1Stdlib.Bytes(3) OCaml library Stdlib.Bytes(3)
2
6 Stdlib.Bytes - no description
7
9 Module Stdlib.Bytes
10
12 Module Bytes
13 : (module Stdlib__Bytes)
14 val length : bytes -> int
23 Return the length (number of bytes) of the argument.
25 val get : bytes -> int -> char
29 get s n returns the byte at index n in argument s .
32 Raises Invalid_argument if n is not a valid index in s .
35 val set : bytes -> int -> char -> unit
39 set s n c modifies s in place, replacing the byte at index n with c .
42 Raises Invalid_argument if n is not a valid index in s .
45 val create : int -> bytes
49 create n returns a new byte sequence of length n . The sequence is
52 uninitialized and contains arbitrary bytes.
53 Raises Invalid_argument if n < 0 or n > Sys.max_string_length .
56 val make : int -> char -> bytes
60 make n c returns a new byte sequence of length n , filled with the byte
63 c .
64 Raises Invalid_argument if n < 0 or n > Sys.max_string_length .
67 val init : int -> (int -> char) -> bytes
71 init n f returns a fresh byte sequence of length n , with character i
74 initialized to the result of f i (in increasing index order).
75 Raises Invalid_argument if n < 0 or n > Sys.max_string_length .
78 val empty : bytes
82 A byte sequence of size 0.
84 val copy : bytes -> bytes
88 Return a new byte sequence that contains the same bytes as the argu‐
90 ment.
91 val of_string : string -> bytes
95 Return a new byte sequence that contains the same bytes as the given
97 string.
98 val to_string : bytes -> string
102 Return a new string that contains the same bytes as the given byte se‐
104 quence.
105 val sub : bytes -> int -> int -> bytes
109 sub s pos len returns a new byte sequence of length len , containing
112 the subsequence of s that starts at position pos and has length len .
113 Raises Invalid_argument if pos and len do not designate a valid range
116 of s .
117 val sub_string : bytes -> int -> int -> string
121 Same as Bytes.sub but return a string instead of a byte sequence.
123 val extend : bytes -> int -> int -> bytes
127 extend s left right returns a new byte sequence that contains the bytes
130 of s , with left uninitialized bytes prepended and right uninitialized
131 bytes appended to it. If left or right is negative, then bytes are re‐
132 moved (instead of appended) from the corresponding side of s .
133 Since 4.05.0 in BytesLabels
136 Raises Invalid_argument if the result length is negative or longer than
139 Sys.max_string_length bytes.
140 val fill : bytes -> int -> int -> char -> unit
144 fill s pos len c modifies s in place, replacing len characters with c ,
147 starting at pos .
148 Raises Invalid_argument if pos and len do not designate a valid range
151 of s .
152 val blit : bytes -> int -> bytes -> int -> int -> unit
156 blit src src_pos dst dst_pos len copies len bytes from sequence src ,
159 starting at index src_pos , to sequence dst , starting at index dst_pos
160 . It works correctly even if src and dst are the same byte sequence,
161 and the source and destination intervals overlap.
162 Raises Invalid_argument if src_pos and len do not designate a valid
165 range of src , or if dst_pos and len do not designate a valid range of
166 dst .
167 val blit_string : string -> int -> bytes -> int -> int -> unit
171 blit src src_pos dst dst_pos len copies len bytes from string src ,
174 starting at index src_pos , to byte sequence dst , starting at index
175 dst_pos .
176 Since 4.05.0 in BytesLabels
179 Raises Invalid_argument if src_pos and len do not designate a valid
182 range of src , or if dst_pos and len do not designate a valid range of
183 dst .
184 val concat : bytes -> bytes list -> bytes
188 concat sep sl concatenates the list of byte sequences sl , inserting
191 the separator byte sequence sep between each, and returns the result as
192 a new byte sequence.
