf], [Sprintf], and [Fprintf], the default behavior is for each
   223  formatting verb to format successive arguments passed in the call.
   224  However, the notation [n] immediately before the verb indicates that the
   225  nth one-indexed argument is to be formatted instead. The same notation
   226  before a '*' for a width or precision selects the argument index holding
   227  the value. After processing a bracketed expression [n], subsequent verbs
   228  will use arguments n+1, n+2, etc. unless otherwise directed.
   229  
   230  For example,
   231  
   232  	fmt.Sprintf("%[2]d %[1]d\n", 11, 22)
   233  
   234  will yield "22 11", while
   235  
   236  	fmt.Sprintf("%[3]*.[2]*[1]f", 12.0, 2, 6)
   237  
   238  equivalent to
   239  
   240  	fmt.Sprintf("%6.2f", 12.0)
   241  
   242  will yield " 12.00". Because an explicit index affects subsequent verbs,
   243  this notation can be used to print the same values multiple times
   244  by resetting the index for the first argument to be repeated:
   245  
   246  	fmt.Sprintf("%d %d %#[1]x %#x", 16, 17)
   247  
   248  will yield "16 17 0x10 0x11".
   249  
   250  # Format errors
   251  
   252  If an invalid argument is given for a verb, such as providing
   253  a string to %d, the generated string will contain a
   254  description of the problem, as in these examples:
   255  
   256  	Wrong type or unknown verb: %!verb(type=value)
   257  		Printf("%d", "hi"):        %!d(string=hi)
   258  	Too many arguments: %!(EXTRA type=value)
   259  		Printf("hi", "guys"):      hi%!(EXTRA string=guys)
   260  	Too few arguments: %!verb(MISSING)
   261  		Printf("hi%d"):            hi%!d(MISSING)
   262  	Non-int for width or precision: %!(BADWIDTH) or %!(BADPREC)
   263  		Printf("%*s", 4.5, "hi"):  %!(BADWIDTH)hi
   264  		Printf("%.*s", 4.5, "hi"): %!(BADPREC)hi
   265  	Invalid or invalid use of argument index: %!(BADINDEX)
   266  		Printf("%*[2]d", 7):       %!d(BADINDEX)
   267  		Printf("%.[2]d", 7):       %!d(BADINDEX)
   268  
   269  All errors begin with the string "%!" followed sometimes
   270  by a single character (the verb) and end with a parenthesized
   271  description.
   272  
   273  If an Error or String method triggers a panic when called by a
   274  print routine, the fmt package reformats the error message
   275  from the panic, decorating it with an indication that it came
   276  through the fmt package.  For example, if a String method
   277  calls panic("bad"), the resulting formatted message will look
   278  like
   279  
   280  	%!s(PANIC=bad)
   281  
   282  The %!s just shows the print verb in use when the failure
   283  occurred. If the panic is caused by a nil receiver to an Error
   284  or String method, however, the output is the undecorated
   285  string, "<nil>".
   286  
   287  # Scanning
   288  
   289  An analogous set of functions scans formatted text to yield
   290  values.  [Scan], [Scanf] and [Scanln] read from [os.Stdin]; [Fscan],
   291  [Fscanf] and [Fscanln] read from a specified [io.Reader]; [Sscan],
   292  [Sscanf] and [Sscanln] read from an argument string.
   293  
   294  [Scan], [Fscan], [Sscan] treat newlines in the input as spaces.
   295  
   296  [Scanln], [Fscanln] and [Sscanln] stop scanning at a newline and
   297  require that the items be followed by a newline or EOF.
   298  
   299  [Scanf], [Fscanf], and [Sscanf] parse the arguments according to a
   300  format string, analogous to that of [Printf]. In the text that
   301  follows, 'space' means any Unicode whitespace character
   302  except newline.
   303  
   304  In the format string, a verb introduced by the % character
   305  consumes and parses input; these verbs are described in more
   306  detail below. A character other than %, space, or newline in
   307  the format consumes exactly that input character, which must
   308  be present. A newline with zero or more spaces before it in
   309  the format string consumes zero or more spaces in the input
   310  followed by a single newline or the end of the input. A space
   311  following a newline in the format string consumes zero or more
   312  spaces in the input. Otherwise, any run of one or more spaces
   313  in the format string consumes as many spaces as possible in
   314  the input. Unless the run of spaces in the format string
   315  appears adjacent to a newline, the run must consume at least
   316  one space from the input or find the end of the input.
   317  
   318  The handling of spaces and newlines differs from that of C's
   319  scanf family: in C, newlines are treated as any other space,
   320  and it is never an error when a run of spaces in the format
   321  string finds no spaces to consume in the input.
   322  
   323  The verbs behave analogously to those of [Printf].
   324  For example, %x will scan an integer as a hexadecimal number,
   325  and %v will scan the default representation format for the value.
   326  The [Printf] verbs %p and %T and the flags # and + are not implemented.
   327  For floating-point and complex values, all valid formatting verbs
   328  (%b %e %E %f %F %g %G %x %X and %v) are equivalent and accept
   329  both decimal and hexadecimal notation (for example: "2.3e+7", "0x4.5p-8")
   330  and digit-separating underscores (for example: "3.14159_26535_89793").
   331  
   332  Input processed by verbs is implicitly space-delimited: the
   333  implementation of every verb except %c starts by discarding
   334  leading spaces from the remaining input, and the %s verb
   335  (and %v reading into a string) stops consuming input at the first
   336  space or newline character.
   337  
   338  The familiar base-setting prefixes 0b (binary), 0o and 0 (octal),
   339  and 0x (hexadecimal) are accepted when scanning integers
   340  without a format or with the %v verb, as are digit-separating
   341  underscores.
   342  
   343  Width is interpreted in the input text but there is no
   344  syntax for scanning with a precision (no %5.2f, just %5f).
   345  If width is provided, it applies after leading spaces are
   346  trimmed and specifies the maximum number of runes to read
   347  to satisfy the verb. For example,
   348  
   349  	Sscanf(" 1234567 ", "%5s%d", &s, &i)
   350  
   351  will set s to "12345" and i to 67 while
   352  
   353  	Sscanf(" 12 34 567 ", "%5s%d", &s, &i)
   354  
   355  will set s to "12" and i to 34.
   356  
   357  In all the scanning functions, a carriage return followed
   358  immediately by a newline is treated as a plain newline
   359  (\r\n means the same as \n).
   360  
   361  In all the scanning functions, if an operand implements method
   362  [Scan] (that is, it implements the [Scanner] interface) that
   363  method will be used to scan the text for that operand.  Also,
   364  if the number of arguments scanned is less than the number of
   365  arguments provided, an error is returned.
   366  
   367  All arguments to be scanned must be either pointers to basic
   368  types or implementations of the [Scanner] interface.
   369  
   370  Like [Scanf] and [Fscanf], [Sscanf] need not consume its entire input.
   371  There is no way to recover how much of the input string [Sscanf] used.
   372  
   373  Note: [Fscan] etc. can read one character (rune) past the input
   374  they return, which means that a loop calling a scan routine
   375  may skip some of the input.  This is usually a problem only
   376  when there is no space between input values.  If the reader
   377  provided to [Fscan] implements ReadRune, that method will be used
   378  to read characters.  If the reader also implements UnreadRune,
   379  that method will be used to save the character and successive
   380  calls will not lose data.  To attach ReadRune and UnreadRune
   381  methods to a reader without that capability, use
   382  [bufio.NewReader].
   383  */
   384  package fmt
   385  

View as plain text