Raise the Fatal_error exception with the given string.

Format the arguments according to the given format string and raise Fatal_error with the resulting string.

try_finally work ~always ~exceptionally is designed to run code in work that may fail with an exception, and has two kind of cleanup routines: always, that must be run after any execution of the function (typically, freeing system resources), and exceptionally, that should be run only if work or always failed with an exception (typically, undoing user-visible state changes that would only make sense if the function completes correctly). For example:

  let objfile = outputprefix ^ ".cmo" in
  let oc = open_out_bin objfile in
  Misc.try_finally
    (fun () ->
       bytecode
       ++ Timings.(accumulate_time (Generate sourcefile))
           (Emitcode.to_file oc modulename objfile);
       Warnings.check_fatal ())
    ~always:(fun () -> close_out oc)
    ~exceptionally:(fun _exn -> remove_file objfile);

If exceptionally fail with an exception, it is propagated as usual.

If always or exceptionally use exceptions internally for control-flow but do not raise, then try_finally is careful to preserve any exception backtrace coming from work or always for easier debugging.

reraise_preserving_backtrace e f is (f (); raise e) except that the current backtrace is preserved, even if f uses exceptions internally.

map_end f l t is map f l @ t, just more efficient.

map_end f l t is map f (rev l) @ t, just more efficient.

Like List.map, with guaranteed left-to-right evaluation order

Same as List.for_all but for a binary predicate. In addition, this for_all2 never fails: given two lists with different lengths, it returns false.

replicate_list elem n is the list with n elements all identical to elem.

list_remove x l returns a copy of l with the first element equal to x removed.

Return the last element and the other elements of the given list.

Detects a repeated label - for use with labeled tuples.

Create a hashtable with the given initial size and fills it with the given bindings.

Search a file in a list of directories.

Search a relative file in a list of directories.

Normalize file name Foo.ml to foo.ml, using NFC and case-folding. Return Error if the input is not a valid utf-8 byte sequence

Same as find_in_path_rel , but search also for normalized unit filename, i.e. if name is Foo.ml, allow /path/Foo.ml and /path/foo.ml to match.

Delete the given file if it exists and is a regular file. Does nothing for other kinds of files. Never raises an error.

expand_directory alt file eventually expands a + at the beginning of file into alt (an alternate root directory)

split_path_contents ?sep s interprets s as the value of a "PATH"-like variable and returns the corresponding list of directories. s is split using the platform-specific delimiter, or ~sep if it is passed.

Returns the empty list if s is empty.

copy_file ic oc reads the contents of file ic and copies them to oc. It stops when encountering EOF on ic.

copy_file_chunk ic oc n reads n bytes from ic and copies them to oc. It raises End_of_file when encountering EOF on ic.

string_of_file ic reads the contents of file ic and copies them to a string. It stops when encountering EOF on ic.

Produce output in temporary file, then rename it (as atomically as possible) to the desired output file name. output_to_file_via_temporary filename fn opens a temporary file which is passed to fn (name + output channel). When fn returns, the channel is closed and the temporary file is renamed to filename.

Open the given filename for writing (in binary mode), pass the out_channel to the given function, then close the channel. If the function raises an exception then filename will be removed.

concat_null_terminated [x1;x2; ... xn] is x1 ^ "\000" ^ x2 ^ "\000" ^ ... ^ xn ^ "\000"

split_null_terminated s is similar String.split_on_char '\000' but ignores the trailing separator, if any

Return the given file name without its extensions. The extensions is the longest suffix starting with a period and not including a directory separator, .xyz.uvw for instance.

Return the given name if it does not contain an extension.

log2 n returns s such that n = 1 lsl s if n is a power of 2

align n a rounds n upwards to a multiple of a (a power of 2).

no_overflow_add n1 n2 returns true if the computation of n1 + n2 does not overflow.

no_overflow_sub n1 n2 returns true if the computation of n1 - n2 does not overflow.

no_overflow_mul n1 n2 returns true if the computation of n1 * n2 does not overflow.

no_overflow_lsl n k returns true if the computation of n lsl k does not overflow.

find_first_mono p takes an integer predicate p : int -> bool that we assume: 1. is monotonic on natural numbers: if a <= b then p a implies p b, 2. is satisfied for some natural numbers in range 0; max_int (this is equivalent to: p max_int = true).

find_first_mono p is the smallest natural number N that satisfies p, computed in O(log(N)) calls to p.