193 Raises Invalid_argument if the result is longer than
196 Sys.max_string_length bytes.
197 val cat : bytes -> bytes -> bytes
201 cat s1 s2 concatenates s1 and s2 and returns the result as a new byte
204 sequence.
205 Since 4.05.0 in BytesLabels
208 Raises Invalid_argument if the result is longer than
211 Sys.max_string_length bytes.
212 val iter : (char -> unit) -> bytes -> unit
216 iter f s applies function f in turn to all the bytes of s . It is
219 equivalent to f (get s 0); f (get s 1); ...; f (get s
220 (length s - 1)); () .
221 val iteri : (int -> char -> unit) -> bytes -> unit
225 Same as Bytes.iter , but the function is applied to the index of the
227 byte as first argument and the byte itself as second argument.
228 val map : (char -> char) -> bytes -> bytes
232 map f s applies function f in turn to all the bytes of s (in increasing
235 index order) and stores the resulting bytes in a new sequence that is
236 returned as the result.
237 val mapi : (int -> char -> char) -> bytes -> bytes
241 mapi f s calls f with each character of s and its index (in increasing
244 index order) and stores the resulting bytes in a new sequence that is
245 returned as the result.
246 val fold_left : ('a -> char -> 'a) -> 'a -> bytes -> 'a
250 fold_left f x s computes f (... (f (f x (get s 0)) (get s 1)) ...) (get
253 s (n-1)) , where n is the length of s .
254 Since 4.13.0
257 val fold_right : (char -> 'a -> 'a) -> bytes -> 'a -> 'a
261 fold_right f s x computes f (get s 0) (f (get s 1) ( ... (f (get s
264 (n-1)) x) ...)) , where n is the length of s .
265 Since 4.13.0
268 val for_all : (char -> bool) -> bytes -> bool
272 for_all p s checks if all characters in s satisfy the predicate p .
275 Since 4.13.0
278 val exists : (char -> bool) -> bytes -> bool
282 exists p s checks if at least one character of s satisfies the predi‐
285 cate p .
286 Since 4.13.0
289 val trim : bytes -> bytes
293 Return a copy of the argument, without leading and trailing whitespace.
295 The bytes regarded as whitespace are the ASCII characters ' ' , '\012'
296 , '\n' , '\r' , and '\t' .
297 val escaped : bytes -> bytes
301 Return a copy of the argument, with special characters represented by
303 escape sequences, following the lexical conventions of OCaml. All
304 characters outside the ASCII printable range (32..126) are escaped, as
305 well as backslash and double-quote.
306 Raises Invalid_argument if the result is longer than
309 Sys.max_string_length bytes.
310 val index : bytes -> char -> int
314 index s c returns the index of the first occurrence of byte c in s .
317 Raises Not_found if c does not occur in s .
320 val index_opt : bytes -> char -> int option
324 index_opt s c returns the index of the first occurrence of byte c in s
327 or None if c does not occur in s .
328 Since 4.05
331 val rindex : bytes -> char -> int
335 rindex s c returns the index of the last occurrence of byte c in s .
338 Raises Not_found if c does not occur in s .
341 val rindex_opt : bytes -> char -> int option
345 rindex_opt s c returns the index of the last occurrence of byte c in s
348 or None if c does not occur in s .
349 Since 4.05
352 val index_from : bytes -> int -> char -> int
356 index_from s i c returns the index of the first occurrence of byte c in
359 s after position i . index s c is equivalent to index_from s 0 c .
360 Raises Invalid_argument if i is not a valid position in s .
363 Raises Not_found if c does not occur in s after position i .
366 val index_from_opt : bytes -> int -> char -> int option
370 index_from_opt s i c returns the index of the first occurrence of byte
373 c in s after position i or None if c does not occur in s after position
374 i . index_opt s c is equivalent to index_from_opt s 0 c .
375 Since 4.05
378 Raises Invalid_argument if i is not a valid position in s .
381 val rindex_from : bytes -> int -> char -> int
385 rindex_from s i c returns the index of the last occurrence of byte c in
388 s before position i+1 . rindex s c is equivalent to rindex_from s
389 (length s - 1) c .
