stringifor_string_t.F90 Source File


This file depends on

sourcefile~~stringifor_string_t.f90~2~~EfferentGraph sourcefile~stringifor_string_t.f90~2 stringifor_string_t.F90 sourcefile~befor64.f90 befor64.F90 sourcefile~stringifor_string_t.f90~2->sourcefile~befor64.f90 sourcefile~face.f90 face.F90 sourcefile~stringifor_string_t.f90~2->sourcefile~face.f90 sourcefile~penf.f90 penf.F90 sourcefile~stringifor_string_t.f90~2->sourcefile~penf.f90 sourcefile~befor64.f90->sourcefile~penf.f90 sourcefile~befor64_pack_data_m.f90 befor64_pack_data_m.F90 sourcefile~befor64.f90->sourcefile~befor64_pack_data_m.f90 sourcefile~penf_global_parameters_variables.f90 penf_global_parameters_variables.F90 sourcefile~penf.f90->sourcefile~penf_global_parameters_variables.f90 sourcefile~penf_b_size.f90 penf_b_size.F90 sourcefile~penf.f90->sourcefile~penf_b_size.f90 sourcefile~penf_stringify.f90 penf_stringify.F90 sourcefile~penf.f90->sourcefile~penf_stringify.f90 sourcefile~befor64_pack_data_m.f90->sourcefile~penf.f90 sourcefile~penf_b_size.f90->sourcefile~penf_global_parameters_variables.f90 sourcefile~penf_stringify.f90->sourcefile~penf_global_parameters_variables.f90 sourcefile~penf_stringify.f90->sourcefile~penf_b_size.f90

Contents


Source Code

!< StringiFor, definition of `string` type.
module stringifor_string_t
!< StringiFor, definition of `string` type.
use, intrinsic :: iso_fortran_env, only : iostat_eor
use befor64, only : b64_decode, b64_encode
use face, only : colorize
use penf, only : I1P, I2P, I4P, I8P, R4P, R8P, R16P, str

implicit none
private
save
! expose StingiFor overloaded builtins and operators
! public :: adjustl, adjustr, count, index, len, len_trim, repeat, scan, trim, verify
public :: adjustl, adjustr, count, index, len_trim, repeat, scan, trim, verify
! expose StingiFor objects
public :: CK
public :: glob
public :: strjoin
public :: string

integer, parameter :: CK = selected_char_kind('DEFAULT') !< Default character kind.

type :: string
  !< OOP designed string class.
  character(kind=CK, len=:), allocatable :: raw !< Raw data.
  contains
    ! public methods
    ! builtins replacements
    procedure, pass(self) :: adjustl  => sadjustl                 !< Adjustl replacement.
    procedure, pass(self) :: adjustr  => sadjustr                 !< Adjustr replacement.
    procedure, pass(self) :: count    => scount                   !< Count replacement.
    generic               :: index    => sindex_string_string, &
                                         sindex_string_character  !< Index replacement.
    procedure, pass(self) :: len      => slen                     !< Len replacement.
    procedure, pass(self) :: len_trim => slen_trim                !< Len_trim replacement.
    generic               :: repeat   => srepeat_string_string, &
                                         srepeat_character_string !< Repeat replacement.
    generic               :: scan     => sscan_string_string,    &
                                         sscan_string_character   !< Scan replacement.
    procedure, pass(self) :: trim     => strim                    !< Trim replacement.
    generic               :: verify   => sverify_string_string, &
                                         sverify_string_character !< Verify replacement.
    ! auxiliary methods
    procedure, pass(self) :: basedir          !< Return the base directory name of a string containing a file name.
    procedure, pass(self) :: basename         !< Return the base file name of a string containing a file name.
    procedure, pass(self) :: camelcase        !< Return a string with all words capitalized without spaces.
    procedure, pass(self) :: capitalize       !< Return a string with its first character capitalized and the rest lowercased.
    procedure, pass(self) :: chars            !< Return the raw characters data.
    generic               :: colorize => &
                             colorize_str     !< Colorize and stylize strings.
    procedure, pass(self) :: decode           !< Decode string.
    procedure, pass(self) :: encode           !< Encode string.
    procedure, pass(self) :: escape           !< Escape backslashes (or custom escape character).
    procedure, pass(self) :: extension        !< Return the extension of a string containing a file name.
    procedure, pass(self) :: fill             !< Pad string on the left (or right) with zeros (or other char) to fill width.
    procedure, pass(self) :: free             !< Free dynamic memory.
    generic               :: glob =>         &
                             glob_character, &
                             glob_string      !< Glob search, finds all the pathnames matching a given pattern.
    generic               :: insert =>      &
                             insert_string, &
                             insert_character !< Insert substring into string at a specified position.
    generic               :: join =>       &
                             join_strings, &
                             join_characters  !< Return a string that is a join of an array of strings or characters.
    generic               :: strjoin =>   &
                             strjoin_strings, &
                             strjoin_characters, &
                             strjoin_strings_array, &
                             strjoin_characters_array  !< Return a string that is a join of an array of strings or characters;
                                                       !< Return join 1D string array of an 2D array of strings or characters in columns or rows.
    procedure, pass(self) :: lower            !< Return a string with all lowercase characters.
    procedure, pass(self) :: partition        !< Split string at separator and return the 3 parts (before, the separator and after).
    procedure, pass(self) :: read_file        !< Read a file a single string stream.
    procedure, pass(self) :: read_line        !< Read line (record) from a connected unit.
    procedure, pass(self) :: read_lines       !< Read (all) lines (records) from a connected unit as a single ascii stream.
    procedure, pass(self) :: replace          !< Return a string with all occurrences of substring old replaced by new.
    procedure, pass(self) :: reverse          !< Return a reversed string.
    procedure, pass(self) :: search           !< Search for *tagged* record into string.
    procedure, pass(self) :: slice            !< Return the raw characters data sliced.
    procedure, pass(self) :: snakecase        !< Return a string with all words lowercase separated by "_".
    procedure, pass(self) :: split            !< Return a list of substring in the string, using sep as the delimiter string.
    procedure, pass(self) :: split_chunked    !< Return a list of substring in the string, using sep as the delimiter string.
    procedure, pass(self) :: startcase        !< Return a string with all words capitalized, e.g. title case.
    procedure, pass(self) :: strip            !< Return a string with the leading and trailing characters removed.
    procedure, pass(self) :: swapcase         !< Return a string with uppercase chars converted to lowercase and vice versa.
    procedure, pass(self) :: tempname         !< Return a safe temporary name suitable for temporary file or directories.
    generic               :: to_number =>   &
                             to_integer_I1P,&
#ifndef _NVF
                             to_integer_I2P,&
#endif
                             to_integer_I4P,&
                             to_integer_I8P,&
#if defined _R16P
                             to_real_R16P,  &
#endif
                             to_real_R8P,   &
                             to_real_R4P      !< Cast string to number.
    procedure, pass(self) :: unescape         !< Unescape double backslashes (or custom escaped character).
    procedure, pass(self) :: unique           !< Reduce to one (unique) multiple occurrences of a substring into a string.
    procedure, pass(self) :: upper            !< Return a string with all uppercase characters.
    procedure, pass(self) :: write_file       !< Write a single string stream into file.
    procedure, pass(self) :: write_line       !< Write line (record) to a connected unit.
    procedure, pass(self) :: write_lines      !< Write lines (records) to a connected unit.
    ! inquire methods
    procedure, pass(self) :: end_with     !< Return true if a string ends with a specified suffix.
    procedure, pass(self) :: is_allocated !< Return true if the string is allocated.
    procedure, pass(self) :: is_digit     !< Return true if all characters in the string are digits.
    procedure, pass(self) :: is_integer   !< Return true if the string contains an integer.
    procedure, pass(self) :: is_lower     !< Return true if all characters in the string are lowercase.
    procedure, pass(self) :: is_number    !< Return true if the string contains a number (real or integer).
    procedure, pass(self) :: is_real      !< Return true if the string contains an real.
    procedure, pass(self) :: is_upper     !< Return true if all characters in the string are uppercase.
    procedure, pass(self) :: start_with   !< Return true if a string starts with a specified prefix.
    ! operators
    generic :: assignment(=) => string_assign_string,      &
                                string_assign_character,   &
                                string_assign_integer_I1P, &
                                string_assign_integer_I2P, &
                                string_assign_integer_I4P, &
                                string_assign_integer_I8P, &
#if defined _R16P
                                string_assign_real_R16P,   &
#endif
                                string_assign_real_R8P,    &
                                string_assign_real_R4P              !< Assignment operator overloading.
    generic :: operator(//) => string_concat_string,    &
                               string_concat_character, &
                               character_concat_string              !< Concatenation operator overloading.
    generic :: operator(.cat.) => string_concat_string_string,    &
                                  string_concat_character_string, &
                                  character_concat_string_string    !< Concatenation operator (string output) overloading.
    generic :: operator(==) => string_eq_string,    &
                               string_eq_character, &
                               character_eq_string                  !< Equal operator overloading.
    generic :: operator(/=) => string_ne_string,    &
                               string_ne_character, &
                               character_ne_string                  !< Not equal operator overloading.
    generic :: operator(<) => string_lt_string,    &
                              string_lt_character, &
                              character_lt_string                   !< Lower than operator overloading.
    generic :: operator(<=) => string_le_string,    &
                               string_le_character, &
                               character_le_string                  !< Lower equal than operator overloading.
    generic :: operator(>=) => string_ge_string,    &
                               string_ge_character, &
                               character_ge_string                  !< Greater equal than operator overloading.
    generic :: operator(>) => string_gt_string,    &
                              string_gt_character, &
                              character_gt_string                   !< Greater than operator overloading.
    ! IO
    generic :: read(formatted) => read_formatted       !< Formatted input.
    generic :: write(formatted) => write_formatted     !< Formatted output.
    generic :: read(unformatted) => read_unformatted   !< Unformatted input.
    generic :: write(unformatted) => write_unformatted !< Unformatted output.
    ! private methods
    ! builtins replacements
    procedure, private, pass(self) :: sindex_string_string     !< Index replacement.
    procedure, private, pass(self) :: sindex_string_character  !< Index replacement.
    procedure, private, pass(self) :: srepeat_string_string    !< Repeat replacement.
    procedure, private, nopass     :: srepeat_character_string !< Repeat replacement.
    procedure, private, pass(self) :: sscan_string_string      !< Scan replacement.
    procedure, private, pass(self) :: sscan_string_character   !< Scan replacement.
    procedure, private, pass(self) :: sverify_string_string    !< Verify replacement.
    procedure, private, pass(self) :: sverify_string_character !< Verify replacement.
    ! auxiliary methods
    procedure, private, pass(self) :: colorize_str     !< Colorize and stylize strings.
    procedure, private, pass(self) :: glob_character   !< Glob search (character output).
    procedure, private, pass(self) :: glob_string      !< Glob search (string output).
    procedure, private, pass(self) :: insert_string    !< Insert substring into string at a specified position.
    procedure, private, pass(self) :: insert_character !< Insert substring into string at a specified position.
    procedure, private, pass(self) :: join_strings     !< Return join string of an array of strings.
    procedure, private, pass(self) :: join_characters  !< Return join string of an array of characters.
    procedure, private, nopass ::     strjoin_strings  !< Return join string of an array of strings.
    procedure, private, nopass ::     strjoin_characters        !< Return join string of an array of strings.
    procedure, private, nopass ::     strjoin_strings_array     !< Return join 1D string array of an 2D array of strings in columns or rows.
    procedure, private, nopass ::     strjoin_characters_array  !< Return join 1D string array of an 2D array of characters in columns or rows.
    procedure, private, pass(self) :: to_integer_I1P   !< Cast string to integer.
#ifndef _NVF
    procedure, private, pass(self) :: to_integer_I2P   !< Cast string to integer.
#endif
    procedure, private, pass(self) :: to_integer_I4P   !< Cast string to integer.
    procedure, private, pass(self) :: to_integer_I8P   !< Cast string to integer.
    procedure, private, pass(self) :: to_real_R4P      !< Cast string to real.
    procedure, private, pass(self) :: to_real_R8P      !< Cast string to real.
    procedure, private, pass(self) :: to_real_R16P     !< Cast string to real.
    ! assignments
    procedure, private, pass(lhs) :: string_assign_string      !< Assignment operator from string input.
    procedure, private, pass(lhs) :: string_assign_character   !< Assignment operator from character input.
    procedure, private, pass(lhs) :: string_assign_integer_I1P !< Assignment operator from integer input.
    procedure, private, pass(lhs) :: string_assign_integer_I2P !< Assignment operator from integer input.
    procedure, private, pass(lhs) :: string_assign_integer_I4P !< Assignment operator from integer input.
    procedure, private, pass(lhs) :: string_assign_integer_I8P !< Assignment operator from integer input.
    procedure, private, pass(lhs) :: string_assign_real_R4P    !< Assignment operator from real input.
    procedure, private, pass(lhs) :: string_assign_real_R8P    !< Assignment operator from real input.
    procedure, private, pass(lhs) :: string_assign_real_R16P   !< Assignment operator from real input.
    ! concatenation operators
    procedure, private, pass(lhs) :: string_concat_string           !< Concatenation with string.
    procedure, private, pass(lhs) :: string_concat_character        !< Concatenation with character.
    procedure, private, pass(rhs) :: character_concat_string        !< Concatenation with character (inverted).
    procedure, private, pass(lhs) :: string_concat_string_string    !< Concatenation with string (string output).
    procedure, private, pass(lhs) :: string_concat_character_string !< Concatenation with character (string output).
    procedure, private, pass(rhs) :: character_concat_string_string !< Concatenation with character (inverted, string output).
    ! logical operators
    procedure, private, pass(lhs) :: string_eq_string    !< Equal to string logical operator.
    procedure, private, pass(lhs) :: string_eq_character !< Equal to character logical operator.
    procedure, private, pass(rhs) :: character_eq_string !< Equal to character (inverted) logical operator.
    procedure, private, pass(lhs) :: string_ne_string    !< Not equal to string logical operator.
    procedure, private, pass(lhs) :: string_ne_character !< Not equal to character logical operator.
    procedure, private, pass(rhs) :: character_ne_string !< Not equal to character (inverted) logical operator.
    procedure, private, pass(lhs) :: string_lt_string    !< Lower than to string logical operator.
    procedure, private, pass(lhs) :: string_lt_character !< Lower than to character logical operator.
    procedure, private, pass(rhs) :: character_lt_string !< Lower than to character (inverted) logical operator.
    procedure, private, pass(lhs) :: string_le_string    !< Lower equal than to string logical operator.
    procedure, private, pass(lhs) :: string_le_character !< Lower equal than to character logical operator.
    procedure, private, pass(rhs) :: character_le_string !< Lower equal than to character (inverted) logical operator.
    procedure, private, pass(lhs) :: string_ge_string    !< Greater equal than to string logical operator.
    procedure, private, pass(lhs) :: string_ge_character !< Greater equal than to character logical operator.
    procedure, private, pass(rhs) :: character_ge_string !< Greater equal than to character (inverted) logical operator.
    procedure, private, pass(lhs) :: string_gt_string    !< Greater than to string logical operator.
    procedure, private, pass(lhs) :: string_gt_character !< Greater than to character logical operator.
    procedure, private, pass(rhs) :: character_gt_string !< Greater than to character (inverted) logical operator.
    ! IO
    procedure, private, pass(dtv) :: read_formatted                !< Formatted input.
    procedure, private, pass(dtv) :: read_delimited                !< Read a delimited input.
    procedure, private, pass(dtv) :: read_undelimited              !< Read an undelimited input.
    procedure, private, pass(dtv) :: read_undelimited_listdirected !< Read an undelimited list directed input.
    procedure, private, pass(dtv) :: write_formatted               !< Formatted output.
    procedure, private, pass(dtv) :: read_unformatted              !< Unformatted input.
    procedure, private, pass(dtv) :: write_unformatted             !< Unformatted output.
    ! miscellanea
    procedure, private, pass(self) :: replace_one_occurrence !< Replace the first occurrence of substring old by new.
endtype string

! internal parameters
character(kind=CK, len=26), parameter :: UPPER_ALPHABET = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' !< Upper case alphabet.
character(kind=CK, len=26), parameter :: LOWER_ALPHABET = 'abcdefghijklmnopqrstuvwxyz' !< Lower case alphabet.
character(kind=CK, len=1),  parameter :: SPACE          = ' '                          !< Space character.
character(kind=CK, len=1),  parameter :: TAB            = achar(9)                     !< Tab character.
character(kind=CK, len=1),  parameter :: UIX_DIR_SEP    = char(47)                     !< Unix/Linux directories separator (/).
character(kind=CK, len=1),  parameter :: BACKSLASH      = char(92)                     !< Backslash character.

interface glob
  !< Overloading glob procedure.
  !<```fortran
  !< type(string)                  :: astring
  !< character(len=:), allocatable :: alist_chr(:)
  !< type(string),     allocatable :: alist_str(:)
  !< integer, parameter            :: Nf=5
  !< character(14)                 :: files(1:Nf)
  !< integer                       :: file_unit
  !< integer                       :: f
  !< integer                       :: ff
  !< logical                       :: test_passed
  !< do f=1, Nf
  !<    files(f) = astring%tempname(prefix='foo-')
  !<    open(newunit=file_unit, file=files(f))
  !<    write(file_unit, *)f
  !<    close(unit=file_unit)
  !< enddo
  !< call glob(self=astring, pattern='foo-*', list=alist_chr)
  !< call glob(self=astring, pattern='foo-*', list=alist_str)
  !< do f=1, Nf
  !<    open(newunit=file_unit, file=files(f))
  !<    close(unit=file_unit, status='delete')
  !< enddo
  !< test_passed = .false.
  !< outer_chr: do f=1, size(alist_chr, dim=1)
  !<    do ff=1, Nf
  !<       test_passed = alist_chr(f) == files(ff)
  !<       if (test_passed) cycle outer_chr
  !<    enddo
  !< enddo outer_chr
  !< if (test_passed) then
  !<    test_passed = .false.
  !<    outer_str: do f=1, size(alist_str, dim=1)
  !<       do ff=1, Nf
  !<          test_passed = alist_str(f) == files(ff)
  !<          if (test_passed) cycle outer_str
  !<       enddo
  !<    enddo outer_str
  !< endif
  !< print '(L1)', test_passed
  !<```
  !=> T <<<
  module procedure glob_character, glob_string
endinterface glob

interface strjoin
  module procedure strjoin_strings, strjoin_characters, strjoin_strings_array, strjoin_characters_array
endinterface strjoin

! builtin overloading
interface adjustl
  !< Builtin adjustl overloading.
  module procedure sadjustl_character
endinterface adjustl

interface adjustr
  !< Builtin adjustr overloading.
  module procedure sadjustr_character
endinterface adjustr

interface count
  !< Builtin count overloading.
  module procedure count_substring
endinterface

interface index
  !< Builtin index overloading.
  module procedure sindex_string_string, sindex_string_character, sindex_character_string
endinterface index

!interface len
!  !< Builtin len overloading.
!  module procedure slen
!endinterface len

interface len_trim
  !< Builtin len_trim overloading.
  module procedure slen_trim
endinterface len_trim

interface repeat
  !< Builtin repeat overloading.
  module procedure srepeat_string_string
endinterface repeat

interface scan
  !< Builtin scan overloading.
  module procedure sscan_string_string, sscan_string_character, sscan_character_string
endinterface scan

interface trim
  !< Builtin trim overloading.
  module procedure strim
endinterface trim

interface verify
  !< Builtin verify overloading.
  module procedure sverify_string_string, sverify_string_character, sverify_character_string
endinterface verify

contains
   ! public non TBP

   ! creator
   pure function string_(c)
   !< Return a string given a character input.
   !<
   !<```fortran
   !< print "(L1)", string('Hello World')//''=='Hello World'
   !<```
   !=> T <<<
   character(*), intent(in) :: c       !< Character.
   type(string)             :: string_ !< String.

   string_%raw = c
   endfunction string_

   ! builtins replacements
   pure function sadjustl_character(s) result(adjusted)
   !< Left adjust a string by removing leading spaces (character output).
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = '   Hello World!'
   !< print "(L1)", adjustl(astring)=='Hello World!   '
   !<```
   !=> T <<<
   class(string), intent(in)              :: s        !< String.
   character(kind=CK, len=:), allocatable :: adjusted !< Adjusted string.

   if (allocated(s%raw)) adjusted = adjustl(s%raw)
   endfunction sadjustl_character

   pure function sadjustr_character(s) result(adjusted)
   !< Right adjust a string by removing leading spaces (character output).
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'Hello World!   '
   !< print "(L1)", adjustr(astring)=='   Hello World!'
   !<```
   !=> T <<<
   class(string), intent(in)              :: s        !< String.
   character(kind=CK, len=:), allocatable :: adjusted !< Adjusted string.

   if (allocated(s%raw)) adjusted = adjustr(s%raw)
   endfunction sadjustr_character

   elemental function count_substring(s, substring) result(No)
   !< Count the number of occurences of a substring into a string.
   !<
   !<```fortran
   !< print "(L1)", count('hello', substring='ll')==1
   !<```
   !=> T <<<
   character(*), intent(in) :: s         !< String.
   character(*), intent(in) :: substring !< Substring.
   integer(I4P)             :: No        !< Number of occurrences.
   integer(I4P)             :: c1        !< Counters.
   integer(I4P)             :: c2        !< Counters.

