/usr/lib/ocaml/netstring/netnumber.mli is in libocamlnet-ocaml-dev 3.7.3-4.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 | (* $Id: netnumber.mli 1701 2012-02-14 14:48:46Z gerd $ *)
(** Binary encodings of numbers *)
(** This is the successor of the {!Rtypes} module *)
(** {2 Numeric types} *)
(* 4- and 8-bytes representation of signed integers *)
type int4
(** 32 bit signed integer *)
type int8
(** 64 bit signed integer *)
(* 4- and 8-bytes representation of non-negative integers *)
type uint4
(** 32 bit unsigned integer *)
type uint8
(** 64 bit unsigned integer *)
(* Floating-point numbers of single and double precision according to IEEE *)
type fp4
(** single precision float (IEEE "float") *)
type fp8
(** double precision float (IEEE "double") *)
exception Cannot_represent of string
(** raised if a conversion can't be done *)
exception Out_of_range
(** raised if string position out of range *)
(** {2 Basic encoding/decoding functions} *)
val mk_int4 : char * char * char * char -> int4
val mk_int8 : char * char * char * char * char * char * char * char -> int8
val mk_uint4 : char * char * char * char -> uint4
val mk_uint8 : char * char * char * char * char * char * char * char -> uint8
(** [mk_]<t> create integer values from character tuples. In these tuples
* the MSB is the first component and the LSB the last.
*)
(* destroy integers and get tuples *)
val dest_int4 : int4 -> char * char * char * char
val dest_int8 : int8 -> char * char * char * char * char * char * char * char
val dest_uint4 : uint4 -> char * char * char * char
val dest_uint8 : uint8 -> char * char * char * char * char * char * char * char
(** [dest_]<t> destroy integer values and returns the corresponding char
* tuples.
*)
val mk_fp4 : char * char * char * char -> fp4
val mk_fp8 : char * char * char * char * char * char * char * char -> fp8
val dest_fp4 : fp4 -> char * char * char * char
val dest_fp8 : fp8 -> char * char * char * char * char * char * char * char
(** {2 Conversions} *)
(** Conversions from int to (u)int and vice versa.
* On 32-bit computers, the type [int] can hold 31-bit signed integers
* (including the sign, i.e. one bit cannot be used).
* On 64-bit computers, the type [int] can hold 63-bit signed integers
* (including the sign, i.e. one bit cannot be used).
* The [int_of_xxx] functions raise [Cannot_represent] if the number to
* convert is too big (or too small) to be represented as [int]. Note
* that this depends on the word size of your architecture.
*)
val int_of_int4 : int4 -> int
val int_of_uint4 : uint4 -> int
val int_of_int8 : int8 -> int
val int_of_uint8 : uint8 -> int
val int4_of_int : int -> int4
val uint4_of_int : int -> uint4
val int8_of_int : int -> int8
val uint8_of_int : int -> uint8
val int32_of_int4 : int4 -> int32
val int32_of_uint4 : uint4 -> int32
val int32_of_int8 : int8 -> int32
val int32_of_uint8 : uint8 -> int32
val int4_of_int32 : int32 -> int4
val uint4_of_int32 : int32 -> uint4
val int8_of_int32 : int32 -> int8
val uint8_of_int32 : int32 -> uint8
val int64_of_int4 : int4 -> int64
val int64_of_uint4 : uint4 -> int64
val int64_of_int8 : int8 -> int64
val int64_of_uint8 : uint8 -> int64
val int4_of_int64 : int64 -> int4
val uint4_of_int64 : int64 -> uint4
val int8_of_int64 : int64 -> int8
val uint8_of_int64 : int64 -> uint8
(** Casts from [uint4]/[uint8] to [int32]/[int64]. Here, the sign is ignored and
* simply considered as a bit.
*)
val logical_uint4_of_int32 : int32 -> uint4
val logical_int32_of_uint4 : uint4 -> int32
val logical_uint8_of_int64 : int64 -> uint8
val logical_int64_of_uint8 : uint8 -> int64
val fp8_of_fp4 : fp4 -> fp8
val fp4_of_fp8 : fp8 -> fp4
(** Note [fp4_of_fp8]: This conversion is not exact. It is quite
* normal that precision is lost. Numbers too small or too large
* for fp4 are converted to the "infinity" value.