390 Raises Invalid_argument if i+1 is not a valid position in s .
393 Raises Not_found if c does not occur in s before position i+1 .
396 val rindex_from_opt : bytes -> int -> char -> int option
400 rindex_from_opt s i c returns the index of the last occurrence of byte
403 c in s before position i+1 or None if c does not occur in s before po‐
404 sition i+1 . rindex_opt s c is equivalent to rindex_from s (length s -
405 1) c .
406 Since 4.05
409 Raises Invalid_argument if i+1 is not a valid position in s .
412 val contains : bytes -> char -> bool
416 contains s c tests if byte c appears in s .
419 val contains_from : bytes -> int -> char -> bool
423 contains_from s start c tests if byte c appears in s after position
426 start . contains s c is equivalent to contains_from
427 s 0 c .
428 Raises Invalid_argument if start is not a valid position in s .
431 val rcontains_from : bytes -> int -> char -> bool
435 rcontains_from s stop c tests if byte c appears in s before position
438 stop+1 .
439 Raises Invalid_argument if stop < 0 or stop+1 is not a valid position
442 in s .
443 val uppercase_ascii : bytes -> bytes
447 Return a copy of the argument, with all lowercase letters translated to
449 uppercase, using the US-ASCII character set.
450 Since 4.03.0 (4.05.0 in BytesLabels)
453 val lowercase_ascii : bytes -> bytes
457 Return a copy of the argument, with all uppercase letters translated to
459 lowercase, using the US-ASCII character set.
460 Since 4.03.0 (4.05.0 in BytesLabels)
463 val capitalize_ascii : bytes -> bytes
467 Return a copy of the argument, with the first character set to upper‐
469 case, using the US-ASCII character set.
470 Since 4.03.0 (4.05.0 in BytesLabels)
473 val uncapitalize_ascii : bytes -> bytes
477 Return a copy of the argument, with the first character set to lower‐
479 case, using the US-ASCII character set.
480 Since 4.03.0 (4.05.0 in BytesLabels)
483 type t = bytes
486 An alias for the type of byte sequences.
489 val compare : t -> t -> int
493 The comparison function for byte sequences, with the same specification
495 as compare . Along with the type t , this function compare allows the
496 module Bytes to be passed as argument to the functors Set.Make and
497 Map.Make .
498 val equal : t -> t -> bool
502 The equality function for byte sequences.
504 Since 4.03.0 (4.05.0 in BytesLabels)
507 val starts_with : prefix:bytes -> bytes -> bool
511 starts_with ~prefix s is true if and only if s starts with prefix .
514 Since 4.13.0
517 val ends_with : suffix:bytes -> bytes -> bool
521 ends_with ~suffix s is true if and only if s ends with suffix .
524 Since 4.13.0
527 Unsafe conversions (for advanced users)
532 This section describes unsafe, low-level conversion functions between
533 bytes and string . They do not copy the internal data; used improperly,
534 they can break the immutability invariant on strings provided by the
535 -safe-string option. They are available for expert library authors, but
536 for most purposes you should use the always-correct Bytes.to_string and
537 Bytes.of_string instead.
538 val unsafe_to_string : bytes -> string
540 Unsafely convert a byte sequence into a string.
542 To reason about the use of unsafe_to_string , it is convenient to con‐
544 sider an "ownership" discipline. A piece of code that manipulates some
545 data "owns" it; there are several disjoint ownership modes, including:
546 -Unique ownership: the data may be accessed and mutated
548 -Shared ownership: the data has several owners, that may only access
550 it, not mutate it.
551 Unique ownership is linear: passing the data to another piece of code
553 means giving up ownership (we cannot write the data again). A unique
554 owner may decide to make the data shared (giving up mutation rights on
555 it), but shared data may not become uniquely-owned again.