   No = 0
   if (len(substring) > len(s)) return
   c1 = 1
   do
     c2 = index(string=s(c1:), substring=substring)
     if (c2==0) return
     No = No + 1
     c1 = c1 + c2 + len(substring)
   enddo
   endfunction count_substring

   elemental function sindex_character_string(s, substring, back) result(i)
   !< Return the position of the start of the first occurrence of string `substring` as a substring in `string`, counting from one.
   !< If `substring` is not present in `string`, zero is returned. If the back argument is present and true, the return value is
   !< the start of the last occurrence rather than the first.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(2)
   !< string1 = 'llo'
   !< test_passed(1) = index(s='Hello World Hello!', substring=string1)==index(string='Hello World Hello!', substring='llo')
   !< test_passed(2) = index(s='Hello World Hello!', substring=string1, back=.true.)==index(string='Hello World Hello!', &
   !<                                                                                       substring='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)           :: s         !< String.
   type(string),              intent(in)           :: substring !< Searched substring.
   logical,                   intent(in), optional :: back      !< Start of the last occurrence rather than the first.
   integer                                         :: i         !< Result of the search.

   if (allocated(substring%raw)) then
     i = index(string=s, substring=substring%raw, back=back)
   else
     i = 0
   endif
   endfunction sindex_character_string

   elemental function sscan_character_string(s, set, back) result(i)
   !< Return the leftmost (if `back` is either absent or equals false, otherwise the rightmost) character of string that is in `set`.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(2)
   !< string1 = 'llo'
   !< test_passed(1) = scan(s='Hello World Hello!', set=string1)==scan(string='Hello World Hello!', set='llo')
   !< test_passed(2) = scan(s='Hello World Hello!', set=string1, back=.true.)==scan(string='Hello World Hello!', &
   !<                                                                               set='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)           :: s    !< String.
   type(string),              intent(in)           :: set  !< Searched set.
   logical,                   intent(in), optional :: back !< Start of the last occurrence rather than the first.
   integer                                         :: i    !< Result of the search.

   if (allocated(set%raw)) then
     i = scan(string=s, set=set%raw, back=back)
   else
     i = 0
   endif
   endfunction sscan_character_string

   elemental function sverify_character_string(s, set, back) result(i)
   !< Return the leftmost (if `back` is either absent or equals false, otherwise the rightmost) character of string that is not
   !< in `set`. If all characters of `string` are found in `set`, the result is zero.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(2)
   !< string1 = 'ell'
   !< test_passed(1) = verify(s='Hello World Hello!', set=string1)==verify(string='Hello World Hello!', set='llo')
   !< test_passed(2) = verify(s='Hello World Hello!', set=string1, back=.true.)==verify(string='Hello World Hello!', set='llo', &
   !<                                                                                   back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)           :: s    !< String.
   type(string),              intent(in)           :: set  !< Searched set.
   logical,                   intent(in), optional :: back !< Start of the last occurrence rather than the first.
   integer                                         :: i    !< Result of the search.

   if (allocated(set%raw)) then
     i = verify(string=s, set=set%raw, back=back)
   else
     i = 0
   endif
   endfunction sverify_character_string

   ! public methods

   ! builtins replacements
   elemental function sadjustl(self) result(adjusted)
   !< Left adjust a string by removing leading spaces.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = '   Hello World!'
   !< print "(L1)", astring%adjustl()//''=='Hello World!   '
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   type(string)              :: adjusted !< Adjusted string.

   adjusted = self
   if (allocated(adjusted%raw)) adjusted%raw = adjustl(adjusted%raw)
   endfunction sadjustl

   elemental function sadjustr(self) result(adjusted)
   !< Right adjust a string by removing leading spaces.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'Hello World!   '
   !< print "(L1)", astring%adjustr()//''=='   Hello World!'
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   type(string)              :: adjusted !< Adjusted string.

   adjusted = self
   if (allocated(adjusted%raw)) adjusted%raw = adjustr(adjusted%raw)
   endfunction sadjustr

   elemental function scount(self, substring, ignore_isolated) result(No)
   !< Count the number of occurences of a substring into a string.
   !<
   !< @note If `ignore_isolated` is set to true the eventual "isolated" occurences are ignored: an isolated occurrences are those
   !< occurrences happening at the start of string (thus not having a left companion) or at the end of the string (thus not having a
   !< right companion).
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(4)
   !< astring = '   Hello World  !    '
   !< test_passed(1) = astring%count(substring=' ')==10
   !< astring = 'Hello World  !    '
   !< test_passed(2) = astring%count(substring=' ', ignore_isolated=.true.)==6
   !< astring = '    Hello World  !'
   !< test_passed(3) = astring%count(substring=' ', ignore_isolated=.true.)==6
   !< astring = '   Hello World  !    '
   !< test_passed(4) = astring%count(substring=' ', ignore_isolated=.true.)==8
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)              :: self             !< The string.
   character(*),  intent(in)              :: substring        !< Substring.
   logical,       intent(in), optional    :: ignore_isolated  !< Ignore "isolated" occurrences.
   integer                                :: No               !< Number of occurrences.
   logical                                :: ignore_isolated_ !< Ignore "isolated" occurrences, local variable.
   integer                                :: c1               !< Counter.
   integer                                :: c2               !< Counter.

   No = 0
   if (allocated(self%raw)) then
      if (len(substring)>len(self%raw)) return
      ignore_isolated_ = .false. ; if (present(ignore_isolated)) ignore_isolated_ = ignore_isolated
      c1 = 1
      do
         c2 = index(string=self%raw(c1:), substring=substring)
         if (c2==0) return
         if (.not.ignore_isolated_) then
            No = No + 1
         else
            if (.not.((c1==1.and.c2==1) .or. (c1==len(self%raw)-len(substring)+1))) then
               No = No + 1
            endif
         endif
         c1 = c1 + c2 - 1 + len(substring)
      enddo
   endif
   endfunction scount

   elemental function sindex_string_string(self, substring, back) result(i)
   !< Return the position of the start of the first occurrence of string `substring` as a substring in `string`, counting from one.
   !< If `substring` is not present in `string`, zero is returned. If the back argument is present and true, the return value is
   !< the start of the last occurrence rather than the first.
   !<
   !<```fortran
   !< type(string) :: string1
   !< type(string) :: string2
   !< logical      :: test_passed(2)
   !< string1 = 'Hello World Hello!'
   !< string2 = 'llo'
   !< test_passed(1) = string1%index(substring=string2)==index(string='Hello World Hello!', substring='llo')
   !< test_passed(2) = string1%index(substring=string2, back=.true.)==index(string='Hello World Hello!', substring='llo', &
   !<                                                                       back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self      !< The string.
   type(string),  intent(in)           :: substring !< Searched substring.
   logical,       intent(in), optional :: back      !< Start of the last occurrence rather than the first.
   integer                             :: i         !< Result of the search.

   if (allocated(self%raw)) then
      i = index(string=self%raw, substring=substring%raw, back=back)
   else
      i = 0
   endif
   endfunction sindex_string_string

   elemental function sindex_string_character(self, substring, back) result(i)
   !< Return the position of the start of the first occurrence of string `substring` as a substring in `string`, counting from one.
   !< If `substring` is not present in `string`, zero is returned. If the back argument is present and true, the return value is
   !< the start of the last occurrence rather than the first.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(2)
   !< string1 = 'Hello World Hello!'
   !< test_passed(1) = string1%index(substring='llo')==index(string='Hello World Hello!', substring='llo')
   !< test_passed(2) = string1%index(substring='llo', back=.true.)==index(string='Hello World Hello!', substring='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   character(kind=CK, len=*), intent(in)           :: substring !< Searched substring.
   logical,                   intent(in), optional :: back      !< Start of the last occurrence rather than the first.
   integer                                         :: i         !< Result of the search.

   if (allocated(self%raw)) then
      i = index(string=self%raw, substring=substring, back=back)
   else
      i = 0
   endif
   endfunction sindex_string_character

   elemental function slen(self) result(l)
   !< Return the length of a string.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'Hello World!   '
   !< print "(L1)", astring%len()==len('Hello World!   ')
   !<```
   !=> T <<<
   class(string), intent(in) :: self !< The string.
   integer                   :: l    !< String length.

   if (allocated(self%raw)) then
      l = len(string=self%raw)
   else
      l = 0
   endif
   endfunction slen

   elemental function slen_trim(self) result(l)
   !< Return the length of a string, ignoring any trailing blanks.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'Hello World!   '
   !< print "(L1)", astring%len_trim()==len_trim('Hello World!   ')
   !<```
   !=> T <<<
   class(string), intent(in) :: self !< The string.
   integer                   :: l    !< String length.

   if (allocated(self%raw)) then
      l = len_trim(string=self%raw)
   else
      l = 0
   endif
   endfunction slen_trim

   elemental function srepeat_string_string(self, ncopies) result(repeated)
   !< Concatenates several copies of an input string.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'x'
   !< print "(L1)", astring%repeat(5)//''=='xxxxx'
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< String to be repeated.
   integer,       intent(in) :: ncopies  !< Number of string copies.
   type(string)              :: repeated !< Repeated string.
#ifdef _NVF
   character(9999)           :: nvf_bug  !< Work around for NVFortran bug.
#endif

#ifdef _NVF
   nvf_bug = self%raw
   repeated%raw = repeat(string=trim(nvf_bug), ncopies=ncopies)
#else
   repeated%raw = repeat(string=self%raw, ncopies=ncopies)
#endif
   endfunction srepeat_string_string

   elemental function srepeat_character_string(rstring, ncopies) result(repeated)
   !< Concatenates several copies of an input string.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'y'
   !< print "(L1)", astring%repeat('x', 5)//''=='xxxxx'
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: rstring  !< String to be repeated.
   integer,                   intent(in) :: ncopies  !< Number of string copies.
   type(string)                          :: repeated !< Repeated string.

   repeated%raw = repeat(string=rstring, ncopies=ncopies)
   endfunction srepeat_character_string

   elemental function sscan_string_string(self, set, back) result(i)
   !< Return the leftmost (if `back` is either absent or equals false, otherwise the rightmost) character of string that is in `set`.
   !<
   !<```fortran
   !< type(string) :: string1
   !< type(string) :: string2
   !< logical      :: test_passed(2)
   !< string1 = 'Hello World Hello!'
   !< string2 = 'llo'
   !< test_passed(1) = string1%scan(set=string2)==scan(string='Hello World Hello!', set='llo')
   !< test_passed(2) = string1%scan(set=string2, back=.true.)==scan(string='Hello World Hello!', set='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self  !< The string.
   type(string),  intent(in)           :: set   !< Searched set.
   logical,       intent(in), optional :: back  !< Start of the last occurrence rather than the first.
   integer                             :: i     !< Result of the search.

   if (allocated(self%raw).and.allocated(set%raw)) then
     i = scan(string=self%raw, set=set%raw, back=back)
   else
     i = 0
   endif
   endfunction sscan_string_string

   elemental function sscan_string_character(self, set, back) result(i)
   !< Return the leftmost (if `back` is either absent or equals false, otherwise the rightmost) character of string that is in `set`.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(2)
   !< string1 = 'Hello World Hello!'
   !< test_passed(1) = string1%scan(set='llo')==scan(string='Hello World Hello!', set='llo')
   !< test_passed(2) = string1%scan(set='llo', back=.true.)==scan(string='Hello World Hello!', set='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self  !< The string.
   character(kind=CK, len=*), intent(in)           :: set   !< Searched set.
   logical,                   intent(in), optional :: back  !< Start of the last occurrence rather than the first.
   integer                                         :: i     !< Result of the search.

   if (allocated(self%raw)) then
     i = scan(string=self%raw, set=set, back=back)
   else
     i = 0
   endif
   endfunction sscan_string_character

   elemental function strim(self) result(trimmed)
   !< Remove trailing spaces.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'Hello World!   '
   !< print "(L1)", astring%trim()==trim('Hello World!   ')
   !<```
   !=> T <<<
   class(string), intent(in) :: self    !< The string.
   type(string)              :: trimmed !< Trimmed string.

   trimmed = self
   if (allocated(trimmed%raw)) trimmed%raw = trim(trimmed%raw)
   endfunction strim

   elemental function sverify_string_string(self, set, back) result(i)
   !< Return the leftmost (if `back` is either absent or equals false, otherwise the rightmost) character of string that is not
   !< in `set`. If all characters of `string` are found in `set`, the result is zero.
   !<
   !<```fortran
   !< type(string) :: string1
   !< type(string) :: string2
   !< logical      :: test_passed(2)
   !< string1 = 'Hello World Hello!'
   !< string2 = 'llo'
   !< test_passed(1) = string1%verify(set=string2)==verify(string='Hello World Hello!', set='llo')
   !< test_passed(2) = string1%verify(set=string2, back=.true.)==verify(string='Hello World Hello!', set='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self  !< The string.
   type(string),  intent(in)           :: set   !< Searched set.
   logical,       intent(in), optional :: back  !< Start of the last occurrence rather than the first.
   integer                             :: i     !< Result of the search.

   if (allocated(self%raw).and.allocated(set%raw)) then
     i = verify(string=self%raw, set=set%raw, back=back)
   else
     i = 0
   endif
   endfunction sverify_string_string

   elemental function sverify_string_character(self, set, back) result(i)
   !< Return the leftmost (if `back` is either absent or equals false, otherwise the rightmost) character of string that is not
   !< in `set`. If all characters of `string` are found in `set`, the result is zero.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(2)
   !< string1 = 'Hello World Hello!'
   !< test_passed(1) = string1%verify(set='llo')==verify(string='Hello World Hello!', set='llo')
   !< test_passed(2) = string1%verify(set='llo', back=.true.)==verify(string='Hello World Hello!', set='llo', back=.true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self  !< The string.
   character(kind=CK, len=*), intent(in)           :: set   !< Searched set.
   logical,                   intent(in), optional :: back  !< Start of the last occurrence rather than the first.
   integer                                         :: i     !< Result of the search.

   if (allocated(self%raw)) then
     i = verify(string=self%raw, set=set, back=back)
   else
     i = 0
   endif
   endfunction sverify_string_character

   ! auxiliary methods
   elemental function basedir(self, sep)
   !< Return the base directory name of a string containing a file name.
   !<
   !<```fortran
   !< type(string) :: string1
   !< logical      :: test_passed(4)
   !< string1 = '/bar/foo.tar.bz2'
   !< test_passed(1) = string1%basedir()//''=='/bar'
   !< string1 = './bar/foo.tar.bz2'
   !< test_passed(2) = string1%basedir()//''=='./bar'
   !< string1 = 'bar/foo.tar.bz2'
   !< test_passed(3) = string1%basedir()//''=='bar'
   !< string1 = '\bar\foo.tar.bz2'
   !< test_passed(4) = string1%basedir(sep='\')//''=='\bar'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self    !< The string.
   character(kind=CK, len=*), intent(in), optional :: sep     !< Directory separator.
   type(string)                                    :: basedir !< Base directory name.
   character(kind=CK, len=:), allocatable          :: sep_    !< Separator, default value.
   integer                                         :: pos     !< Character position.

   if (allocated(self%raw)) then
     sep_ = UIX_DIR_SEP ; if (present(sep)) sep_ = sep
     basedir = self
     pos = index(self%raw, sep_, back=.true.)
     if (pos>0) basedir%raw = self%raw(1:pos-1)
   endif
   endfunction basedir

   elemental function basename(self, sep, extension, strip_last_extension)
   !< Return the base file name of a string containing a file name.
   !<
   !< Optionally, the extension is also stripped if provided or the last one if required, e.g.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(5)
   !< astring = 'bar/foo.tar.bz2'
   !< test_passed(1) = astring%basename()//''=='foo.tar.bz2'
   !< test_passed(2) = astring%basename(extension='.tar.bz2')//''=='foo'
   !< test_passed(3) = astring%basename(strip_last_extension=.true.)//''=='foo.tar'
   !< astring = '\bar\foo.tar.bz2'
   !< test_passed(4) = astring%basename(sep='\')//''=='foo.tar.bz2'
   !< astring = 'bar'
   !< test_passed(5) = astring%basename(strip_last_extension=.true.)//''=='bar'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self                 !< The string.
   character(kind=CK, len=*), intent(in), optional :: sep                  !< Directory separator.
   character(kind=CK, len=*), intent(in), optional :: extension            !< File extension.
   logical,                   intent(in), optional :: strip_last_extension !< Flag to enable the stripping of last extension.
   type(string)                                    :: basename             !< Base file name.
   character(kind=CK, len=:), allocatable          :: sep_                 !< Separator, default value.
   integer                                         :: pos                  !< Character position.