*)
val float_of_fp4 : fp4 -> float
val float_of_fp8 : fp8 -> float
val fp4_of_float : float -> fp4
val fp8_of_float : float -> fp8
(** Note fp4_of_float: The same problems as in fp4_of_fp8 may arise *)
(** {2 Comparisons} *)
(** The comparisons "=" and "<>" work for all numbers.
For signed integers, the operators "<", "<=", ">", and ">=" work, too.
The unsigned integer type use representation that are not compatible
with these operators, and the following functions need to be called.
For [fp4] and [fp8] there are no comparison functions - convert to
[float] first and compare then.
*)
val lt_uint4 : uint4 -> uint4 -> bool
(** [lt_uint4] is true iff the first value is less than the second value
as unsigned int
*)
val le_uint4 : uint4 -> uint4 -> bool
val gt_uint4 : uint4 -> uint4 -> bool
val ge_uint4 : uint4 -> uint4 -> bool
(** Other comparisons *)
val lt_uint8 : uint8 -> uint8 -> bool
(** [lt_uint8] is true iff the first value is less than the second value
as unsigned int
*)
val le_uint8 : uint8 -> uint8 -> bool
val gt_uint8 : uint8 -> uint8 -> bool
val ge_uint8 : uint8 -> uint8 -> bool
(** Other comparisons *)
(** {2 Minimum/maximum values} *)
val min_int4 : int4
val min_uint4 : uint4
val min_int8 : int8
val min_uint8 : uint8
val max_int4 : int4
val max_uint4 : uint4
val max_int8 : int8
val max_uint8 : uint8
module type ENCDEC = sig
(** Encode/decode numbers as strings. These functions exist in two
flavors:
- {!Netnumber.BE} implements network byte order (big endian)
- {!Netnumber.LE} implements little endian
*)
val read_int4 : string -> int -> int4
val read_int8 : string -> int -> int8
val read_uint4 : string -> int -> uint4
val read_uint8 : string -> int -> uint8
(** [read_]<t> create integer values from the characters found at a
certain position in the string. Raises [Out_of_range] if the position
is bad
*)
val read_int4_unsafe : string -> int -> int4
val read_int8_unsafe : string -> int -> int8
val read_uint4_unsafe : string -> int -> uint4
val read_uint8_unsafe : string -> int -> uint8
(** Same, but no index check *)
val write_int4 : string -> int -> int4 -> unit
val write_int8 : string -> int -> int8 -> unit
val write_uint4 : string -> int -> uint4 -> unit
val write_uint8 : string -> int -> uint8 -> unit
(** [write_]<t> copies the characters corresponding to the integer values
into the string at the given positions. Raises [Out_of_range] if the
position is bad.
*)
val write_int4_unsafe : string -> int -> int4 -> unit
val write_int8_unsafe : string -> int -> int8 -> unit
val write_uint4_unsafe : string -> int -> uint4 -> unit
val write_uint8_unsafe : string -> int -> uint8 -> unit
(** [write_]<t>[_unsafe]: Same, but no index check. *)
val int4_as_string : int4 -> string
val int8_as_string : int8 -> string
val uint4_as_string : uint4 -> string
val uint8_as_string : uint8 -> string
(** <t>[_as_string]: Returns the corresponding string for an integer value
*)
val write_fp4 : string -> int -> fp4 -> unit
val write_fp8 : string -> int -> fp8 -> unit
val fp4_as_string : fp4 -> string
val fp8_as_string : fp8 -> string
val read_fp4 : string -> int -> fp4
val read_fp8 : string -> int -> fp8
end
module BE : ENCDEC
(** Encoders/decoders for big endian - network byte order *)
module LE : ENCDEC
(** Encoders/decoders for little endian *)
module HO : ENCDEC
(** Encoders/decoders for host byte order - which is either little
endian or big endian, depending on the CPU (or CPU mode)
*)
|