556 unsafe_to_string s can only be used when the caller owns the byte se‐
559 quence s -- either uniquely or as shared immutable data. The caller
560 gives up ownership of s , and gains ownership of the returned string.
561 There are two valid use-cases that respect this ownership discipline:
563 1. Creating a string by initializing and mutating a byte sequence that
565 is never changed after initialization is performed.
566 let string_init len f : string =
569 let s = Bytes.create len in
570 for i = 0 to len - 1 do Bytes.set s i (f i) done;
571 Bytes.unsafe_to_string s
572 This function is safe because the byte sequence s will never be ac‐
575 cessed or mutated after unsafe_to_string is called. The string_init
576 code gives up ownership of s , and returns the ownership of the result‐
577 ing string to its caller.
578 Note that it would be unsafe if s was passed as an additional parameter
580 to the function f as it could escape this way and be mutated in the fu‐
581 ture -- string_init would give up ownership of s to pass it to f , and
582 could not call unsafe_to_string safely.
583 We have provided the String.init , String.map and String.mapi functions
585 to cover most cases of building new strings. You should prefer those
586 over to_string or unsafe_to_string whenever applicable.
587 2. Temporarily giving ownership of a byte sequence to a function that
589 expects a uniquely owned string and returns ownership back, so that we
590 can mutate the sequence again after the call ended.
591 let bytes_length (s : bytes) =
594 String.length (Bytes.unsafe_to_string s)
595 In this use-case, we do not promise that s will never be mutated after
598 the call to bytes_length s . The String.length function temporarily
599 borrows unique ownership of the byte sequence (and sees it as a string
600 ), but returns this ownership back to the caller, which may assume that
601 s is still a valid byte sequence after the call. Note that this is only
602 correct because we know that String.length does not capture its argu‐
603 ment -- it could escape by a side-channel such as a memoization combi‐
604 nator.
605 The caller may not mutate s while the string is borrowed (it has tempo‐
607 rarily given up ownership). This affects concurrent programs, but also
608 higher-order functions: if String.length returned a closure to be
609 called later, s should not be mutated until this closure is fully ap‐
610 plied and returns ownership.
611 val unsafe_of_string : string -> bytes
615 Unsafely convert a shared string to a byte sequence that should not be
617 mutated.
618 The same ownership discipline that makes unsafe_to_string correct ap‐
620 plies to unsafe_of_string : you may use it if you were the owner of the
621 string value, and you will own the return bytes in the same mode.
622 In practice, unique ownership of string values is extremely difficult
624 to reason about correctly. You should always assume strings are shared,
625 never uniquely owned.
626 For example, string literals are implicitly shared by the compiler, so
628 you never uniquely own them.
629 let incorrect = Bytes.unsafe_of_string "hello"
632 let s = Bytes.of_string "hello"
633 The first declaration is incorrect, because the string literal "hello"
636 could be shared by the compiler with other parts of the program, and
637 mutating incorrect is a bug. You must always use the second version,
638 which performs a copy and is thus correct.
639 Assuming unique ownership of strings that are not string literals, but
641 are (partly) built from string literals, is also incorrect. For exam‐
642 ple, mutating unsafe_of_string ("foo" ^ s) could mutate the shared
643 string "foo" -- assuming a rope-like representation of strings. More
644 generally, functions operating on strings will assume shared ownership,
645 they do not preserve unique ownership. It is thus incorrect to assume
646 unique ownership of the result of unsafe_of_string .
647 The only case we have reasonable confidence is safe is if the produced
649 bytes is shared -- used as an immutable byte sequence. This is possibly
650 useful for incremental migration of low-level programs that manipulate
651 immutable sequences of bytes (for example Marshal.from_bytes ) and pre‐
652 viously used the string type for this purpose.
653 val split_on_char : char -> bytes -> bytes list
657 split_on_char sep s returns the list of all (possibly empty) subse‐
660 quences of s that are delimited by the sep character.
661 The function's output is specified by the following invariants:
663 -The list is not empty.