   if (allocated(self%raw)) then
      sep_ = UIX_DIR_SEP ; if (present(sep)) sep_ = sep
      basename = self
      pos = index(basename%raw, sep_, back=.true.)
      if (pos>0) basename%raw = self%raw(pos+1:)
      if (present(extension)) then
         pos = index(basename%raw, extension, back=.true.)
         if (pos>0) basename%raw = basename%raw(1:pos-1)
      elseif (present(strip_last_extension)) then
         if (strip_last_extension) then
            pos = index(basename%raw, '.', back=.true.)
            if (pos>0) basename%raw = basename%raw(1:pos-1)
         endif
      endif
   endif
   endfunction basename

   elemental function camelcase(self, sep)
   !< Return a string with all words capitalized without spaces.
   !<
   !< @note Multiple subsequent separators are collapsed to one occurence.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'caMeL caSe var'
   !< print '(L1)', astring%camelcase()//''=='CamelCaseVar'
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   type(string)                                    :: camelcase !< Camel case string.
   type(string), allocatable                       :: tokens(:) !< String tokens.

   if (allocated(self%raw)) then
     call self%split(tokens=tokens, sep=sep)
     tokens = tokens%capitalize()
     camelcase = camelcase%join(array=tokens)
   endif
   endfunction camelcase

   elemental function capitalize(self) result(capitalized)
   !< Return a string with its first character capitalized and the rest lowercased.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'say all Hello WorLD!'
   !< print '(L1)', astring%capitalize()//''=='Say all hello world!'
   !<```
   !=> T <<<
   class(string), intent(in) :: self        !< The string.
   type(string)              :: capitalized !< Upper case string.
   integer                   :: c           !< Character counter.

   if (allocated(self%raw)) then
     capitalized = self%lower()
     c = index(LOWER_ALPHABET, capitalized%raw(1:1))
     if (c>0) capitalized%raw(1:1) = UPPER_ALPHABET(c:c)
   endif
   endfunction capitalize

   pure function chars(self) result(raw)
   !< Return the raw characters data.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'say all Hello WorLD!'
   !< print '(L1)', astring%chars()=='say all Hello WorLD!'
   !<```
   !=> T <<<
   class(string), intent(in)              :: self !< The string.
   character(kind=CK, len=:), allocatable :: raw  !< Raw characters data.

   if (allocated(self%raw)) then
     raw = self%raw
   else
     raw = ''
   endif
   endfunction chars

   pure function colorize_str(self, color_fg, color_bg, style) result(colorized)
   !< Colorize and stylize strings, DEFAULT kind.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'say all Hello WorLD!'
   !< print '(L1)', astring%colorize(color_fg='red')=='say all Hello WorLD!'
   !<```
   !=> T <<<
   class(string),    intent(in)           :: self      !< The string.
   character(len=*), intent(in), optional :: color_fg  !< Foreground color definition.
   character(len=*), intent(in), optional :: color_bg  !< Background color definition.
   character(len=*), intent(in), optional :: style     !< Style definition.
   character(len=:), allocatable          :: colorized !< Colorized string.

   colorized = colorize(string=self%chars(), color_fg=color_fg, color_bg=color_bg, style=style)
   endfunction colorize_str

   elemental function decode(self, codec) result(decoded)
   !< Return a string decoded accordingly the codec.
   !<
   !< @note Only BASE64 codec is currently available.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'SG93IGFyZSB5b3U/'
   !< print '(L1)', astring%decode(codec='base64')//''=='How are you?'
   !<```
   !=> T <<<
   class(string),             intent(in) :: self    !< The string.
   character(kind=CK, len=*), intent(in) :: codec   !< Encoding codec.
   type(string)                          :: decoded !< Decoded string.
   type(string)                          :: codec_u !< Encoding codec in upper case string.

   if (allocated(self%raw)) then
     decoded = self
     codec_u = codec
     select case(codec_u%upper()//'')
     case('BASE64')
       call b64_decode(code=self%raw, s=decoded%raw)
     endselect
     decoded = decoded%strip(remove_nulls=.true.)
   endif
   endfunction decode

   elemental function encode(self, codec) result(encoded)
   !< Return a string encoded accordingly the codec.
   !<
   !< @note Only BASE64 codec is currently available.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'How are you?'
   !< print '(L1)', astring%encode(codec='base64')//''=='SG93IGFyZSB5b3U/'
   !<```
   !=> T <<<
   class(string),             intent(in) :: self    !< The string.
   character(kind=CK, len=*), intent(in) :: codec   !< Encoding codec.
   type(string)                          :: encoded !< Encoded string.

   if (allocated(self%raw)) then
     encoded = codec
     select case(encoded%upper()//'')
     case('BASE64')
       call b64_encode(s=self%raw, code=encoded%raw)
     endselect
   endif
   endfunction encode

   elemental function escape(self, to_escape, esc) result(escaped)
   !< Escape backslashes (or custom escape character).
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(2)
   !< astring = '^\s \d+\s*'
   !< test_passed(1) = astring%escape(to_escape='\')//''=='^\\s \\d+\\s*'
   !< test_passed(2) = astring%escape(to_escape='\', esc='|')//''=='^|\s |\d+|\s*'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   character(kind=CK, len=1), intent(in)           :: to_escape !< Character to be escaped.
   character(kind=CK, len=*), intent(in), optional :: esc       !< Character used to escape.
   type(string)                                    :: escaped   !< Escaped string.
   character(kind=CK, len=:), allocatable          :: esc_      !< Character to escape, local variable.
   integer                                         :: c         !< Character counter.

   if (allocated(self%raw)) then
     esc_ = BACKSLASH ; if (present(esc)) esc_ = esc
     escaped%raw = ''
     do c=1, len(self%raw)
       if (self%raw(c:c)==to_escape) then
         escaped%raw = escaped%raw//esc_//to_escape
       else
         escaped%raw = escaped%raw//self%raw(c:c)
       endif
     enddo
   endif
   endfunction escape

   elemental function extension(self)
   !< Return the extension of a string containing a file name.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = '/bar/foo.tar.bz2'
   !< print '(L1)', astring%extension()//''=='.bz2'
   !<```
   !=> T <<<
   class(string), intent(in)              :: self      !< The string.
   type(string)                           :: extension !< Extension file name.
   integer                                :: pos       !< Character position.

   if (allocated(self%raw)) then
      extension = ''
      pos = index(self%raw, '.', back=.true.)
      if (pos>0) extension%raw = self%raw(pos:)
   endif
   endfunction extension

   elemental function fill(self, width, right, filling_char) result(filled)
   !< Pad string on the left (or right) with zeros (or other char) to fill width.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(4)
   !< astring = 'this is string example....wow!!!'
   !< test_passed(1) = astring%fill(width=40)//''=='00000000this is string example....wow!!!'
   !< test_passed(2) = astring%fill(width=50)//''=='000000000000000000this is string example....wow!!!'
   !< test_passed(3) = astring%fill(width=50, right=.true.)//''=='this is string example....wow!!!000000000000000000'
   !< test_passed(4) = astring%fill(width=40, filling_char='*')//''=='********this is string example....wow!!!'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self          !< The string.
   integer,                   intent(in)           :: width         !< Final width of filled string.
   logical,                   intent(in), optional :: right         !< Fill on the right instead of left.
   character(kind=CK, len=1), intent(in), optional :: filling_char  !< Filling character (default "0").
   type(string)                                    :: filled        !< Filled string.
   logical                                         :: right_        !< Fill on the right instead of left, local variable.
   character(kind=CK, len=1)                       :: filling_char_ !< Filling character (default "0"), local variable.

   if (allocated(self%raw)) then
      if (width>len(self%raw)) then
         right_ = .false. ; if (present(right)) right_ = right
         filling_char_ = '0' ; if (present(filling_char)) filling_char_ = filling_char
         if (.not.right_) then
            filled%raw = repeat(filling_char_, width-len(self%raw))//self%raw
         else
            filled%raw = self%raw//repeat(filling_char_, width-len(self%raw))
         endif
      endif
   endif
   endfunction fill

   elemental subroutine free(self)
   !< Free dynamic memory.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'this is string example....wow!!!'
   !< call astring%free
   !< print '(L1)', astring%is_allocated().eqv..false.
   !<```
   !=> T <<<
   class(string), intent(inout) :: self !< The string.

   if (allocated(self%raw)) deallocate(self%raw)
   endsubroutine free

   subroutine glob_character(self, pattern, list)
   !< Glob search (character output), finds all the pathnames matching a given pattern according to the rules used by the Unix shell.
   !<
   !< @note Method not portable: works only on Unix/GNU Linux OS.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: alist_chr(:)
   !< integer, parameter            :: Nf=5
   !< character(14)                 :: files(1:Nf)
   !< integer                       :: file_unit
   !< integer                       :: f
   !< integer                       :: ff
   !< logical                       :: test_passed
   !< do f=1, Nf
   !<    files(f) = astring%tempname(prefix='foo-')
   !<    open(newunit=file_unit, file=files(f))
   !<    write(file_unit, *)f
   !<    close(unit=file_unit)
   !< enddo
   !< call astring%glob(pattern='foo-*', list=alist_chr)
   !< do f=1, Nf
   !<    open(newunit=file_unit, file=files(f))
   !<    close(unit=file_unit, status='delete')
   !< enddo
   !< test_passed = .false.
   !< outer_chr: do f=1, size(alist_chr, dim=1)
   !<    do ff=1, Nf
   !<       test_passed = alist_chr(f) == files(ff)
   !<       if (test_passed) cycle outer_chr
   !<    enddo
   !< enddo outer_chr
   !< print '(L1)', test_passed
   !<```
   !=> T <<<
   class(string),                 intent(in)  :: self           !< The string.
   character(*),                  intent(in)  :: pattern        !< Given pattern.
   character(len=:), allocatable, intent(out) :: list(:)        !< List of matching pathnames.
   type(string), allocatable                  :: list_(:)       !< List of matching pathnames.
   integer(I4P)                               :: max_len        !< Maximum length.
   integer(I4P)                               :: matches_number !< Matches number.
   integer(I4P)                               :: m              !< Counter.

   call self%glob(pattern=pattern, list=list_)
   if (allocated(list_)) then
      matches_number = size(list_, dim=1)
      max_len = 0
      do m=1, matches_number
         max_len = max(max_len, list_(m)%len())
      enddo
      allocate(character(max_len) :: list(1:matches_number))
      do m=1, matches_number
         list(m) = list_(m)%chars()
      enddo
   endif
   endsubroutine glob_character

   subroutine glob_string(self, pattern, list)
   !< Glob search (string output), finds all the pathnames matching a given pattern according to the rules used by the Unix shell.
   !<
   !< @note Method not portable: works only on Unix/GNU Linux OS.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< type(string),     allocatable :: alist_str(:)
   !< integer, parameter            :: Nf=5
   !< character(14)                 :: files(1:Nf)
   !< integer                       :: file_unit
   !< integer                       :: f
   !< integer                       :: ff
   !< logical                       :: test_passed
   !<
   !< do f=1, Nf
   !<    files(f) = astring%tempname(prefix='foo-')
   !<    open(newunit=file_unit, file=files(f))
   !<    write(file_unit, *)f
   !<    close(unit=file_unit)
   !< enddo
   !< call astring%glob(pattern='foo-*', list=alist_str)
   !< do f=1, Nf
   !<    open(newunit=file_unit, file=files(f))
   !<    close(unit=file_unit, status='delete')
   !< enddo
   !< test_passed = .false.
   !< outer_str: do f=1, size(alist_str, dim=1)
   !<    do ff=1, Nf
   !<       test_passed = alist_str(f) == files(ff)
   !<       if (test_passed) cycle outer_str
   !<    enddo
   !< enddo outer_str
   !< print '(L1)', test_passed
   !<```
   !=> T <<<
   class(string),             intent(in)  :: self     !< The string.
   character(*),              intent(in)  :: pattern  !< Given pattern.
   type(string), allocatable, intent(out) :: list(:)  !< List of matching pathnames.
   type(string)                           :: tempfile !< Safe temporary file.
   character(len=:), allocatable          :: tempname !< Safe temporary name.
   integer(I4P)                           :: tempunit !< Unit of temporary file.

   tempname = self%tempname()
   call execute_command_line('ls -1 '//trim(adjustl(pattern))//' > '//tempname)
   call tempfile%read_file(file=tempname)
   call tempfile%split(sep=new_line('a'), tokens=list)
   open(newunit=tempunit, file=tempname)
   close(unit=tempunit, status='delete')
   endsubroutine glob_string

   elemental function insert_character(self, substring, pos) result(inserted)
   !< Insert substring into string at a specified position.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(5)
   !< astring = 'this is string example wow!!!'
   !< acharacter = '... '
   !< test_passed(1) = astring%insert(substring=acharacter, pos=1)//''=='... this is string example wow!!!'
   !< test_passed(2) = astring%insert(substring=acharacter, pos=23)//''=='this is string example...  wow!!!'
   !< test_passed(3) = astring%insert(substring=acharacter, pos=29)//''=='this is string example wow!!!... '
   !< test_passed(4) = astring%insert(substring=acharacter, pos=-1)//''=='... this is string example wow!!!'
   !< test_passed(5) = astring%insert(substring=acharacter, pos=100)//''=='this is string example wow!!!... '
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),    intent(in) :: self      !< The string.
   character(len=*), intent(in) :: substring !< Substring.
   integer,          intent(in) :: pos       !< Position from which insert substring.
   type(string)                 :: inserted  !< Inserted string.
   integer                      :: safepos   !< Safe position from which insert substring.

   if (allocated(self%raw)) then
      inserted = self
      safepos = min(max(1, pos), len(self%raw))
      if (safepos==1) then
         inserted%raw = substring//self%raw
      elseif (safepos==len(self%raw)) then
         inserted%raw = self%raw//substring
      else
         inserted%raw = self%raw(1:safepos-1)//substring//self%raw(safepos:)
      endif
   else
      inserted%raw = substring
   endif
   endfunction insert_character

   elemental function insert_string(self, substring, pos) result(inserted)
   !< Insert substring into string at a specified position.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(5)
   !< astring = 'this is string example wow!!!'
   !< anotherstring = '... '
   !< test_passed(1) = astring%insert(substring=anotherstring, pos=1)//''=='... this is string example wow!!!'
   !< test_passed(2) = astring%insert(substring=anotherstring, pos=23)//''=='this is string example...  wow!!!'
   !< test_passed(3) = astring%insert(substring=anotherstring, pos=29)//''=='this is string example wow!!!... '
   !< test_passed(4) = astring%insert(substring=anotherstring, pos=-1)//''=='... this is string example wow!!!'
   !< test_passed(5) = astring%insert(substring=anotherstring, pos=100)//''=='this is string example wow!!!... '
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   type(string),  intent(in) :: substring !< Substring.
   integer,       intent(in) :: pos       !< Position from which insert substring.
   type(string)              :: inserted  !< Inserted string.
   integer                   :: safepos   !< Safe position from which insert substring.

   if (allocated(self%raw)) then
      inserted = self
      if (allocated(substring%raw)) then
         safepos = min(max(1, pos), len(self%raw))
         if (safepos==1) then
            inserted%raw = substring%raw//self%raw
         elseif (safepos==len(self%raw)) then
            inserted%raw = self%raw//substring%raw
         else
            inserted%raw = self%raw(1:safepos-1)//substring%raw//self%raw(safepos:)
         endif
      endif
   else
      if (allocated(substring%raw)) inserted%raw = substring%raw
   endif
   endfunction insert_string

   pure function join_strings(self, array, sep) result(join)
   !< Return a string that is a join of an array of strings.
   !<
   !< The join-separator is set equals to self if self has a value or it is set to a null string ''. This value can be overridden
   !< passing a custom separator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: strings(3)
   !< logical      :: test_passed(5)
   !< strings(1) = 'one'
   !< strings(2) = 'two'
   !< strings(3) = 'three'
   !< test_passed(1) = (astring%join(array=strings)//''==strings(1)//strings(2)//strings(3))
   !< test_passed(2) = (astring%join(array=strings, sep='-')//''==strings(1)//'-'//strings(2)//'-'//strings(3))
   !< call strings(1)%free
   !< strings(2) = 'two'
   !< strings(3) = 'three'
   !< test_passed(3) = (astring%join(array=strings, sep='-')//''==strings(2)//'-'//strings(3))
   !< strings(1) = 'one'
   !< strings(2) = 'two'
   !< call strings(3)%free
   !< test_passed(4) = (astring%join(array=strings, sep='-')//''==strings(1)//'-'//strings(2))
   !< strings(1) = 'one'
   !< call strings(2)%free
   !< strings(3) = 'three'
   !< test_passed(5) = (astring%join(array=strings, sep='-')//''==strings(1)//'-'//strings(3))
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   type(string),              intent(in)           :: array(1:) !< Array to be joined.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   type(string)                                    :: join      !< The join of array.
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   integer                                         :: a         !< Counter.

   if (allocated(self%raw)) then
      sep_ = self%raw
   else
      sep_ = ''
   endif
   if (present(sep)) sep_ = sep
   join = ''
   do a=2, size(array, dim=1)
      if (allocated(array(a)%raw)) join%raw = join%raw//sep_//array(a)%raw
   enddo
   if (allocated(array(1)%raw)) then
      join%raw = array(1)%raw//join%raw
   else
      join%raw = join%raw(len(sep_)+1:len(join%raw))
   endif
   endfunction join_strings

   pure function join_characters(self, array, sep) result(join)
   !< Return a string that is a join of an array of characters.
   !<
   !< The join-separator is set equals to self if self has a value or it is set to a null string ''. This value can be overridden
   !< passing a custom separator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< character(5) :: characters(3)
   !< logical      :: test_passed(6)
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< test_passed(1) = (astring%join(array=characters)//''==characters(1)//characters(2)//characters(3))
   !< test_passed(2) = (astring%join(array=characters, sep='-')//''==characters(1)//'-'//characters(2)//'-'//characters(3))
   !< characters(1) = ''
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< test_passed(3) = (astring%join(array=characters, sep='-')//''==characters(2)//'-'//characters(3))
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = ''
   !< test_passed(4) = (astring%join(array=characters, sep='-')//''==characters(1)//'-'//characters(2))
   !< characters(1) = 'one'
   !< characters(2) = ''
   !< characters(3) = 'three'
   !< test_passed(5) = (astring%join(array=characters, sep='-')//''==characters(1)//'-'//characters(3))
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< astring = '_'
   !< test_passed(6) = (astring%join(array=characters)//''==characters(1)//'_'//characters(2)//'_'//characters(3))
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   character(kind=CK, len=*), intent(in)           :: array(1:) !< Array to be joined.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   type(string)                                    :: join      !< The join of array.
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   integer                                         :: a         !< Counter.