Our implementation supports two cases where the preconditions on p are not respected:

• If p is always false, we silently return max_int instead of looping or crashing.
• If p is non-monotonic but eventually true, we return some satisfying value.

search_substring pat str start returns the position of the first occurrence of string pat in string str. Search starts at offset start in str. Raise Not_found if pat does not occur.

replace_substring ~before ~after str replaces all occurrences of before with after in str and returns the resulting string.

rev_split_words s splits s in blank-separated words, and returns the list of words in reverse order.

String.cut_at s c returns a pair containing the sub-string before the first occurrence of c in s, and the sub-string after the first occurrence of c in s. let (before, after) = String.cut_at s c in before ^ String.make 1 c ^ after is the identity if s contains c.

Raise Not_found if the character does not appear in the string

• since 4.01

ordinal_suffix n is the appropriate suffix to append to the numeral n as an ordinal number: 1 -> "st", 2 -> "nd", 3 -> "rd", 4 -> "th", and so on. Handles larger numbers (e.g., 42 -> "nd") and the numbers 11--13 (which all get "th") correctly.

normalise_eol s returns a fresh copy of s with any '\r' characters removed. Intended for pre-processing text which will subsequently be printed on a channel which performs EOL transformations (i.e. Windows)

delete_eol_spaces s returns a fresh copy of s with any end of line spaces removed. Intended to normalize the output of the toplevel for tests.

protect_refs l f temporarily sets r to v for each R (r, v) in l while executing f. The previous contents of the references is restored even if f raises an exception, without altering the exception backtrace.

get_ref lr returns the content of the list reference lr and reset its content to the empty list.

set_or_ignore f opt x sets opt to f x if it returns Some _, or leaves it unmodified if it returns None.

edit_distance a b cutoff computes the edit distance between strings a and b. To help efficiency, it uses a cutoff: if the distance d is smaller than cutoff, it returns Some d, else None.

The distance algorithm currently used is Damerau-Levenshtein: it computes the number of insertion, deletion, substitution of letters, or swapping of adjacent letters to go from one word to the other. The particular algorithm may change in the future.

spellcheck env name takes a list of names env that exist in the current environment and an erroneous name, and returns a list of suggestions taken from env, that are close enough to name that it may be a typo for one of them.

aligned_hint main hint vertically aligns a main message and a hint message. The vertical alignment is controlled by the use of @{<ralign> ... @} boxes: the start of one box, in either the hint or the main message, will be shifted on the left to ensure that the end of the two boxes are vertically aligned, taking in account a pre-existing prefix before the main message. For instance,

let main, sub =
  align_hint
    ~prefix:"Error: "
    (doc_printf "@{<ralign>The value @}%a is not an instance variable"
      Style.inline_code "foobar"
    )
    (doc_printf
      "@{<ralign>Did you mean @}%a" Style.inline_code "foobaz"
    ) in
   printf "Error: %a%a" pp_doc main pp_doc sub

produces the following text:

Error:   The value "foobaz" is not an instance variable
Hint: Did you mean "foobar"?

where the main message has been shifted to the left to align "foobaz" and "foobar".

Same as align_hint ~prefix:"Error: "

aligned_hint ~prefix fmt ... hint align the potential hint with the main error message generated by the format string fmt before printing the two message.

did_you_mean ~pp choices hints that the user may have meant one of the option in choices.

Each choice is printed with the pp function, or Style.inline_code if pp=None.

print_if ppf flag fmt x prints x with fmt on ppf if b is true.

See manual section

Display the values of all compiler configuration variables from module Config, then exit the program with code 0.

Display the value of the given configuration variable, then exit the program with code 0.

Returns the map encoded in the BUILD_PATH_PREFIX_MAP environment variable.

Returns the list of --debug-prefix-map flags to be passed to the assembler, built from the BUILD_PATH_PREFIX_MAP environment variable.

a typical magic number is "Caml1999I011"; it is formed of an alphanumeric prefix, here Caml1990I, followed by a version, here 011. The prefix identifies the kind of the versioned data: here the I indicates that it is the magic number for .cmi files.