666 -Concatenating its elements using sep as a separator returns a byte se‐
668 quence equal to the input ( Bytes.concat (Bytes.make 1 sep)
669 (Bytes.split_on_char sep s) = s ).
670 -No byte sequence in the result contains the sep character.
672 Since 4.13.0
676 Iterators
681 val to_seq : t -> char Seq.t
682 Iterate on the string, in increasing index order. Modifications of the
684 string during iteration will be reflected in the sequence.
685 Since 4.07
688 val to_seqi : t -> (int * char) Seq.t
692 Iterate on the string, in increasing order, yielding indices along
694 chars
695 Since 4.07
698 val of_seq : char Seq.t -> t
702 Create a string from the generator
704 Since 4.07
707 UTF codecs and validations
712 UTF-8
713 val get_utf_8_uchar : t -> int -> Uchar.utf_decode
714 get_utf_8_uchar b i decodes an UTF-8 character at index i in b .
717 val set_utf_8_uchar : t -> int -> Uchar.t -> int
721 set_utf_8_uchar b i u UTF-8 encodes u at index i in b and returns the
724 number of bytes n that were written starting at i . If n is 0 there was
725 not enough space to encode u at i and b was left untouched. Otherwise a
726 new character can be encoded at i + n .
727 val is_valid_utf_8 : t -> bool
731 is_valid_utf_8 b is true if and only if b contains valid UTF-8 data.
734 UTF-16BE
739 val get_utf_16be_uchar : t -> int -> Uchar.utf_decode
740 get_utf_16be_uchar b i decodes an UTF-16BE character at index i in b .
743 val set_utf_16be_uchar : t -> int -> Uchar.t -> int
747 set_utf_16be_uchar b i u UTF-16BE encodes u at index i in b and returns
750 the number of bytes n that were written starting at i . If n is 0 there
751 was not enough space to encode u at i and b was left untouched. Other‐
752 wise a new character can be encoded at i + n .
753 val is_valid_utf_16be : t -> bool
757 is_valid_utf_16be b is true if and only if b contains valid UTF-16BE
760 data.
761 UTF-16LE
766 val get_utf_16le_uchar : t -> int -> Uchar.utf_decode
767 get_utf_16le_uchar b i decodes an UTF-16LE character at index i in b .
770 val set_utf_16le_uchar : t -> int -> Uchar.t -> int
774 set_utf_16le_uchar b i u UTF-16LE encodes u at index i in b and returns
777 the number of bytes n that were written starting at i . If n is 0 there
778 was not enough space to encode u at i and b was left untouched. Other‐
779 wise a new character can be encoded at i + n .
780 val is_valid_utf_16le : t -> bool
784 is_valid_utf_16le b is true if and only if b contains valid UTF-16LE
787 data.
788 Binary encoding/decoding of integers
793 The functions in this section binary encode and decode integers to and
794 from byte sequences.
795 All following functions raise Invalid_argument if the space needed at
797 index i to decode or encode the integer is not available.
798 Little-endian (resp. big-endian) encoding means that least (resp. most)
800 significant bytes are stored first. Big-endian is also known as net‐
801 work byte order. Native-endian encoding is either little-endian or
802 big-endian depending on Sys.big_endian .
803 32-bit and 64-bit integers are represented by the int32 and int64
805 types, which can be interpreted either as signed or unsigned numbers.
806 8-bit and 16-bit integers are represented by the int type, which has
808 more bits than the binary encoding. These extra bits are handled as
809 follows:
810 -Functions that decode signed (resp. unsigned) 8-bit or 16-bit integers
812 represented by int values sign-extend (resp. zero-extend) their result.
813 -Functions that encode 8-bit or 16-bit integers represented by int val‐
815 ues truncate their input to their least significant bytes.
816 val get_uint8 : bytes -> int -> int
819 get_uint8 b i is b 's unsigned 8-bit integer starting at byte index i .