   if (allocated(self%raw)) then
      sep_ = self%raw
   else
      sep_ = ''
   endif
   if (present(sep)) sep_ = sep
   join = ''
   do a=2, size(array, dim=1)
      if (array(a)/='') join%raw = join%raw//sep_//array(a)
   enddo
   if (array(1)/='') then
      join%raw = array(1)//join%raw
   else
      join%raw = join%raw(len(sep_)+1:len(join%raw))
   endif
   endfunction join_characters

   pure function strjoin_strings(array, sep) result(join)
   !< Return a string that is a join of an array of strings.
   !<
   !< The join-separator is set equals to a null string '' if custom separator isn't specified.
   !<
   !<```fortran
   !< type(string)     :: strings(3)
   !< logical          :: test_passed(5)
   !< strings(1) = 'one'
   !< strings(2) = 'two'
   !< strings(3) = 'three'
   !< test_passed(1) = (strjoin(array=strings)//''==strings(1)//strings(2)//strings(3))
   !< test_passed(2) = (strjoin(array=strings, sep='-')//''==strings(1)//'-'//strings(2)//'-'//strings(3))
   !< call strings(1)%free
   !< strings(2) = 'two'
   !< strings(3) = 'three'
   !< test_passed(3) = (strjoin(array=strings, sep='-')//''==strings(2)//'-'//strings(3))
   !< strings(1) = 'one'
   !< strings(2) = 'two'
   !< call strings(3)%free
   !< test_passed(4) = (strjoin(array=strings, sep='-')//''==strings(1)//'-'//strings(2))
   !< strings(1) = 'one'
   !< call strings(2)%free
   !< strings(3) = 'three'
   !< test_passed(5) = (strjoin(array=strings, sep='-')//''==strings(1)//'-'//strings(3))
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: array(1:) !< Array to be joined.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   type(string)                                    :: join      !< The join of array.
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   integer                                         :: a         !< Counter.

   sep_ = ''
   if (present(sep)) sep_ = sep
   join = ''
   do a=2, size(array, dim=1)
      if (allocated(array(a)%raw))join%raw = join%raw//sep_//array(a)%raw
   enddo
   if (allocated(array(1)%raw)) then
      join%raw = array(1)%raw//join%raw
   else
      join%raw = join%raw(len(sep_)+1:len(join%raw))
   endif
   endfunction strjoin_strings

  pure function strjoin_characters(array, sep, is_trim) result(join)
   !< Return a string that is a join of an array of characters.
   !<
   !< The join-separator is set equals to a null string '' if custom separator isn't specified.
   !< The trim function is applied to array items if optional logical is_trim variable isn't set to .false.
   !<
   !<```fortran
   !< character(5) :: characters(3)
   !< logical      :: test_passed(13)
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< test_passed(1) = (strjoin(array=characters)//''==trim(characters(1))//trim(characters(2))//trim(characters(3)))
   !< test_passed(2) = (strjoin(array=characters, sep='-')//''==trim(characters(1))//'-'//trim(characters(2))//'-'//trim(characters(3)))
   !< test_passed(3) = ( strjoin(array=characters, is_trim=.false.)//''==characters(1)//characters(2)//characters(3))
   !< test_passed(4) = ( strjoin(array=characters, sep='-', is_trim=.false.)//''==characters(1)//'-'//characters(2)//'-'//characters(3))
   !< characters(1) = ''
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< test_passed(5) = (strjoin(array=characters)//''==trim(characters(2))//trim(characters(3)))
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = ''
   !< test_passed(6) = (strjoin(array=characters)//''==trim(characters(1))//trim(characters(2)))
   !< characters(1) = 'one'
   !< characters(2) = ''
   !< characters(3) = 'three'
   !< test_passed(7) = (strjoin(array=characters)//''==trim(characters(1))//trim(characters(3)))
   !< characters(1) = ''
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< test_passed(8) = (strjoin(array=characters, sep='-')//''==trim(characters(2))//'-'//trim(characters(3)))
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = ''
   !< test_passed(9) = (strjoin(array=characters, sep='-')//''==trim(characters(1))//'-'//trim(characters(2)))
   !< characters(1) = 'one'
   !< characters(2) = ''
   !< characters(3) = 'three'
   !< test_passed(10) = (strjoin(array=characters, sep='-')//''==trim(characters(1))//'-'//trim(characters(3)))
   !< characters(1) = ''
   !< characters(2) = 'two'
   !< characters(3) = 'three'
   !< test_passed(11) = (strjoin(array=characters, sep='-', is_trim=.false.)//''==characters(2)//'-'//characters(3))
   !< characters(1) = 'one'
   !< characters(2) = 'two'
   !< characters(3) = ''
   !< test_passed(12) = (strjoin(array=characters, sep='-', is_trim=.false.)//''==characters(1)//'-'//characters(2))
   !< characters(1) = 'one'
   !< characters(2) = ''
   !< characters(3) = 'three'
   !< test_passed(13) = (strjoin(array=characters, sep='-', is_trim=.false.)//''==characters(1)//'-'//characters(3))
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)           :: array(1:) !< Array to be joined.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   logical,                   intent(in), optional :: is_trim   !< Flag to setup trim character or not
   type(string)                                    :: join      !< The join of array.
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   logical                                         :: is_trim_  !< Flag to setup trim character or not
   integer                                         :: a         !< Counter.

   sep_ = ''
   if (present(sep)) sep_ = sep
   is_trim_ = .true. ; if (present(is_trim)) is_trim_ = is_trim
   join = ''

   if (is_trim_) then
       do a=2, size(array, dim=1)
          if (trim(array(a))/='') join%raw = join%raw//sep_//trim(array(a))
       enddo
       if (trim(array(1))/='') then
          join%raw = trim(array(1))//join%raw
       else
          join%raw = join%raw(len(sep_)+1:len(join%raw))
       endif
   else
       do a=2, size(array, dim=1)
          if (array(a)/='') join%raw = join%raw//sep_//array(a)
       enddo
       if (array(1)/='') then
          join%raw = array(1)//join%raw
       else
          join%raw = join%raw(len(sep_)+1:len(join%raw))
       endif
   endif
   endfunction strjoin_characters

   pure function strjoin_strings_array(array, sep, is_col) result(join)
   !< Return a string that is a join of columns or rows of an array of strings.
   !<
   !< The join-separator is set equals to a null string '' if custom separator isn't specified.
   !< The is_col is setup the direction of join: within default columns (.true.) or rows(.false.).
   !<
   !<```fortran
   !< type(string), allocatable :: strings_arr(:, :)
   !< logical                   :: test_passed(5)
   !<
   !< strings_arr = reshape( source = &
   !<                        [string('one'), string('two'), string('three'),  &
   !<                         string('ONE'), string('TWO'), string('THREE')], &
   !<                        shape = [3, 2] )
   !<
   !< test_passed(1) = all( strjoin(array=strings_arr) == &
   !<                       reshape([string('onetwothree'), string('ONETWOTHREE')], &
   !<                       shape = [2]) )
   !<
   !< test_passed(2) = all( strjoin(array=strings_arr, sep='_') == &
   !<                       reshape([string('one_two_three'), string('ONE_TWO_THREE')], &
   !<                       shape = [2]) )
   !<
   !<  test_passed(3) = all( strjoin(array=strings_arr, is_col=.false.) == &
   !<                        reshape([string('oneONE'), string('twoTWO'), string('threeTHREE')], &
   !<                        shape = [3]) )
   !<
   !<  test_passed(4) = all( strjoin(array=strings_arr, sep='_', is_col=.false.) == &
   !<                        reshape([string('one_ONE'), string('two_TWO'), string('three_THREE')], &
   !<                        shape = [3]) )
   !<
   !< call strings_arr(2, 1)%free
   !< test_passed(5) = all( strjoin(array=strings_arr, sep='_', is_col=.false.) == &
   !<                  reshape([string('one_ONE'), string('TWO'), string('three_THREE')], &
   !<                  shape = [3]) )
   !<
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: array(1:, 1:) !< Array to be joined.
   character(kind=CK, len=*), intent(in), optional :: sep  !< Separator.
   logical,                   intent(in), optional :: is_col  !< Direction: 'columns' if .true. or 'rows' if .false.
   type(string),              allocatable          :: join(:)       !< The join of array.
   type(string),              allocatable          :: slice(:)      !< The column or row slice of array
   character(kind=CK, len=:), allocatable          :: sep_          !< Separator, default value.
   logical                                         :: is_col_       !< Direction, default value.
   integer                                         :: a, join_size, slice_size  !< Counter, sizes of join vector and of slice of array

   sep_    = ''     ; if (present(sep)) sep_ = sep
   is_col_ = .true. ; if (present(is_col)) is_col_ = is_col

   if (is_col_) then
       join_size  = size(array, dim=2)
       slice_size = size(array, dim=1)

       if (.not.allocated(join))  allocate(join(join_size))
       if (.not.allocated(slice)) allocate(slice(slice_size))
       do a = 1, join_size
           slice(:) = array(:, a)
           join(a)  = strjoin_strings(slice, sep_)
       end do
   else
       join_size  = size(array, dim=1)
       slice_size = size(array, dim=2)

       if (.not.allocated(join))  allocate(join(join_size))
       if (.not.allocated(slice)) allocate(slice(slice_size))
       do a = 1, join_size
           slice(:) = array(a, :)
           join(a)  = strjoin_strings(slice, sep_)
       end do
   endif
   endfunction strjoin_strings_array

  pure function strjoin_characters_array(array, sep, is_trim, is_col) result(join)
   !< Return a string that is a join of columns or rows of an array of characters.
   !<
   !< The join-separator is set equals to a null string '' if custom separator isn't specified.
   !< The trim function is applied to array items if optional logical is_trim variable isn't set to .false.
   !< The is_col is setup the direction of join: within default columns (.true.) or rows(.false.).
   !<
   !<```fortran
   !< character(len=10)         :: chars_arr(3, 2)
   !< logical                   :: test_passed(9)
   !< chars_arr(:, 1) = ['one       ', 'two       ', 'three     ']
   !< chars_arr(:, 2) = ['ONE       ', 'TWO       ', 'THREE     ']
   !<
   !< test_passed(1) = all( strjoin(array=chars_arr) == &
   !<                       reshape([string('onetwothree'), string('ONETWOTHREE')], &
   !<                       shape = [2]) )
   !<
   !< test_passed(2) = all( strjoin(array=chars_arr, is_trim=.false.) ==  &
   !<                       reshape([string('one       two       three     '),  &
   !<                                string('ONE       TWO       THREE     ')], &
   !<                       shape = [2]) )
   !<
   !< test_passed(3) = all( strjoin(array=chars_arr, sep='_') == &
   !<                       reshape([string('one_two_three'), string('ONE_TWO_THREE')], &
   !<                       shape = [2]) )
   !<
   !< test_passed(4) = all( strjoin(array=chars_arr, sep='_', is_trim=.false.) ==  &
   !<                       reshape([string('one       _two       _three     '),  &
   !<                                string('ONE       _TWO       _THREE     ')], &
   !<                       shape = [2]) )
   !<
   !< test_passed(5) = all( strjoin(array=chars_arr, is_col=.false.) == &
   !<                       reshape([string('oneONE'), string('twoTWO'), string('threeTHREE')], &
   !<                       shape = [3]) )
   !<
   !< test_passed(6) = all( strjoin(array=chars_arr, is_trim=.false., is_col=.false.) ==  &
   !<                       reshape([string('one       ONE       '),  &
   !<                                string('two       TWO       '),  &
   !<                                string('three     THREE     ')], &
   !<                       shape = [3]) )
   !<
   !< test_passed(7) = all( strjoin(array=chars_arr, sep='_', is_col=.false.) == &
   !<                       reshape([string('one_ONE'), string('two_TWO'), string('three_THREE')], &
   !<                       shape = [3]) )
   !<
   !< test_passed(8) = all( strjoin(array=chars_arr, sep='_', is_trim=.false., is_col=.false.) ==  &
   !<                       reshape([string('one       _ONE       '),  &
   !<                                string('two       _TWO       '),  &
   !<                                string('three     _THREE     ')], &
   !<                       shape = [3]) )
   !<
   !< chars_arr(2,1) = ''
   !< test_passed(9) = all( strjoin(array=chars_arr, sep='_', is_col=.false.) ==  &
   !<                       reshape([string('one_ONE'),  &
   !<                                string('TWO'),  &
   !<                                string('three_THREE')], &
   !<                       shape = [3]) )
   !<
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)           :: array(1:, 1:) !< Array to be joined.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   logical,                   intent(in), optional :: is_trim   !< Flag to setup trim character or not
   logical,                   intent(in), optional :: is_col    !< Direction: 'columns' if .true. or 'rows' if .false.
   type(string),              allocatable          :: join(:)   !< The join of array.
   character(kind=CK, len=:), allocatable          :: slice(:)  !< The column or row slice of array
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   logical                                         :: is_trim_  !< Flag to setup trim character or not
   logical                                         :: is_col_   !< Direction, default value.
   integer                                         :: a, join_size, slice_size  !< Counter, sizes of join vector and of slice of array
   integer                                         :: item_len  !< Length of array item (all items of character array have equal lengths)

   item_len = len(array(1,1)) !< all items of character array have equal lengths
   sep_     = ''     ; if (present(sep)) sep_ = sep
   is_trim_ = .true. ; if (present(is_trim)) is_trim_ = is_trim
   is_col_  = .true. ; if (present(is_col)) is_col_ = is_col

   if (is_col_) then
       join_size  = size(array, dim=2)
       slice_size = size(array, dim=1)

       if (.not.allocated(join))  allocate(join(join_size))
       if (.not.allocated(slice)) allocate(character(len=item_len) :: slice(slice_size))
       do a = 1, join_size
           slice(:) = array(:, a)
           join(a)  = strjoin_characters(slice, sep_, is_trim_)
       end do
   else
       join_size  = size(array, dim=1)
       slice_size = size(array, dim=2)

       if (.not.allocated(join))  allocate(join(join_size))
       if (.not.allocated(slice)) allocate(character(len=item_len) :: slice(slice_size))
       do a = 1, join_size
           slice(:) = array(a, :)
           join(a)  = strjoin_characters(slice, sep_, is_trim_)
       end do
   endif
   endfunction strjoin_characters_array

   elemental function lower(self)
   !< Return a string with all lowercase characters.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 'Hello WorLD!'
   !< test_passed(1) = astring%lower()//''=='hello world!'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self  !< The string.
   type(string)              :: lower !< Upper case string.
   integer                   :: n1    !< Characters counter.
   integer                   :: n2    !< Characters counter.

   if (allocated(self%raw)) then
      lower = self
      do n1=1, len(self%raw)
         n2 = index(UPPER_ALPHABET, self%raw(n1:n1))
         if (n2>0) lower%raw(n1:n1) = LOWER_ALPHABET(n2:n2)
      enddo
   endif
   endfunction lower

   pure function partition(self, sep) result(partitions)
   !< Split string at separator and return the 3 parts (before, the separator and after).
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: strings(3)
   !< logical      :: test_passed(3)
   !< astring = 'Hello WorLD!'
   !< strings = astring%partition(sep='lo Wo')
   !< test_passed(1) = (strings(1)//''=='Hel'.and.strings(2)//''=='lo Wo'.and.strings(3)//''=='rLD!')
   !< strings = astring%partition(sep='Hello')
   !< test_passed(2) = (strings(1)//''==''.and.strings(2)//''=='Hello'.and.strings(3)//''==' WorLD!')
   !< astring = 'Hello WorLD!'
   !< strings = astring%partition()
   !< test_passed(3) = (strings(1)//''=='Hello'.and.strings(2)//''==' '.and.strings(3)//''=='WorLD!')
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self            !< The string.
   character(kind=CK, len=*), intent(in), optional :: sep             !< Separator.
   type(string)                                    :: partitions(1:3) !< Partions: before the separator, the separator itsels and
                                                                      !< after the separator.
   character(kind=CK, len=:), allocatable          :: sep_            !< Separator, default value.
   integer                                         :: c               !< Character counter.

   if (allocated(self%raw)) then
      sep_ = SPACE ; if (present(sep)) sep_ = sep

      partitions(1) = self
      partitions(2) = sep_
      partitions(3) = ''
      if (len(sep_)>=len(self%raw)) return
      c = index(self%raw, sep_)
      if (c>0) then
         partitions(1)%raw = self%raw(1:c-1)
         partitions(2)%raw = self%raw(c:c+len(sep_)-1)
         partitions(3)%raw = self%raw(c+len(sep_):)
      endif
   endif
   endfunction partition

   subroutine read_file(self, file, is_fast, form, iostat, iomsg)
   !< Read a file as a single string stream.
   !<
   !< @note All the lines are stored into the string self as a single ascii stream. Each line (record) is separated by a `new_line`
   !< character.
   !<
   !< @note For unformatted read only `access='stream'` is supported with new_line as line terminator.
   !<
   !< @note *Fast* file reading allows a very efficient reading of streamed file, but it dumps file as single streamed string.
   !<
   !<```fortran
   !< type(string)              :: astring
   !< type(string), allocatable :: strings(:)
   !< type(string)              :: line(3)
   !< integer                   :: iostat
   !< character(len=99)         :: iomsg
   !< integer                   :: scratch
   !< integer                   :: l
   !< logical                   :: test_passed(9)
   !< line(1) = ' Hello World!   '
   !< line(2) = 'How are you?  '
   !< line(3) = '   All say: "Fine thanks"'
   !< open(newunit=scratch, file='read_file_test.tmp')
   !< write(scratch, "(A)") line(1)%chars()
   !< write(scratch, "(A)") line(2)%chars()
   !< write(scratch, "(A)") line(3)%chars()
   !< close(scratch)
   !< call astring%read_file(file='read_file_test.tmp', iostat=iostat, iomsg=iomsg)
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(1) = (size(strings, dim=1)==size(line, dim=1))
   !< do l=1, size(strings, dim=1)
   !<   test_passed(l+1) = (strings(l)==line(l))
   !< enddo
   !< open(newunit=scratch, file='read_file_test.tmp', form='UNFORMATTED', access='STREAM')
   !< write(scratch) line(1)%chars()//new_line('a')
   !< write(scratch) line(2)%chars()//new_line('a')
   !< write(scratch) line(3)%chars()//new_line('a')
   !< close(scratch)
   !< call astring%read_file(file='read_file_test.tmp', form='unformatted', iostat=iostat, iomsg=iomsg)
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(5) = (size(strings, dim=1)==size(line, dim=1))
   !< do l=1, size(strings, dim=1)
   !<   test_passed(l+5) = (strings(l)==line(l))
   !< enddo
   !< open(newunit=scratch, file='read_file_test.tmp', form='UNFORMATTED', access='STREAM')
   !< close(scratch, status='DELETE')
   !< call astring%read_file(file='read_file_test.tmp', iostat=iostat)
   !< test_passed(9) = (iostat/=0)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),    intent(inout)           :: self       !< The string.
   character(len=*), intent(in)              :: file       !< File name.
   logical,          intent(in),    optional :: is_fast    !< Flag to enable (super) fast file reading.
   character(len=*), intent(in),    optional :: form       !< Format of unit.
   integer,          intent(out),   optional :: iostat     !< IO status code.
   character(len=*), intent(inout), optional :: iomsg      !< IO status message.
   logical                                   :: is_fast_   !< Flag to enable (super) fast file reading, local variable.
   type(string)                              :: form_      !< Format of unit, local variable.
   integer                                   :: iostat_    !< IO status code, local variable.
   character(len=:), allocatable             :: iomsg_     !< IO status message, local variable.
   integer                                   :: unit       !< Logical unit.
   logical                                   :: does_exist !< Check if file exist.
   integer(I4P)                              :: filesize   !< Size of the file for fast reading.