822 Since 4.08
825 val get_int8 : bytes -> int -> int
829 get_int8 b i is b 's signed 8-bit integer starting at byte index i .
832 Since 4.08
835 val get_uint16_ne : bytes -> int -> int
839 get_uint16_ne b i is b 's native-endian unsigned 16-bit integer start‐
842 ing at byte index i .
843 Since 4.08
846 val get_uint16_be : bytes -> int -> int
850 get_uint16_be b i is b 's big-endian unsigned 16-bit integer starting
853 at byte index i .
854 Since 4.08
857 val get_uint16_le : bytes -> int -> int
861 get_uint16_le b i is b 's little-endian unsigned 16-bit integer start‐
864 ing at byte index i .
865 Since 4.08
868 val get_int16_ne : bytes -> int -> int
872 get_int16_ne b i is b 's native-endian signed 16-bit integer starting
875 at byte index i .
876 Since 4.08
879 val get_int16_be : bytes -> int -> int
883 get_int16_be b i is b 's big-endian signed 16-bit integer starting at
886 byte index i .
887 Since 4.08
890 val get_int16_le : bytes -> int -> int
894 get_int16_le b i is b 's little-endian signed 16-bit integer starting
897 at byte index i .
898 Since 4.08
901 val get_int32_ne : bytes -> int -> int32
905 get_int32_ne b i is b 's native-endian 32-bit integer starting at byte
908 index i .
909 Since 4.08
912 val get_int32_be : bytes -> int -> int32
916 get_int32_be b i is b 's big-endian 32-bit integer starting at byte in‐
919 dex i .
920 Since 4.08
923 val get_int32_le : bytes -> int -> int32
927 get_int32_le b i is b 's little-endian 32-bit integer starting at byte
930 index i .
931 Since 4.08
934 val get_int64_ne : bytes -> int -> int64
938 get_int64_ne b i is b 's native-endian 64-bit integer starting at byte
941 index i .
942 Since 4.08
945 val get_int64_be : bytes -> int -> int64
949 get_int64_be b i is b 's big-endian 64-bit integer starting at byte in‐
952 dex i .
953 Since 4.08
956 val get_int64_le : bytes -> int -> int64
960 get_int64_le b i is b 's little-endian 64-bit integer starting at byte
963 index i .
964 Since 4.08
967 val set_uint8 : bytes -> int -> int -> unit
971 set_uint8 b i v sets b 's unsigned 8-bit integer starting at byte index
974 i to v .
975 Since 4.08
978 val set_int8 : bytes -> int -> int -> unit
982 set_int8 b i v sets b 's signed 8-bit integer starting at byte index i
985 to v .
986 Since 4.08
989 val set_uint16_ne : bytes -> int -> int -> unit
993 set_uint16_ne b i v sets b 's native-endian unsigned 16-bit integer
996 starting at byte index i to v .
997 Since 4.08
1000 val set_uint16_be : bytes -> int -> int -> unit
1004 set_uint16_be b i v sets b 's big-endian unsigned 16-bit integer start‐
1007 ing at byte index i to v .
1008 Since 4.08
1011 val set_uint16_le : bytes -> int -> int -> unit
1015 set_uint16_le b i v sets b 's little-endian unsigned 16-bit integer
1018 starting at byte index i to v .
1019 Since 4.08
1022 val set_int16_ne : bytes -> int -> int -> unit
1026 set_int16_ne b i v sets b 's native-endian signed 16-bit integer start‐
1029 ing at byte index i to v .
1030 Since 4.08
1033 val set_int16_be : bytes -> int -> int -> unit
1037 set_int16_be b i v sets b 's big-endian signed 16-bit integer starting
1040 at byte index i to v .
1041 Since 4.08
1044 val set_int16_le : bytes -> int -> int -> unit
1048 set_int16_le b i v sets b 's little-endian signed 16-bit integer start‐
1051 ing at byte index i to v .
1052 Since 4.08
1055 val set_int32_ne : bytes -> int -> int32 -> unit
1059 set_int32_ne b i v sets b 's native-endian 32-bit integer starting at
1062 byte index i to v .
1063 Since 4.08
1066 val set_int32_be : bytes -> int -> int32 -> unit
1070 set_int32_be b i v sets b 's big-endian 32-bit integer starting at byte
1073 index i to v .
1074 Since 4.08
1077 val set_int32_le : bytes -> int -> int32 -> unit
1081 set_int32_le b i v sets b 's little-endian 32-bit integer starting at
1084 byte index i to v .
1085 Since 4.08
1088 val set_int64_ne : bytes -> int -> int64 -> unit
1092 set_int64_ne b i v sets b 's native-endian 64-bit integer starting at
1095 byte index i to v .
1096 Since 4.08
1099 val set_int64_be : bytes -> int -> int64 -> unit
1103 set_int64_be b i v sets b 's big-endian 64-bit integer starting at byte
1106 index i to v .
1107 Since 4.08
1110 val set_int64_le : bytes -> int -> int64 -> unit
1114 set_int64_le b i v sets b 's little-endian 64-bit integer starting at
1117 byte index i to v .
1118 Since 4.08
1121 Byte sequences and concurrency safety
1126 Care must be taken when concurrently accessing byte sequences from mul‐
1127 tiple domains: accessing a byte sequence will never crash a program,
1128 but unsynchronized accesses might yield surprising (non-sequen‐
1129 tially-consistent) results.
1130 Atomicity
1133 Every byte sequence operation that accesses more than one byte is not
1134 atomic. This includes iteration and scanning.
1135 For example, consider the following program:
1137 let size = 100_000_000
1138 let b = Bytes.make size ' '
1139 let update b f () =
1140 Bytes.iteri (fun i x -> Bytes.set b i (Char.chr (f (Char.code x)))) b
1141 let d1 = Domain.spawn (update b (fun x -> x + 1))
1142 let d2 = Domain.spawn (update b (fun x -> 2 * x + 1))
1143 let () = Domain.join d1; Domain.join d2
1144 the bytes sequence b may contain a non-deterministic mixture of '!' ,
1146 'A' , 'B' , and 'C' values.
1147 After executing this code, each byte of the sequence b is either '!' ,
1149 'A' , 'B' , or 'C' . If atomicity is required, then the user must im‐
1150 plement their own synchronization (for example, using Mutex.t ).
1151 Data races
1154 If two domains only access disjoint parts of a byte sequence, then the
1155 observed behaviour is the equivalent to some sequential interleaving of
1156 the operations from the two domains.
1157 A data race is said to occur when two domains access the same byte
1159 without synchronization and at least one of the accesses is a write.
1160 In the absence of data races, the observed behaviour is equivalent to
1161 some sequential interleaving of the operations from different domains.
1162 Whenever possible, data races should be avoided by using synchroniza‐
1164 tion to mediate the accesses to the elements of the sequence.
1165 Indeed, in the presence of data races, programs will not crash but the
1167 observed behaviour may not be equivalent to any sequential interleaving
1168 of operations from different domains. Nevertheless, even in the pres‐
1169 ence of data races, a read operation will return the value of some
1170 prior write to that location.
1171 Mixed-size accesses
1174 Another subtle point is that if a data race involves mixed-size writes
1175 and reads to the same location, the order in which those writes and
1176 reads are observed by domains is not specified. For instance, the fol‐
1177 lowing code write sequentially a 32-bit integer and a char to the same
1178 index
1179 let b = Bytes.make 10 '\000'
1180 let d1 = Domain.spawn (fun () -> Bytes.set_int32_ne b 0 100; b.[0] <- 'd' )
1181 In this situation, a domain that observes the write of 'd' to b. 0 is
1184 not guaranteed to also observe the write to indices 1 , 2 , or 3 .
1185OCamldoc 2023-07-20 Stdlib.Bytes(3)