   iomsg_ = repeat(' ', 99) ; if (present(iomsg)) iomsg_ = iomsg
   inquire(file=file, iomsg=iomsg_, iostat=iostat_, exist=does_exist)
   if (does_exist) then
      is_fast_ = .false. ; if (present(is_fast)) is_fast_ = is_fast
      if (is_fast_) then
         open(newunit=unit, file=file, access='STREAM', form='UNFORMATTED', iomsg=iomsg_, iostat=iostat_)
         inquire(file=file, size=filesize)
         if (allocated(self%raw)) deallocate(self%raw)
         allocate(character(len=filesize):: self%raw)
         read(unit=unit, iostat=iostat_, iomsg=iomsg_) self%raw
         close(unit)
      else
         form_ = 'FORMATTED' ; if (present(form)) form_ = form ; form_ = form_%upper()
         select case(form_%chars())
         case('FORMATTED')
            open(newunit=unit, file=file, status='OLD', action='READ', iomsg=iomsg_, iostat=iostat_, err=10)
         case('UNFORMATTED')
            open(newunit=unit, file=file, status='OLD', action='READ', form='UNFORMATTED', access='STREAM', &
                 iomsg=iomsg_, iostat=iostat_, err=10)
         endselect
         call self%read_lines(unit=unit, form=form, iomsg=iomsg_, iostat=iostat_)
         10 close(unit)
      endif
   else
      iostat_ = 1
      iomsg_ = 'file not found'
   endif
   if (present(iostat)) iostat = iostat_
   if (present(iomsg)) iomsg = iomsg_
   endsubroutine read_file

   subroutine read_line(self, unit, form, iostat, iomsg)
   !< Read line (record) from a connected unit.
   !<
   !< The line is read as an ascii stream read until the eor is reached.
   !<
   !< @note For unformatted read only `access='stream'` is supported with new_line as line terminator.
   !<
   !<```fortran
   !< type(string)      :: astring
   !< type(string)      :: line(3)
   !< integer           :: iostat
   !< character(len=99) :: iomsg
   !< integer           :: scratch
   !< integer           :: l
   !< logical           :: test_passed(6)
   !< line(1) = ' Hello World!   '
   !< line(2) = 'How are you?  '
   !< line(3) = '   All say: "Fine thanks"'
   !< open(newunit=scratch, status='SCRATCH')
   !< write(scratch, "(A)") line(1)%chars()
   !< write(scratch, "(A)") line(2)%chars()
   !< write(scratch, "(A)") line(3)%chars()
   !< rewind(scratch)
   !< l = 0
   !< iostat = 0
   !< do
   !<   l = l + 1
   !<   call astring%read_line(unit=scratch, iostat=iostat, iomsg=iomsg)
   !<   if (iostat/=0.and..not.is_iostat_eor(iostat)) then
   !<     exit
   !<   else
   !<     test_passed(l) = (astring==line(l))
   !<   endif
   !< enddo
   !< close(scratch)
   !< open(newunit=scratch, status='SCRATCH', form='UNFORMATTED', access='STREAM')
   !< write(scratch) line(1)%chars()//new_line('a')
   !< write(scratch) line(2)%chars()//new_line('a')
   !< write(scratch) line(3)%chars()//new_line('a')
   !< rewind(scratch)
   !< l = 0
   !< iostat = 0
   !< do
   !<   l = l + 1
   !<   call astring%read_line(unit=scratch, iostat=iostat, iomsg=iomsg, form='UnfORMatteD')
   !<   if (iostat/=0.and..not.is_iostat_eor(iostat)) then
   !<     exit
   !<   else
   !<     test_passed(l+3) = (astring==line(l))
   !<   endif
   !< enddo
   !< close(scratch)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),    intent(inout)           :: self    !< The string.
   integer,          intent(in)              :: unit    !< Logical unit.
   character(len=*), intent(in),    optional :: form    !< Format of unit.
   integer,          intent(out),   optional :: iostat  !< IO status code.
   character(len=*), intent(inout), optional :: iomsg   !< IO status message.
   type(string)                              :: form_   !< Format of unit, local variable.
   integer                                   :: iostat_ !< IO status code, local variable.
   character(len=:),          allocatable    :: iomsg_  !< IO status message, local variable.
   character(kind=CK, len=:), allocatable    :: line    !< Line storage.
   character(kind=CK, len=1)                 :: ch      !< Character storage.

   form_ = 'FORMATTED' ; if (present(form)) form_ = form ; form_ = form_%upper()
   iomsg_ = repeat(' ', 99) ; if (present(iomsg)) iomsg_ = iomsg
   line = ''
   select case(form_%chars())
   case('FORMATTED')
      do
         read(unit, "(A)", advance='no', iostat=iostat_, iomsg=iomsg_, err=10, end=10, eor=10) ch
         line = line//ch
      enddo
   case('UNFORMATTED')
      do
         read(unit, iostat=iostat_, iomsg=iomsg_, err=10, end=10) ch
         if (ch==new_line('a')) then
            iostat_ = iostat_eor
            exit
         endif
         line = line//ch
      enddo
   endselect
   10 if (line/='') self%raw = line
   if (present(iostat)) iostat = iostat_
   if (present(iomsg)) iomsg = iomsg_
   endsubroutine read_line

   subroutine read_lines(self, unit, form, iostat, iomsg)
   !< Read (all) lines (records) from a connected unit as a single ascii stream.
   !<
   !< @note All the lines are stored into the string self as a single ascii stream. Each line (record) is separated by a `new_line`
   !< character. The line is read as an ascii stream read until the eor is reached.
   !<
   !< @note The connected unit is rewinded. At a successful exit current record is at eof, at the beginning otherwise.
   !<
   !< @note For unformatted read only `access='stream'` is supported with new_line as line terminator.
   !<
   !<```fortran
   !< type(string)              :: astring
   !< type(string), allocatable :: strings(:)
   !< type(string)              :: line(3)
   !< integer                   :: iostat
   !< character(len=99)         :: iomsg
   !< integer                   :: scratch
   !< integer                   :: l
   !< logical                   :: test_passed(8)
   !<
   !< line(1) = ' Hello World!   '
   !< line(2) = 'How are you?  '
   !< line(3) = '   All say: "Fine thanks"'
   !< open(newunit=scratch, status='SCRATCH')
   !< write(scratch, "(A)") line(1)%chars()
   !< write(scratch, "(A)") line(2)%chars()
   !< write(scratch, "(A)") line(3)%chars()
   !< call astring%read_lines(unit=scratch, iostat=iostat, iomsg=iomsg)
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(1) = (size(strings, dim=1)==size(line, dim=1))
   !< do l=1, size(strings, dim=1)
   !<   test_passed(l+1) = (strings(l)==line(l))
   !< enddo
   !< close(scratch)
   !< open(newunit=scratch, status='SCRATCH', form='UNFORMATTED', access='STREAM')
   !< write(scratch) line(1)%chars()//new_line('a')
   !< write(scratch) line(2)%chars()//new_line('a')
   !< write(scratch) line(3)%chars()//new_line('a')
   !< call astring%read_lines(unit=scratch, form='unformatted', iostat=iostat, iomsg=iomsg)
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(5) = (size(strings, dim=1)==size(line, dim=1))
   !< do l=1, size(strings, dim=1)
   !<   test_passed(l+5) = (strings(l)==line(l))
   !< enddo
   !< close(scratch)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),    intent(inout)           :: self    !< The string.
   integer,          intent(in)              :: unit    !< Logical unit.
   character(len=*), intent(in),    optional :: form    !< Format of unit.
   integer,          intent(out),   optional :: iostat  !< IO status code.
   character(len=*), intent(inout), optional :: iomsg   !< IO status message.
   integer                                   :: iostat_ !< IO status code, local variable.
   character(len=:), allocatable             :: iomsg_  !< IO status message, local variable.
   type(string)                              :: lines   !< Lines storage.
   type(string)                              :: line    !< Line storage.

   iomsg_ = repeat(' ', 99) ; if (present(iomsg)) iomsg_ = iomsg
   rewind(unit)
   iostat_ = 0
   lines%raw = ''
   do
      line%raw = ''
      call line%read_line(unit=unit, form=form, iostat=iostat_, iomsg=iomsg_)
      if (iostat_/=0.and..not.is_iostat_eor(iostat_)) then
         exit
      elseif (line/='') then
         lines%raw = lines%raw//line%raw//new_line('a')
      endif
   enddo
   if (lines%raw/='') self%raw = lines%raw
   if (present(iostat)) iostat = iostat_
   if (present(iomsg)) iomsg = iomsg_
   endsubroutine read_lines

   elemental function replace(self, old, new, count) result(replaced)
   !< Return a string with all occurrences of substring old replaced by new.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(3)
   !< astring = 'When YOU are sad YOU should think to me :-)'
   !< test_passed(1) = (astring%replace(old='YOU', new='THEY')//''=='When THEY are sad THEY should think to me :-)')
   !< test_passed(2) = (astring%replace(old='YOU', new='THEY', count=1)//''=='When THEY are sad YOU should think to me :-)')
   !< astring = repeat(new_line('a')//'abcd', 20)
   !< astring = astring%replace(old=new_line('a'), new='|cr|')
   !< astring = astring%replace(old='|cr|', new=new_line('a')//'    ')
   !< test_passed(3) = (astring//''==repeat(new_line('a')//'    '//'abcd', 20))
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self     !< The string.
   character(kind=CK, len=*), intent(in)           :: old      !< Old substring.
   character(kind=CK, len=*), intent(in)           :: new      !< New substring.
   integer,                   intent(in), optional :: count    !< Number of old occurences to be replaced.
   type(string)                                    :: replaced !< The string with old replaced by new.
   integer                                         :: r        !< Counter.

   if (allocated(self%raw)) then
      replaced = self
      r = 0
      do
         if (index(replaced%raw, old)>0) then
            replaced = replaced%replace_one_occurrence(old=old, new=new)
            r = r + 1
            if (present(count)) then
               if (r>=count) exit
            endif
         else
            exit
         endif
      enddo
   endif
   endfunction replace

   elemental function reverse(self) result(reversed)
   !< Return a reversed string.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(2)
   !< astring = 'abcdefghilmnopqrstuvz'
   !< test_passed(1) = (astring%reverse()//''=='zvutsrqponmlihgfedcba')
   !< astring = '0123456789'
   !< test_passed(2) = (astring%reverse()//''=='9876543210')
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   type(string)              :: reversed !< The reversed string.
   integer                   :: length   !< Length of the string.
   integer                   :: c        !< Counter.

   if (allocated(self%raw)) then
      reversed = self
      length = len(self%raw)
      do c=1, length
         reversed%raw(c:c) = self%raw(length-c+1:length-c+1)
      enddo
   endif
   endfunction reverse

   function search(self, tag_start, tag_end, in_string, in_character, istart, iend) result(tag)
   !< Search for *tagged* record into string, return the first record found (if any) matching the tags.
   !<
   !< Optionally, returns the indexes of tag start/end, thus this is not an `elemental` function.
   !<
   !< @note The tagged record is searched into self if allocated otherwise into `in_string` if passed or, eventually, into
   !< `in_character` is passed. If tag is not found the return string is not allocated and the start/end indexes (if requested) are
   !< zero.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< type(string)                  :: anotherstring
   !< character(len=:), allocatable :: acharacter
   !< integer                       :: istart
   !< integer                       :: iend
   !< logical                       :: test_passed(5)
   !< astring = '<test> <first> hello </first> <first> not the first </first> </test>'
   !< anotherstring = astring%search(tag_start='<first>', tag_end='</first>')
   !< test_passed(1) = anotherstring//''=='<first> hello </first>'
   !< astring = '<test> <a> <a> <a> the nested a </a> </a> </a> </test>'
   !< anotherstring = astring%search(tag_start='<a>', tag_end='</a>')
   !< test_passed(2) = anotherstring//''=='<a> <a> <a> the nested a </a> </a> </a>'
   !< call astring%free
   !< anotherstring = '<test> <a> <a> <a> the nested a </a> </a> </a> </test>'
   !< astring = astring%search(in_string=anotherstring, tag_start='<a>', tag_end='</a>')
   !< test_passed(3) = astring//''=='<a> <a> <a> the nested a </a> </a> </a>'
   !< call astring%free
   !< acharacter = '<test> <a> <a> <a> the nested a </a> </a> </a> </test>'
   !< astring = astring%search(in_character=acharacter, tag_start='<a>', tag_end='</a>')
   !< test_passed(4) = astring//''=='<a> <a> <a> the nested a </a> </a> </a>'
   !< acharacter = '<test> <first> hello </first> <sec> <sec>not the first</sec> </sec> </test>'
   !< astring = astring%search(in_character=acharacter, tag_start='<sec>', tag_end='</sec>', istart=istart, iend=iend)
   !< test_passed(5) = astring//''==acharacter(31:67)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)            :: self         !< The string.
   character(kind=CK, len=*), intent(in)            :: tag_start    !< Start tag.
   character(kind=CK, len=*), intent(in)            :: tag_end      !< End tag.
   type(string),              intent(in),  optional :: in_string    !< Search into this string.
   character(kind=CK, len=*), intent(in),  optional :: in_character !< Search into this character string.
   integer,                   intent(out), optional :: istart       !< Starting index of tag inside the string.
   integer,                   intent(out), optional :: iend         !< Ending index of tag inside the string.
   type(string)                                     :: tag          !< First tag found.
   character(kind=CK, len=:), allocatable           :: raw          !< Raw string into which search the tag.
   integer                                          :: istart_      !< Starting index of tag inside the string, local variable.
   integer                                          :: iend_        !< Ending index of tag inside the string, local variable.
   integer                                          :: nested_tags  !< Number of nested tags inside tag.
   integer                                          :: t            !< Counter.

   raw = ''
   if (present(in_string)) then
      raw = in_string%raw
   elseif (present(in_character)) then
      raw = in_character
   else
      if (allocated(self%raw)) raw = self%raw
   endif
   istart_ = 0
   iend_ = 0
   if (raw/='') then
      istart_ = index(raw, tag_start)
      iend_ = index(raw, tag_end)
      if (istart_>0.and.iend_>0) then
         iend_ = iend_ + len(tag_end) - 1
         tag%raw = raw(istart_:iend_)
         nested_tags = tag%count(tag_start)
         if (nested_tags>1) then
            do t=2, nested_tags
               iend_ = iend_ + len(tag_end) - 1 + index(raw(iend_+1:), tag_end)
            enddo
            tag%raw = raw(istart_:iend_)
         endif
      endif
   endif
   if (present(istart)) istart = istart_
   if (present(iend)) iend = iend_
   endfunction search

   pure function slice(self, istart, iend) result(raw)
   !< Return the raw characters data sliced.
   !<
   !<```fortran
   !< type(string) :: astring
   !< astring = 'the Quick Brown fox Jumps over the Lazy Dog.'
   !< print "(A)", astring%slice(11,25)
   !<```
   !=> Brown fox Jumps <<<
   class(string), intent(in)              :: self   !< The string.
   integer,       intent(in)              :: istart !< Slice start index.
   integer,       intent(in)              :: iend   !< Slice end   index.
   character(kind=CK, len=:), allocatable :: raw    !< Raw characters data.

   if (allocated(self%raw)) then
      raw = self%raw(istart:iend)
   else
      raw = ''
   endif
   endfunction slice

   elemental function snakecase(self, sep)
   !< Return a string with all words lowercase separated by "_".
   !<
   !< @note Multiple subsequent separators are collapsed to one occurence.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 'the Quick Brown fox Jumps over the Lazy Dog.'
   !< test_passed(1) = astring%snakecase()//''=='the_quick_brown_fox_jumps_over_the_lazy_dog.'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   type(string)                                    :: snakecase !< Snake case string.
   type(string), allocatable                       :: tokens(:) !< String tokens.

   if (allocated(self%raw)) then
      call self%split(tokens=tokens, sep=sep)
      tokens = tokens%lower()
      snakecase = snakecase%join(array=tokens, sep='_')
   endif
   endfunction snakecase

   pure subroutine split(self, tokens, sep, max_tokens)
   !< Return a list of substring in the string, using sep as the delimiter string.
   !<
   !< @note Multiple subsequent separators are collapsed to one occurrence.
   !<
   !< @note If `max_tokens` is passed the returned number of tokens is either `max_tokens` or `max_tokens + 1`.
   !<
   !<```fortran
   !< type(string)              :: astring
   !< type(string), allocatable :: strings(:)
   !< logical                   :: test_passed(11)
   !< astring = '+ab-++cre-++cre-ab+'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(1) = (strings(1)//''=='ab-'.and.strings(2)//''=='cre-'.and.strings(3)//''=='cre-ab')
   !< astring = 'ab-++cre-++cre-ab+'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(2) = (strings(1)//''=='ab-'.and.strings(2)//''=='cre-'.and.strings(3)//''=='cre-ab')
   !< astring = 'ab-++cre-++cre-ab'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(3) = (strings(1)//''=='ab-'.and.strings(2)//''=='cre-'.and.strings(3)//''=='cre-ab')
   !< astring = 'Hello '//new_line('a')//'World!'
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(4) = (strings(1)//''=='Hello '.and.strings(2)//''=='World!')
   !< astring = 'Hello World!'
   !< call astring%split(tokens=strings)
   !< test_passed(5) = (strings(1)//''=='Hello'.and.strings(2)//''=='World!')
   !< astring = '+ab-'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(6) = (strings(1)//''=='ab-')
   !< astring = '+ab-'
   !< call astring%split(tokens=strings, sep='-')
   !< test_passed(7) = (strings(1)//''=='+ab')
   !< astring = '+ab-+cd-'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(8) = (strings(1)//''=='ab-'.and.strings(2)//''=='cd-')
   !< astring = 'ab-+cd-+'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(9) = (strings(1)//''=='ab-'.and.strings(2)//''=='cd-')
   !< astring = '+ab-+cd-+'
   !< call astring%split(tokens=strings, sep='+')
   !< test_passed(10) = (strings(1)//''=='ab-'.and.strings(2)//''=='cd-')
   !< astring = '1-2-3-4-5-6-7-8'
   !< call astring%split(tokens=strings, sep='-', max_tokens=3)
   !< test_passed(11) = (strings(1)//''=='1'.and.strings(2)//''=='2'.and.strings(3)//''=='3'.and.strings(4)//''=='4-5-6-7-8')
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self           !< The string.
   type(string), allocatable, intent(out)          :: tokens(:)      !< Tokens substring.
   character(kind=CK, len=*), intent(in), optional :: sep            !< Separator.
   integer,                   intent(in), optional :: max_tokens     !< Fix the maximum number of returned tokens.
   character(kind=CK, len=:), allocatable          :: sep_           !< Separator, default value.
   integer                                         :: No             !< Number of occurrences of sep.
   integer                                         :: t              !< Character counter.
   type(string)                                    :: temporary      !< Temporary storage.
   type(string), allocatable                       :: temp_toks(:,:) !< Temporary tokens substring.

   if (allocated(self%raw)) then
     sep_ = SPACE ; if (present(sep)) sep_ = sep

     temporary = self%unique(sep_)
     No = temporary%count(sep_)

     if (No>0) then
       if (present(max_tokens)) then
         if (max_tokens < No.and.max_tokens > 0) No = max_tokens
       endif
       allocate(temp_toks(3, No))
       temp_toks(:, 1) = temporary%partition(sep_)
       if (No>1) then
         do t=2, No
           temp_toks(:, t) = temp_toks(3, t-1)%partition(sep_)
         enddo
       endif

       if (temp_toks(1, 1)%raw/=''.and.temp_toks(3, No)%raw/='') then
         allocate(tokens(No+1))
         do t=1, No
           if (t==No) then
             tokens(t  ) = temp_toks(1, t)
             tokens(t+1) = temp_toks(3, t)
           else
             tokens(t) = temp_toks(1, t)
           endif
         enddo
       elseif (temp_toks(1, 1)%raw/='') then
         allocate(tokens(No))
         do t=1, No
           tokens(t) = temp_toks(1, t)
         enddo
       elseif (temp_toks(3, No)%raw/='') then
         allocate(tokens(No))
         do t=1, No-1
           tokens(t) = temp_toks(1, t+1)
         enddo
         tokens(No) = temp_toks(3, No)
       else
         allocate(tokens(No-1))
         do t=2, No
           tokens(t-1) = temp_toks(1, t)
         enddo
       endif

     else
       allocate(tokens(1))
       tokens(1) = self
     endif
   endif
   endsubroutine split

   pure subroutine split_chunked(self, tokens, chunks, sep)
   !< Return a list of substring in the string, using sep as the delimiter string, chunked (memory-efficient) algorithm.
   !<
   !< @note Multiple subsequent separators are collapsed to one occurrence.
   !<
   !< @note The split is performed in chunks of `#chunks` to avoid excessive memory consumption.
   !<
   !<```fortran
   !< type(string)              :: astring
   !< type(string), allocatable :: strings(:)
   !< logical                   :: test_passed(1)
   !< astring = '-1-2-3-4-5-6-7-8-'
   !< call astring%split_chunked(tokens=strings, sep='-', chunks=3)
   !< test_passed(1) = (strings(1)//''=='1'.and.strings(2)//''=='2'.and.strings(3)//''=='3'.and.strings(4)//''=='4'.and. &
   !<                   strings(5)//''=='5'.and.strings(6)//''=='6'.and.strings(7)//''=='7'.and.strings(8)//''=='8')
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   type(string), allocatable, intent(out)          :: tokens(:) !< Tokens substring.
   integer,                   intent(in)           :: chunks    !< Number of chunks.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   integer                                         :: Nt        !< Number of actual tokens.
   integer                                         :: t         !< Counter.
   logical                                         :: isok

   if (allocated(self%raw)) then
     sep_ = SPACE ; if (present(sep)) sep_ = sep

     Nt = self%count(sep_)
     if (self%start_with(prefix=sep_)) Nt = Nt - 1
     if (self%end_with(suffix=sep_)) Nt = Nt - 1
     t = 0
     call self%split(tokens=tokens, sep=sep_, max_tokens=chunks)
     do
       t = size(tokens, dim=1)
       if (t > Nt) exit
       call split_last_token(tokens=tokens, max_tokens=chunks,isok=isok)
       if(isok)then
       else
            exit
       endif
     enddo

     t = size(tokens, dim=1)
     if (tokens(t)%count(sep_) > 0) then
        call split_last_token(tokens=tokens,isok=isok)
     endif
   endif

   contains
      pure subroutine split_last_token(tokens, max_tokens,isok)
      !< Split last token.
      type(string), allocatable, intent(inout)        :: tokens(:)      !< Tokens substring.
      integer,                   intent(in), optional :: max_tokens     !< Max tokens returned.
      type(string), allocatable                       :: tokens_(:)     !< Temporary tokens.
      type(string), allocatable                       :: tokens_swap(:) !< Swap tokens.
      integer                                         :: Nt_            !< Number of last created tokens.
      logical,intent(out)                             :: isok

      isok=.true.
      call tokens(t)%split(tokens=tokens_, sep=sep_, max_tokens=max_tokens)
      if (allocated(tokens_)) then
        Nt_ = size(tokens_, dim=1)
        if (Nt_ >= 1) then
          allocate(tokens_swap(1:t-1+Nt_))
          tokens_swap(1:t-1) = tokens(1:t-1)
          tokens_swap(t:)    = tokens_(:)
          call move_alloc(from=tokens_swap, to=tokens)
        endif
        if (Nt_ == 1) then
            isok=.false.
        end if
        deallocate(tokens_)
      endif
      endsubroutine split_last_token
   endsubroutine split_chunked

   elemental function startcase(self, sep)
   !< Return a string with all words capitalized, e.g. title case.
   !<
   !< @note Multiple subsequent separators are collapsed to one occurence.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 'the Quick Brown fox Jumps over the Lazy Dog.'
   !< test_passed(1) = astring%startcase()//''=='The Quick Brown Fox Jumps Over The Lazy Dog.'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self      !< The string.
   character(kind=CK, len=*), intent(in), optional :: sep       !< Separator.
   type(string)                                    :: startcase !< Start case string.
   character(kind=CK, len=:), allocatable          :: sep_      !< Separator, default value.
   type(string), allocatable                       :: tokens(:) !< String tokens.

   if (allocated(self%raw)) then
      sep_ = SPACE ; if (present(sep)) sep_ = sep
      call self%split(tokens=tokens, sep=sep_)
      tokens = tokens%capitalize()
      startcase = startcase%join(array=tokens, sep=sep_)
   endif
   endfunction startcase

   elemental function strip(self, remove_nulls)
   !< Return a copy of the string with the leading and trailing characters removed.
   !<
   !< @note Multiple subsequent separators are collapsed to one occurence.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = '  Hello World!   '
   !< test_passed(1) = astring%strip()//''=='Hello World!'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self         !< The string.
   logical,       intent(in), optional :: remove_nulls !< Remove null characters at the end.
   type(string)                        :: strip        !< The stripped string.
   integer                             :: c            !< Counter.

   if (allocated(self%raw)) then
      strip = self%adjustl()
      strip = strip%trim()
      if (present(remove_nulls)) then
         if (remove_nulls) then
            c = index(self%raw, char(0))
            if (c>0) strip%raw = strip%raw(1:c-1)
         endif
      endif
   endif
   endfunction strip

   elemental function swapcase(self)
   !< Return a copy of the string with uppercase characters converted to lowercase and vice versa.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = '  Hello World!   '
   !< test_passed(1) = astring%swapcase()//''=='  hELLO wORLD!   '
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   type(string)              :: swapcase !< Upper case string.
   integer                   :: n1       !< Characters counter.
   integer                   :: n2       !< Characters counter.

   if (allocated(self%raw)) then
      swapcase = self
      do n1=1, len(self%raw)
         n2 = index(UPPER_ALPHABET, self%raw(n1:n1))
         if (n2>0) then
            swapcase%raw(n1:n1) = LOWER_ALPHABET(n2:n2)
         else
            n2 = index(LOWER_ALPHABET, self%raw(n1:n1))
            if (n2>0) swapcase%raw(n1:n1) = UPPER_ALPHABET(n2:n2)
         endif
      enddo
   endif
   endfunction swapcase

   function tempname(self, is_file, prefix, path)
   !< Return a safe temporary name suitable for temporary file or directories.
   !<
   !<```fortran
   !< type(string) :: astring
   !< character(len=:), allocatable :: tmpname
   !< logical                       :: test_passed(5)
   !< tmpname = astring%tempname()
   !< inquire(file=tmpname, exist=test_passed(1))
   !< test_passed(1) = .not.test_passed(1)
   !< tmpname = astring%tempname(is_file=.false.)
   !< inquire(file=tmpname, exist=test_passed(2))
   !< test_passed(2) = .not.test_passed(2)
   !< tmpname = astring%tempname(path='./')
   !< inquire(file=tmpname, exist=test_passed(3))
   !< test_passed(3) = .not.test_passed(3)
   !< astring = 'me-'
   !< tmpname = astring%tempname()
   !< inquire(file=tmpname, exist=test_passed(4))
   !< test_passed(4) = .not.test_passed(4)
   !< tmpname = astring%tempname(prefix='you-')
   !< inquire(file=tmpname, exist=test_passed(5))
   !< test_passed(5) = .not.test_passed(5)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self                   !< The string.
   logical,       intent(in), optional :: is_file                !< True if tempname should be used for file (the default).
   character(*),  intent(in), optional :: prefix                 !< Name prefix, otherwise self is used (if allocated).
   character(*),  intent(in), optional :: path                   !< Path where file/directory should be used, default `./`.
   character(len=:), allocatable       :: tempname               !< Safe (unique) temporary name.
   logical                             :: is_file_               !< True if tempname should be used for file (the default).
   character(len=:), allocatable       :: prefix_                !< Name prefix, otherwise self is used (if allocated).
   character(len=:), allocatable       :: path_                  !< Path where file/directory should be used, default `./`.
   logical, save                       :: is_initialized=.false. !< Status of random seed initialization.
   real(R4P)                           :: random_real            !< Random number (real).
   integer(I4P)                        :: random_integer         !< Random number (integer).
   logical                             :: is_hold                !< Flag to check if a safe tempname has been found.

   is_file_ = .true. ; if (present(is_file)) is_file_ = is_file
   path_ = '' ; if (present(path)) path_ = path
   prefix_ = ''
   if (present(prefix)) then
      prefix_ = prefix
   elseif (allocated(self%raw)) then
      prefix_ = self%raw
   endif
   if (.not.is_initialized) then
      call random_seed
      is_initialized = .true.
   endif
   tempname = repeat(' ', len(path_) + len(prefix_) + 10) ! [path_] + [prefix_] + 6 random chars + [.tmp]
   do
      call random_number(random_real)
      random_integer = transfer(random_real, random_integer)
      random_integer = iand(random_integer, 16777215_I4P)
      if (is_file_)  then
         write(tempname, '(A,Z6.6,A)') path_//prefix_, random_integer, '.tmp'
      else
         write(tempname, '(A,Z6.6)') path_//prefix_, random_integer
         tempname = trim(tempname)
      endif
      inquire(file=tempname, exist=is_hold)
      if (.not.is_hold) exit
   enddo
   endfunction tempname

   elemental function to_integer_I1P(self, kind) result(to_number)
   !< Cast string to integer (I1P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< integer(I1P) :: integer_
   !< logical      :: test_passed(1)
   !< astring = '127'
   !< integer_ = astring%to_number(kind=1_I1P)
   !< test_passed(1) = integer_==127_I1P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   integer(I1P),  intent(in) :: kind      !< Mold parameter for kind detection.
   integer(I1P)              :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_integer()) read(self%raw, *) to_number
   endif
   endfunction to_integer_I1P

#ifndef _NVF
   elemental function to_integer_I2P(self, kind) result(to_number)
   !< Cast string to integer (I2P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< integer(I2P) :: integer_
   !< logical      :: test_passed(1)
   !< astring = '127'
   !< integer_ = astring%to_number(kind=1_I2P)
   !< test_passed(1) = integer_==127_I2P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   integer(I2P),  intent(in) :: kind      !< Mold parameter for kind detection.
   integer(I2P)              :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_integer()) read(self%raw, *) to_number
   endif
   endfunction to_integer_I2P
#endif

   elemental function to_integer_I4P(self, kind) result(to_number)
   !< Cast string to integer (I4P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< integer(I4P) :: integer_
   !< logical      :: test_passed(1)
   !< astring = '127'
   !< integer_ = astring%to_number(kind=1_I4P)
   !< test_passed(1) = integer_==127_I4P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   integer(I4P),  intent(in) :: kind      !< Mold parameter for kind detection.
   integer(I4P)              :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_integer()) read(self%raw, *) to_number
   endif
   endfunction to_integer_I4P

   elemental function to_integer_I8P(self, kind) result(to_number)
   !< Cast string to integer (I8P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< integer(I8P) :: integer_
   !< logical      :: test_passed(1)
   !< astring = '127'
   !< integer_ = astring%to_number(kind=1_I8P)
   !< test_passed(1) = integer_==127_I8P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   integer(I8P),  intent(in) :: kind      !< Mold parameter for kind detection.
   integer(I8P)              :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_integer()) read(self%raw, *) to_number
   endif
   endfunction to_integer_I8P

   elemental function to_real_R4P(self, kind) result(to_number)
   !< Cast string to real (R4P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< real(R4P)    :: real_
   !< logical      :: test_passed(1)
   !< astring = '3.4e9'
   !< real_ = astring%to_number(kind=1._R4P)
   !< test_passed(1) = real_==3.4e9_R4P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   real(R4P),     intent(in) :: kind      !< Mold parameter for kind detection.
   real(R4P)                 :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_real()) read(self%raw, *) to_number
   endif
   endfunction to_real_R4P

   elemental function to_real_R8P(self, kind) result(to_number)
   !< Cast string to real (R8P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< real(R8P)    :: real_
   !< logical      :: test_passed(1)
   !< astring = '3.4e9'
   !< real_ = astring%to_number(kind=1._R8P)
   !< test_passed(1) = real_==3.4e9_R8P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   real(R8P),     intent(in) :: kind      !< Mold parameter for kind detection.
   real(R8P)                 :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_real()) read(self%raw, *) to_number
   endif
   endfunction to_real_R8P

   elemental function to_real_R16P(self, kind) result(to_number)
   !< Cast string to real (R16P).
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< real(R16P)   :: real_
   !< logical      :: test_passed(1)
   !< astring = '3.4e9'
   !< real_ = astring%to_number(kind=1._R16P)
   !< test_passed(1) = real_==3.4e9_R16P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self      !< The string.
   real(R16P),    intent(in) :: kind      !< Mold parameter for kind detection.
   real(R16P)                :: to_number !< The number into the string.

   if (allocated(self%raw)) then
     if (self%is_real()) read(self%raw, *) to_number
   endif
   endfunction to_real_R16P

   elemental function unescape(self, to_unescape, unesc) result(unescaped)
   !< Unescape double backslashes (or custom escaped character).
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(2)
   !< astring = '^\\s \\d+\\s*'
   !< test_passed(1) = (astring%unescape(to_unescape='\')//''=='^\s \d+\s*')
   !< test_passed(2) = (astring%unescape(to_unescape='s')//''=='^\s \\d+\s*')
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self        !< The string.
   character(kind=CK, len=1), intent(in)           :: to_unescape !< Character to be unescaped.
   character(kind=CK, len=*), intent(in), optional :: unesc       !< Character used to unescape.
   type(string)                                    :: unescaped   !< Escaped string.
   character(kind=CK, len=:), allocatable          :: unesc_      !< Character to unescape, local variable.
   integer                                         :: c           !< Character counter.

   if (allocated(self%raw)) then
      unesc_ = '' ; if (present(unesc)) unesc_ = unesc
      unescaped%raw = ''
      c = 1
      do
         if (c>len(self%raw)) exit
         if (c==len(self%raw)) then
            unescaped%raw = unescaped%raw//self%raw(c:c)
            exit
         else
            if (self%raw(c:c+1)==BACKSLASH//to_unescape) then
               unescaped%raw = unescaped%raw//to_unescape
               c = c + 2
            else
               unescaped%raw = unescaped%raw//self%raw(c:c)
               c = c + 1
            endif
         endif
      enddo
   endif
   endfunction unescape

   elemental function unique(self, substring) result(uniq)
   !< Reduce to one (unique) multiple (sequential) occurrences of a substring into a string.
   !<
   !< For example the string ' ab-cre-cre-ab' is reduce to 'ab-cre-ab' if the substring is '-cre'.
   !< @note Eventual multiple trailing white space are not reduced to one occurrence.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = '+++ab-++cre-++cre-ab+++++'
   !< test_passed(1) = astring%unique(substring='+')//''=='+ab-+cre-+cre-ab+'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self       !< The string.
   character(kind=CK, len=*), intent(in), optional :: substring  !< Substring which multiple occurences must be reduced to one.
   character(kind=CK, len=:), allocatable          :: substring_ !< Substring, default value.
   type(string)                                    :: uniq       !< String parsed.
#ifdef _NVF
   character(9999)                                 :: nvf_bug  !< Work around for NVFortran bug.
#endif

   if (allocated(self%raw)) then
     substring_ = SPACE ; if (present(substring)) substring_ = substring

     uniq = self
     do
#ifdef _NVF
       nvf_bug = substring_
       if (.not.uniq%index(repeat(trim(nvf_bug), 2))>0) exit
       uniq = uniq%replace(old=repeat(trim(nvf_bug), 2), new=substring_)
#else
       if (.not.uniq%index(repeat(substring_, 2))>0) exit
       uniq = uniq%replace(old=repeat(substring_, 2), new=substring_)
#endif
     enddo
   endif
   endfunction unique

   elemental function upper(self)
   !< Return a string with all uppercase characters.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 'Hello WorLD!'
   !< test_passed(1) = astring%upper()//''=='HELLO WORLD!'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self  !< The string.
   type(string)              :: upper !< Upper case string.
   integer                   :: n1    !< Characters counter.
   integer                   :: n2    !< Characters counter.

   if (allocated(self%raw)) then
      upper = self
      do n1=1, len(self%raw)
         n2 = index(LOWER_ALPHABET, self%raw(n1:n1))
         if (n2>0) upper%raw(n1:n1) = UPPER_ALPHABET(n2:n2)
      enddo
   endif
   endfunction upper

   subroutine write_file(self, file, form, iostat, iomsg)
   !< Write a single string stream into file.
   !<
   !< @note For unformatted read only `access='stream'` is supported with new_line as line terminator.
   !<
   !<```fortran
   !< type(string)              :: astring
   !< type(string)              :: anotherstring
   !< type(string), allocatable :: strings(:)
   !< type(string)              :: line(3)
   !< integer                   :: iostat
   !< character(len=99)         :: iomsg
   !< integer                   :: scratch
   !< integer                   :: l
   !< logical                   :: test_passed(8)
   !< line(1) = ' Hello World!   '
   !< line(2) = 'How are you?  '
   !< line(3) = '   All say: "Fine thanks"'
   !< anotherstring = anotherstring%join(array=line, sep=new_line('a'))
   !< call anotherstring%write_file(file='write_file_test.tmp', iostat=iostat, iomsg=iomsg)
   !< call astring%read_file(file='write_file_test.tmp', iostat=iostat, iomsg=iomsg)
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(1) = (size(strings, dim=1)==size(line, dim=1))
   !< do l=1, size(strings, dim=1)
   !<   test_passed(l+1) = (strings(l)==line(l))
   !< enddo
   !< call anotherstring%write_file(file='write_file_test.tmp', form='unformatted', iostat=iostat, iomsg=iomsg)
   !< call astring%read_file(file='write_file_test.tmp', form='unformatted', iostat=iostat, iomsg=iomsg)
   !< call astring%split(tokens=strings, sep=new_line('a'))
   !< test_passed(5) = (size(strings, dim=1)==size(line, dim=1))
   !< do l=1, size(strings, dim=1)
   !<   test_passed(l+5) = (strings(l)==line(l))
   !< enddo
   !< open(newunit=scratch, file='write_file_test.tmp')
   !< close(unit=scratch, status='delete')
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),    intent(in)              :: self    !< The string.
   character(len=*), intent(in)              :: file    !< File name.
   character(len=*), intent(in),    optional :: form    !< Format of unit.
   integer,          intent(out),   optional :: iostat  !< IO status code.
   character(len=*), intent(inout), optional :: iomsg   !< IO status message.
   type(string)                              :: form_   !< Format of unit, local variable.
   integer                                   :: iostat_ !< IO status code, local variable.
   character(len=:), allocatable             :: iomsg_  !< IO status message, local variable.
   integer                                   :: unit    !< Logical unit.

   iomsg_ = repeat(' ', 99) ; if (present(iomsg)) iomsg_ = iomsg
   form_ = 'FORMATTED' ; if (present(form)) form_ = form ; form_ = form_%upper()
   select case(form_%chars())
   case('FORMATTED')
      open(newunit=unit, file=file, action='WRITE', iomsg=iomsg_, iostat=iostat_, err=10)
   case('UNFORMATTED')
      open(newunit=unit, file=file, action='WRITE', form='UNFORMATTED', access='STREAM', iomsg=iomsg_, iostat=iostat_, err=10)
   endselect
   call self%write_lines(unit=unit, form=form, iomsg=iomsg_, iostat=iostat_)
   10 close(unit)
   if (present(iostat)) iostat = iostat_
   if (present(iomsg)) iomsg = iomsg_
   endsubroutine write_file

   subroutine write_line(self, unit, form, iostat, iomsg)
   !< Write line (record) to a connected unit.
   !<
   !< @note If the connected unit is unformatted a `new_line()` character is added at the end (if necessary) to mark the end of line.
   !<
   !< @note There is no doctests, this being tested by means of [[string:write_file]] doctests.
   class(string),    intent(in)              :: self    !< The string.
   integer,          intent(in)              :: unit    !< Logical unit.
   character(len=*), intent(in),    optional :: form    !< Format of unit.
   integer,          intent(out),   optional :: iostat  !< IO status code.
   character(len=*), intent(inout), optional :: iomsg   !< IO status message.
   type(string)                              :: form_   !< Format of unit, local variable.
   integer                                   :: iostat_ !< IO status code, local variable.
   character(len=:), allocatable             :: iomsg_  !< IO status message, local variable.

   iostat_ = 0
   iomsg_ = repeat(' ', 99) ; if (present(iomsg)) iomsg_ = iomsg
   if (allocated(self%raw)) then
      form_ = 'FORMATTED' ; if (present(form)) form_ = form ; form_ = form_%upper()
      select case(form_%chars())
      case('FORMATTED')
         write(unit, "(A)", iostat=iostat_, iomsg=iomsg_) self%raw
      case('UNFORMATTED')
         if (self%end_with(new_line('a'))) then
            write(unit, iostat=iostat_, iomsg=iomsg_) self%raw
         else
            write(unit, iostat=iostat_, iomsg=iomsg_) self%raw//new_line('a')
         endif
      endselect
   endif
   if (present(iostat)) iostat = iostat_
   if (present(iomsg)) iomsg = iomsg_
   endsubroutine write_line

   subroutine write_lines(self, unit, form, iostat, iomsg)
   !< Write lines (records) to a connected unit.
   !<
   !< This method checks if self contains more than one line (records) and writes them as lines (records).
   !<
   !< @note If the connected unit is unformatted a `new_line()` character is added at the end (if necessary) to mark the end of line.
   !<
   !< @note There is no doctests, this being tested by means of [[string:write_file]] doctests.
   class(string),    intent(in)              :: self     !< The string.
   integer,          intent(in)              :: unit     !< Logical unit.
   character(len=*), intent(in),    optional :: form     !< Format of unit.
   integer,          intent(out),   optional :: iostat   !< IO status code.
   character(len=*), intent(inout), optional :: iomsg    !< IO status message.
   type(string), allocatable                 :: lines(:) !< Lines.
   integer                                   :: l        !< Counter.

   if (allocated(self%raw)) then
      call self%split(tokens=lines, sep=new_line('a'))
      do l=1, size(lines, dim=1)
         call lines(l)%write_line(unit=unit, form=form, iostat=iostat, iomsg=iomsg)
      enddo
   endif
   endsubroutine write_lines

   ! inquire
   elemental function end_with(self, suffix, start, end, ignore_null_eof)
   !< Return true if a string ends with a specified suffix.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(5)
   !< astring = 'Hello WorLD!'
   !< test_passed(1) = astring%end_with(suffix='LD!').eqv..true.
   !< test_passed(2) = astring%end_with(suffix='lD!').eqv..false.
   !< test_passed(3) = astring%end_with(suffix='orLD!', start=5).eqv..true.
   !< test_passed(4) = astring%end_with(suffix='orLD!', start=8, end=12).eqv..true.
   !< test_passed(5) = astring%end_with(suffix='!').eqv..true.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self             !< The string.
   character(kind=CK, len=*), intent(in)           :: suffix           !< Searched suffix.
   integer,                   intent(in), optional :: start            !< Start position into the string.
   integer,                   intent(in), optional :: end              !< End position into the string.
   logical,                   intent(in), optional :: ignore_null_eof  !< Ignore null character at the end of file.
   logical                                         :: end_with         !< Result of the test.
   integer                                         :: start_           !< Start position into the string, local variable.
   integer                                         :: end_             !< End position into the string, local variable.
   logical                                         :: ignore_null_eof_ !< Ignore null character at the end of file, local variable.

   end_with = .false.
   if (allocated(self%raw)) then
      start_           = 1             ; if (present(start))           start_           = start
      end_             = len(self%raw) ; if (present(end))             end_             = end
      ignore_null_eof_ = .false.       ; if (present(ignore_null_eof)) ignore_null_eof_ = ignore_null_eof
      if (ignore_null_eof_.and.(self%raw(end_:end_) == char(0))) end_ = end_ - 1
      if (len(suffix) <= len(self%raw(start_:end_))) then
         end_with = self%raw(end_-len(suffix)+1:end_) == suffix
      endif
   endif
   endfunction end_with

   elemental function is_allocated(self)
   !< Return true if the string is allocated.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(2)
   !< test_passed(1) = astring%is_allocated().eqv..false.
   !< astring = 'hello'
   !< test_passed(2) = astring%is_allocated().eqv..true.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   logical                   :: is_allocated !< Result of the test.

   is_allocated = allocated(self%raw)
   endfunction is_allocated

   elemental function is_digit(self)
   !< Return true if all characters in the string are digits.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(2)
   !< astring = '   -1212112.3 '
   !< test_passed(1) = astring%is_digit().eqv..false.
   !< astring = '12121123'
   !< test_passed(2) = astring%is_digit().eqv..true.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   logical                   :: is_digit !< Result of the test.
   integer                   :: c        !< Character counter.

   is_digit = .false.
   if (allocated(self%raw)) then
      do c=1, len(self%raw)
         select case (self%raw(c:c))
         case ('0':'9')
            is_digit = .true.
         case default
            is_digit = .false.
            exit
         end select
      enddo
   endif
   endfunction is_digit

   elemental function is_integer(self, allow_spaces)
   !< Return true if the string contains an integer.
   !<
   !< The regular expression is `\s*[\+\-]?\d+([eE]\+?\d+)?\s*`. The parse algorithm is done in stages:
   !<
   !< | S0  | S1      | S2  | S3   | S4  | S5  | S6  |
   !< |-----|---------|-----|------|-----|-----|-----|
   !< |`\s*`|`[\+\-]?`|`\d+`|`[eE]`|`\+?`|`\d+`|`\s*`|
   !<
   !< Exit on stages-parsing results in:
   !<
   !< | S0 | S1 | S2 | S3 | S4 | S5 | S6 |
   !< |----|----|----|----|----|----|----|
   !< |  F |  F |  T |  F |  F |  T |  T |
   !<
   !< @note This implementation is courtesy of
   !< [tomedunn](https://github.com/tomedunn/fortran-string-utility-module/blob/master/src/string_utility_module.f90#L294)
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(6)
   !< astring = '   -1212112 '
   !< test_passed(1) = astring%is_integer().eqv..true.
   !< astring = '   -1212112'
   !< test_passed(2) = astring%is_integer(allow_spaces=.false.).eqv..false.
   !< astring = '-1212112   '
   !< test_passed(3) = astring%is_integer(allow_spaces=.false.).eqv..false.
   !< astring = '+2e20'
   !< test_passed(4) = astring%is_integer().eqv..true.
   !< astring = ' -2E13 '
   !< test_passed(5) = astring%is_integer().eqv..true.
   !< astring = ' -2 E13 '
   !< test_passed(6) = astring%is_integer().eqv..false.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self          !< The string.
   logical,       intent(in), optional :: allow_spaces  !< Allow leading-trailing spaces.
   logical                             :: is_integer    !< Result of the test.
   logical                             :: allow_spaces_ !< Allow leading-trailing spaces, local variable.
   integer                             :: stage         !< Stages counter.
   integer                             :: c             !< Character counter.

   if (allocated(self%raw)) then
      allow_spaces_ = .true. ; if (present(allow_spaces)) allow_spaces_ = allow_spaces
      stage = 0
      is_integer = .true.
      do c=1, len(self%raw)
         select case(self%raw(c:c))
         case(SPACE, TAB)
            select case(stage)
            case(0, 6)
               is_integer = allow_spaces_
            case(2, 5)
               is_integer = allow_spaces_
               stage = 6
            case default
               is_integer = .false.
            endselect
         case('-')
            select case(stage)
            case(0)
               stage = 1
            case default
               is_integer = .false.
            end select
         case('+')
            select case(stage)
            case(0)
               stage = 1
            case(3)
               stage = 4
            case default
               is_integer = .false.
            endselect
         case('0':'9')
            select case(stage)
            case(0:1)
               stage = 2
            case(3:4)
               stage = 5
            case default
               continue
            endselect
         case ('e','E')
            select case(stage)
            case(2)
               stage = 3
            case default
               is_integer = .false.
            endselect
         case default
            is_integer = .false.
         endselect
         if (.not.is_integer) exit
      enddo
   endif
   if (is_integer) then
      select case(stage)
      case(2, 5, 6)
         is_integer = .true.
      case default
         is_integer = .false.
      end select
   endif
   endfunction is_integer

   elemental function is_lower(self)
   !< Return true if all characters in the string are lowercase.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(3)
   !< astring = ' Hello World'
   !< test_passed(1) = astring%is_lower().eqv..false.
   !< astring = ' HELLO WORLD'
   !< test_passed(2) = astring%is_lower().eqv..false.
   !< astring = ' hello world'
   !< test_passed(3) = astring%is_lower().eqv..true.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   logical                   :: is_lower !< Result of the test.
   integer                   :: c        !< Character counter.

   is_lower = .false.
   if (allocated(self%raw)) then
      is_lower = .true.
      do c=1, len(self%raw)
         if (index(UPPER_ALPHABET, self%raw(c:c))>0) then
            is_lower = .false.
            exit
         endif
      enddo
   endif
   endfunction is_lower

   elemental function is_number(self, allow_spaces)
   !< Return true if the string contains a number (real or integer).
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(7)
   !< astring = '   -1212112 '
   !< test_passed(1) = astring%is_number().eqv..true.
   !< astring = '   -121.2112 '
   !< test_passed(2) = astring%is_number().eqv..true.
   !< astring = '   -1212112'
   !< test_passed(3) = astring%is_number(allow_spaces=.false.).eqv..false.
   !< astring = '-12121.12   '
   !< test_passed(4) = astring%is_number(allow_spaces=.false.).eqv..false.
   !< astring = '+2e20'
   !< test_passed(5) = astring%is_number().eqv..true.
   !< astring = ' -2.4E13 '
   !< test_passed(6) = astring%is_number().eqv..true.
   !< astring = ' -2 E13 '
   !< test_passed(7) = astring%is_number().eqv..false.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self         !< The string.
   logical,       intent(in), optional :: allow_spaces !< Allow leading-trailing spaces.
   logical                             :: is_number    !< Result of the test.

   is_number = (self%is_integer(allow_spaces=allow_spaces).or.self%is_real(allow_spaces=allow_spaces))
   endfunction is_number

   elemental function is_real(self, allow_spaces)
   !< Return true if the string contains a real.
   !<
   !< The regular expression is `\s*[\+\-]?\d*(|\.?\d*([deDE][\+\-]?\d+)?)\s*`. The parse algorithm is done in stages:
   !<
   !< | S0  | S1      | S2  | S3  | S4  | S5     | S6      | S7  | S8  |
   !< |-----|---------|-----|-----|-----|--------|---------|-----|-----|
   !< |`\s*`|`[\+\-]?`|`\d*`|`\.?`|`\d*`|`[deDE]`|`[\+\-]?`|`\d*`|`\s*`|
   !<
   !< Exit on stages-parsing results in:
   !<
   !< | S0 | S1 | S2 | S3 | S4 | S5 | S6 | S7 | S8 |
   !< |----|----|----|----|----|----|----|----|----|
   !  |  F |  F |  T |  T |  T |  F |  F |  T |  T |
   !<
   !< @note This implementation is courtesy of
   !< [tomedunn](https://github.com/tomedunn/fortran-string-utility-module/blob/master/src/string_utility_module.f90#L614)
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(6)
   !< astring = '   -1212112.d0 '
   !< test_passed(1) = astring%is_real().eqv..true.
   !< astring = '   -1212112.d0'
   !< test_passed(2) = astring%is_real(allow_spaces=.false.).eqv..false.
   !< astring = '-1212112.d0   '
   !< test_passed(3) = astring%is_real(allow_spaces=.false.).eqv..false.
   !< astring = '+2.e20'
   !< test_passed(4) = astring%is_real().eqv..true.
   !< astring = ' -2.01E13 '
   !< test_passed(5) = astring%is_real().eqv..true.
   !< astring = ' -2.01 E13 '
   !< test_passed(6) = astring%is_real().eqv..false.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)           :: self              !< The string.
   logical,       intent(in), optional :: allow_spaces      !< Allow leading-trailing spaces.
   logical                             :: is_real           !< Result of the test.
   logical                             :: allow_spaces_     !< Allow leading-trailing spaces, local variable.
   logical                             :: has_leading_digit !< Check the presence of leading digits.
   integer                             :: stage             !< Stages counter.
   integer                             :: c                 !< Character counter.

   if (allocated(self%raw)) then
      allow_spaces_ = .true. ; if (present(allow_spaces)) allow_spaces_ = allow_spaces
      stage = 0
      is_real = .true.
      has_leading_digit = .false.
      do c=1, len(self%raw)
         select case(self%raw(c:c))
         case(SPACE, TAB)
            select case(stage)
            case(0, 8)
               is_real = allow_spaces_
               continue
            case(2:4, 7)
               is_real = allow_spaces_
               stage = 8
            case default
               is_real = .false.
            endselect
         case('+', '-')
            select case(stage)
            case(0)
               stage = 1
            case(5)
               stage = 6
            case default
               is_real = .false.
            endselect
         case('0':'9')
            select case(stage)
            case(0:1)
               stage = 2
               has_leading_digit = .true.
            case(3)
               stage = 4
            case(5:6)
               stage = 7
            case default
               continue
            endselect
         case('.')
            select case(stage)
            case(0:2)
               stage = 3
            case default
               is_real = .false.
            endselect
         case('e','E','d','D')
            select case(stage)
            case(2:4)
               stage = 5
            case default
               is_real = .false.
            endselect
         case default
            is_real = .false.
         endselect
         if (.not.is_real) exit
      enddo
   endif
   if (is_real) then
      select case(stage)
      case(2, 4, 7, 8)
         is_real = .true.
      case(3)
         is_real = has_leading_digit
      case default
         is_real = .false.
      endselect
   endif
   endfunction is_real

   elemental function is_upper(self)
   !< Return true if all characters in the string are uppercase.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(3)
   !< astring = ' Hello World'
   !< test_passed(1) = astring%is_upper().eqv..false.
   !< astring = ' HELLO WORLD'
   !< test_passed(2) = astring%is_upper().eqv..true.
   !< astring = ' hello world'
   !< test_passed(3) = astring%is_upper().eqv..false.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: self     !< The string.
   logical                   :: is_upper !< Result of the test.
   integer                   :: c        !< Character counter.

   is_upper = .false.
   if (allocated(self%raw)) then
      is_upper = .true.
      do c=1, len(self%raw)
         if (index(LOWER_ALPHABET, self%raw(c:c))>0) then
            is_upper = .false.
            exit
         endif
      enddo
   endif
   endfunction is_upper

   elemental function start_with(self, prefix, start, end)
   !< Return true if a string starts with a specified prefix.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(4)
   !< astring = 'Hello WorLD!'
   !< test_passed(1) = astring%start_with(prefix='Hello').eqv..true.
   !< test_passed(2) = astring%start_with(prefix='hell').eqv..false.
   !< test_passed(3) = astring%start_with(prefix='llo Wor', start=3).eqv..true.
   !< test_passed(4) = astring%start_with(prefix='lo W', start=4, end=7).eqv..true.
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)           :: self       !< The string.
   character(kind=CK, len=*), intent(in)           :: prefix     !< Searched prefix.
   integer,                   intent(in), optional :: start      !< Start position into the string.
   integer,                   intent(in), optional :: end        !< End position into the string.
   logical                                         :: start_with !< Result of the test.
   integer                                         :: start_     !< Start position into the string, local variable.
   integer                                         :: end_       !< End position into the string, local variable.

   start_with = .false.
   if (allocated(self%raw)) then
      start_ = 1             ; if (present(start)) start_ = start
      end_   = len(self%raw) ; if (present(end))   end_   = end
      if (len(prefix)<=len(self%raw(start_:end_))) then
         start_with = index(self%raw(start_:end_), prefix)==1
      endif
   endif
   endfunction start_with

   ! private methods

   ! assignments
   pure subroutine string_assign_string(lhs, rhs)
   !< Assignment operator from string input.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(1)
   !< astring = 'hello'
   !< anotherstring = astring
   !< test_passed(1) = astring%chars()==anotherstring%chars()
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   type(string),  intent(in)    :: rhs !< Right hand side.

   if (allocated(rhs%raw)) lhs%raw = rhs%raw
   endsubroutine string_assign_string

   pure subroutine string_assign_character(lhs, rhs)
   !< Assignment operator from character input.
   !<
   !<```fortran
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 'hello'
   !< test_passed(1) = astring%chars()=='hello'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(inout) :: lhs !< Left hand side.
   character(kind=CK, len=*), intent(in)    :: rhs !< Right hand side.

   lhs%raw = rhs
   endsubroutine string_assign_character

   pure subroutine string_assign_integer_I1P(lhs, rhs)
   !< Assignment operator from integer input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 127_I1P
   !< test_passed(1) = astring%to_number(kind=1_I1P)==127_I1P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   integer(I1P),  intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_integer_I1P

   pure subroutine string_assign_integer_I2P(lhs, rhs)
   !< Assignment operator from integer input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 127_I2P
   !< test_passed(1) = astring%to_number(kind=1_I2P)==127_I2P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   integer(I2P),  intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_integer_I2P

   pure subroutine string_assign_integer_I4P(lhs, rhs)
   !< Assignment operator from integer input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 127_I4P
   !< test_passed(1) = astring%to_number(kind=1_I4P)==127_I4P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   integer(I4P),  intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_integer_I4P

   pure subroutine string_assign_integer_I8P(lhs, rhs)
   !< Assignment operator from integer input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 127_I8P
   !< test_passed(1) = astring%to_number(kind=1_I8P)==127_I8P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   integer(I8P),  intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_integer_I8P

   pure subroutine string_assign_real_R4P(lhs, rhs)
   !< Assignment operator from real input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 3.021e6_R4P
   !< test_passed(1) = astring%to_number(kind=1._R4P)==3.021e6_R4P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   real(R4P),     intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_real_R4P

   pure subroutine string_assign_real_R8P(lhs, rhs)
   !< Assignment operator from real input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 3.021e6_R8P
   !< test_passed(1) = astring%to_number(kind=1._R8P)==3.021e6_R8P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   real(R8P),     intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_real_R8P

   pure subroutine string_assign_real_R16P(lhs, rhs)
   !< Assignment operator from real input.
   !<
   !<```fortran
   !< use penf
   !< type(string) :: astring
   !< logical      :: test_passed(1)
   !< astring = 3.021e6_R8P
   !< test_passed(1) = astring%to_number(kind=1._R8P)==3.021e6_R8P
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(inout) :: lhs !< Left hand side.
   real(R16P),    intent(in)    :: rhs !< Right hand side.

   lhs%raw = trim(str(rhs))
   endsubroutine string_assign_real_R16P

   ! contatenation operators
   pure function string_concat_string(lhs, rhs) result(concat)
   !< Concatenation with string.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(1)
   !< astring = 'Hello '
   !< anotherstring = 'Bye bye'
   !< test_passed(1) = astring//anotherstring=='Hello Bye bye'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)              :: lhs    !< Left hand side.
   type(string),  intent(in)              :: rhs    !< Right hand side.
   character(kind=CK, len=:), allocatable :: concat !< Concatenated string.

   concat = ''
   if (allocated(lhs%raw)) concat = lhs%raw
   if (allocated(rhs%raw)) concat = concat//rhs%raw
   endfunction string_concat_string

   pure function string_concat_character(lhs, rhs) result(concat)
   !< Concatenation with character.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(1)
   !< astring = 'Hello '
   !< acharacter = 'World!'
   !< test_passed(1) = astring//acharacter=='Hello World!'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)  :: lhs    !< Left hand side.
   character(kind=CK, len=*), intent(in)  :: rhs    !< Right hand side.
   character(kind=CK, len=:), allocatable :: concat !< Concatenated string.

   if (allocated(lhs%raw)) then
      concat = lhs%raw//rhs
   else
      concat = rhs
   endif
   endfunction string_concat_character

   pure function character_concat_string(lhs, rhs) result(concat)
   !< Concatenation with character (inverted).
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(1)
   !< astring = 'Hello '
   !< acharacter = 'World!'
   !< test_passed(1) = acharacter//astring=='World!Hello '
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)  :: lhs    !< Left hand side.
   class(string),             intent(in)  :: rhs    !< Right hand side.
   character(kind=CK, len=:), allocatable :: concat !< Concatenated string.

   if (allocated(rhs%raw)) then
      concat = lhs//rhs%raw
   else
      concat = lhs
   endif
   endfunction character_concat_string

   elemental function string_concat_string_string(lhs, rhs) result(concat)
   !< Concatenation with string.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< type(string) :: yetanotherstring
   !< logical      :: test_passed(1)
   !< astring = 'Hello '
   !< anotherstring = 'Bye bye'
   !< yetanotherstring = astring.cat.anotherstring
   !< test_passed(1) = yetanotherstring%chars()=='Hello Bye bye'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in)              :: lhs       !< Left hand side.
   type(string),  intent(in)              :: rhs       !< Right hand side.
   type(string)                           :: concat    !< Concatenated string.
   character(kind=CK, len=:), allocatable :: temporary !< Temporary concatenated string.

   temporary = ''
   if (allocated(lhs%raw)) temporary = lhs%raw
   if (allocated(rhs%raw)) temporary = temporary//rhs%raw
   if (temporary/='') concat%raw = temporary
   endfunction string_concat_string_string

   elemental function string_concat_character_string(lhs, rhs) result(concat)
   !< Concatenation with character.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< type(string)                  :: yetanotherstring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(1)
   !< astring = 'Hello '
   !< acharacter = 'World!'
   !< yetanotherstring = astring.cat.acharacter
   !< test_passed(1) = yetanotherstring%chars()=='Hello World!'
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in)  :: lhs    !< Left hand side.
   character(kind=CK, len=*), intent(in)  :: rhs    !< Right hand side.
   type(string)                           :: concat !< Concatenated string.

   if (allocated(lhs%raw)) then
      concat%raw = lhs%raw//rhs
   else
      concat%raw = rhs
   endif
   endfunction string_concat_character_string

   elemental function character_concat_string_string(lhs, rhs) result(concat)
   !< Concatenation with character (inverted).
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< type(string)                  :: yetanotherstring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(1)
   !< astring = 'Hello '
   !< acharacter = 'World!'
   !< yetanotherstring = acharacter.cat.astring
   !< test_passed(1) = yetanotherstring%chars()=='World!Hello '
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in)  :: lhs    !< Left hand side.
   class(string),             intent(in)  :: rhs    !< Right hand side.
   type(string)                           :: concat !< Concatenated string.

   if (allocated(rhs%raw)) then
     concat%raw = lhs//rhs%raw
   else
     concat%raw = lhs
   endif
   endfunction character_concat_string_string

   ! logical operators
   elemental function string_eq_string(lhs, rhs) result(is_it)
   !< Equal to string logical operator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(2)
   !< astring = '  one '
   !< anotherstring = 'two'
   !< test_passed(1) = ((astring==anotherstring).eqv..false.)
   !< astring = 'the same '
   !< anotherstring = 'the same '
   !< test_passed(2) = ((astring==anotherstring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: lhs   !< Left hand side.
   type(string),  intent(in) :: rhs   !< Right hand side.
   logical                   :: is_it !< Opreator test result.

   is_it = lhs%raw == rhs%raw
   endfunction string_eq_string

   elemental function string_eq_character(lhs, rhs) result(is_it)
   !< Equal to character logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = '  one '
   !< acharacter = 'three'
   !< test_passed(1) = ((astring==acharacter).eqv..false.)
   !< astring = 'the same '
   !< acharacter = 'the same '
   !< test_passed(2) = ((astring==acharacter).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in) :: lhs   !< Left hand side.
   character(kind=CK, len=*), intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs%raw == rhs
   endfunction string_eq_character

   elemental function character_eq_string(lhs, rhs) result(is_it)
   !< Equal to character (inverted) logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = '  one '
   !< acharacter = 'three'
   !< test_passed(1) = ((acharacter==astring).eqv..false.)
   !< astring = 'the same '
   !< acharacter = 'the same '
   !< test_passed(2) = ((acharacter==astring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: lhs   !< Left hand side.
   class(string),             intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = rhs%raw == lhs
   endfunction character_eq_string

   elemental function string_ne_string(lhs, rhs) result(is_it)
   !< Not equal to string logical operator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(2)
   !< astring = '  one '
   !< anotherstring = 'two'
   !< test_passed(1) = ((astring/=anotherstring).eqv..true.)
   !< astring = 'the same '
   !< anotherstring = 'the same '
   !< test_passed(2) = ((astring/=anotherstring).eqv..false.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: lhs   !< Left hand side.
   type(string),  intent(in) :: rhs   !< Right hand side.
   logical                   :: is_it !< Opreator test result.

   is_it = lhs%raw /= rhs%raw
   endfunction string_ne_string

   elemental function string_ne_character(lhs, rhs) result(is_it)
   !< Not equal to character logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = '  one '
   !< acharacter = 'three'
   !< test_passed(1) = ((astring/=acharacter).eqv..true.)
   !< astring = 'the same '
   !< acharacter = 'the same '
   !< test_passed(2) = ((astring/=acharacter).eqv..false.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in) :: lhs   !< Left hand side.
   character(kind=CK, len=*), intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs%raw /= rhs
   endfunction string_ne_character

   elemental function character_ne_string(lhs, rhs) result(is_it)
   !< Not equal to character (inverted) logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = '  one '
   !< acharacter = 'three'
   !< test_passed(1) = ((acharacter/=astring).eqv..true.)
   !< astring = 'the same '
   !< acharacter = 'the same '
   !< test_passed(2) = ((acharacter/=astring).eqv..false.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: lhs   !< Left hand side.
   class(string),             intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = rhs%raw /= lhs
   endfunction character_ne_string

   elemental function string_lt_string(lhs, rhs) result(is_it)
   !< Lower than to string logical operator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(2)
   !< astring = 'one'
   !< anotherstring = 'ONE'
   !< test_passed(1) = ((astring<anotherstring).eqv..false.)
   !< astring = 'ONE'
   !< anotherstring = 'one'
   !< test_passed(2) = ((astring<anotherstring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: lhs   !< Left hand side.
   type(string),  intent(in) :: rhs   !< Right hand side.
   logical                   :: is_it !< Opreator test result.

   is_it = lhs%raw < rhs%raw
   endfunction string_lt_string

   elemental function string_lt_character(lhs, rhs) result(is_it)
   !< Lower than to character logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((astring<acharacter).eqv..false.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((astring<acharacter).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in) :: lhs   !< Left hand side.
   character(kind=CK, len=*), intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs%raw < rhs
   endfunction string_lt_character

   elemental function character_lt_string(lhs, rhs) result(is_it)
   !< Lower than to character (inverted) logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((acharacter<astring).eqv..true.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((acharacter<astring).eqv..false.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: lhs   !< Left hand side.
   class(string),             intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs < rhs%raw
   endfunction character_lt_string

   elemental function string_le_string(lhs, rhs) result(is_it)
   !< Lower equal than to string logical operator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(3)
   !< astring = 'one'
   !< anotherstring = 'ONE'
   !< test_passed(1) = ((astring<=anotherstring).eqv..false.)
   !< astring = 'ONE'
   !< anotherstring = 'one'
   !< test_passed(2) = ((astring<=anotherstring).eqv..true.)
   !< astring = 'ONE'
   !< anotherstring = 'ONE'
   !< test_passed(3) = ((astring<=anotherstring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: lhs   !< Left hand side.
   type(string),  intent(in) :: rhs   !< Right hand side.
   logical                   :: is_it !< Opreator test result.

   is_it = lhs%raw <= rhs%raw
   endfunction string_le_string

   elemental function string_le_character(lhs, rhs) result(is_it)
   !< Lower equal than to character logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(3)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((astring<=acharacter).eqv..false.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((astring<=acharacter).eqv..true.)
   !< astring = 'ONE'
   !< acharacter = 'ONE'
   !< test_passed(3) = ((astring<=acharacter).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in) :: lhs   !< Left hand side.
   character(kind=CK, len=*), intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs%raw <= rhs
   endfunction string_le_character

   elemental function character_le_string(lhs, rhs) result(is_it)
   !< Lower equal than to character (inverted) logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(3)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((acharacter<=astring).eqv..true.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((acharacter<=astring).eqv..false.)
   !< astring = 'ONE'
   !< acharacter = 'ONE'
   !< test_passed(3) = ((acharacter<=astring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: lhs   !< Left hand side.
   class(string),             intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs <= rhs%raw
   endfunction character_le_string

   elemental function string_ge_string(lhs, rhs) result(is_it)
   !< Greater equal than to string logical operator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(3)
   !< astring = 'one'
   !< anotherstring = 'ONE'
   !< test_passed(1) = ((astring>=anotherstring).eqv..true.)
   !< astring = 'ONE'
   !< anotherstring = 'one'
   !< test_passed(2) = ((astring>=anotherstring).eqv..false.)
   !< astring = 'ONE'
   !< anotherstring = 'ONE'
   !< test_passed(3) = ((astring>=anotherstring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: lhs   !< Left hand side.
   type(string),  intent(in) :: rhs   !< Right hand side.
   logical                   :: is_it !< Opreator test result.

   is_it = lhs%raw >= rhs%raw
   endfunction string_ge_string

   elemental function string_ge_character(lhs, rhs) result(is_it)
   !< Greater equal than to character logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(3)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((astring>=acharacter).eqv..true.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((astring>=acharacter).eqv..false.)
   !< astring = 'ONE'
   !< acharacter = 'ONE'
   !< test_passed(3) = ((astring>=acharacter).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in) :: lhs   !< Left hand side.
   character(kind=CK, len=*), intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs%raw >= rhs
   endfunction string_ge_character

   elemental function character_ge_string(lhs, rhs) result(is_it)
   !< Greater equal than to character (inverted) logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(3)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((acharacter>=astring).eqv..false.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((acharacter>=astring).eqv..true.)
   !< astring = 'ONE'
   !< acharacter = 'ONE'
   !< test_passed(3) = ((acharacter>=astring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: lhs   !< Left hand side.
   class(string),             intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs >= rhs%raw
   endfunction character_ge_string

   elemental function string_gt_string(lhs, rhs) result(is_it)
   !< Greater than to string logical operator.
   !<
   !<```fortran
   !< type(string) :: astring
   !< type(string) :: anotherstring
   !< logical      :: test_passed(2)
   !< astring = 'one'
   !< anotherstring = 'ONE'
   !< test_passed(1) = ((astring>anotherstring).eqv..true.)
   !< astring = 'ONE'
   !< anotherstring = 'one'
   !< test_passed(2) = ((astring>anotherstring).eqv..false.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string), intent(in) :: lhs   !< Left hand side.
   type(string),  intent(in) :: rhs   !< Right hand side.
   logical                   :: is_it !< Opreator test result.

   is_it = lhs%raw > rhs%raw
   endfunction string_gt_string

   elemental function string_gt_character(lhs, rhs) result(is_it)
   !< Greater than to character logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((astring>acharacter).eqv..true.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((astring>acharacter).eqv..false.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   class(string),             intent(in) :: lhs   !< Left hand side.
   character(kind=CK, len=*), intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs%raw > rhs
   endfunction string_gt_character

   elemental function character_gt_string(lhs, rhs) result(is_it)
   !< Greater than to character (inverted) logical operator.
   !<
   !<```fortran
   !< type(string)                  :: astring
   !< character(len=:), allocatable :: acharacter
   !< logical                       :: test_passed(2)
   !< astring = 'one'
   !< acharacter = 'ONE'
   !< test_passed(1) = ((acharacter>astring).eqv..false.)
   !< astring = 'ONE'
   !< acharacter = 'one'
   !< test_passed(2) = ((acharacter>astring).eqv..true.)
   !< print '(L1)', all(test_passed)
   !<```
   !=> T <<<
   character(kind=CK, len=*), intent(in) :: lhs   !< Left hand side.
   class(string),             intent(in) :: rhs   !< Right hand side.
   logical                               :: is_it !< Opreator test result.

   is_it = lhs > rhs%raw
   endfunction character_gt_string

   ! IO
   subroutine read_formatted(dtv, unit, iotype, v_list, iostat, iomsg)
   !< Formatted input.
   !<
   !< @bug Change temporary acks: find a more precise length of the input string and avoid the trimming!
   !<
   !< @bug Read listdirected with and without delimiters does not work.
   class(string),             intent(inout) :: dtv         !< The string.
   integer,                   intent(in)    :: unit        !< Logical unit.
   character(len=*),          intent(in)    :: iotype      !< Edit descriptor.
   integer,                   intent(in)    :: v_list(:)   !< Edit descriptor list.
   integer,                   intent(out)   :: iostat      !< IO status code.
   character(len=*),          intent(inout) :: iomsg       !< IO status message.
   character(len=len(iomsg))                :: local_iomsg !< Local variant of iomsg, so it doesn't get inappropriately redefined.
   character(kind=CK, len=1)                :: delim       !< String delimiter, if any.
   character(kind=CK, len=100)              :: temporary   !< Temporary storage string.

   if (iotype == 'LISTDIRECTED') then
      call get_next_non_blank_character_any_record(unit=unit, ch=delim, iostat=iostat, iomsg=iomsg)
      if (iostat/=0) return
      if (delim=='"'.OR.delim=="'") then
         call dtv%read_delimited(unit=unit, delim=delim, iostat=iostat, iomsg=local_iomsg)
      else
         ! step back before the non-blank
         read(unit, "(TL1)", iostat=iostat, iomsg=iomsg)
         if (iostat /= 0) return
         call dtv%read_undelimited_listdirected(unit=unit, iostat=iostat, iomsg=local_iomsg)
      endif
      if (is_iostat_eor(iostat)) then
         ! suppress IOSTAT_EOR
         iostat = 0
      elseif (iostat /= 0) then
         iomsg = local_iomsg
      endif
      return
   else
      read(unit, "(A)", iostat=iostat, iomsg=iomsg)temporary
      dtv%raw = trim(temporary)
   endif
   endsubroutine read_formatted

   subroutine read_delimited(dtv, unit, delim, iostat, iomsg)
   !< Read a delimited string from a unit connected for formatted input.
   !<
   !< If the closing delimiter is followed by end of record, then we return end of record.
   !<
   !< @note This does not need a doctest, it being tested by [[string::read_formatted]].
   class(string),             intent(out)   :: dtv       !< The string.
   integer,                   intent(in)    :: unit      !< Logical unit.
   character(kind=CK, len=1), intent(in)    :: delim     !< String delimiter.
   integer,                   intent(out)   :: iostat    !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg     !< IO status message.
   character(kind=CK, len=1)                :: ch        !< A character read.
   logical                                  :: was_delim !< Indicates that the last character read was a delimiter.

   was_delim = .false.
   dtv%raw = ''
   do
      read(unit, "(A)", iostat=iostat, iomsg=iomsg) ch
      if (is_iostat_eor(iostat)) then
         if (was_delim) then
           ! end of delimited string followed by end of record is end of the string. Pass back the
           ! end of record condition to the caller
           return
         else
           ! end of record without terminating delimiter - move along
           cycle
         endif
      elseif (iostat /= 0) THEN
        return
      endif
      if (ch == delim) then
         if (was_delim) then
            ! doubled delimiter is one delimiter in the value
            dtv%raw = dtv%raw // ch
            was_delim = .false.
         else
            ! need to test next character to see what is happening
            was_delim = .true.
         endif
      elseif (was_delim) then
         ! the previous character was actually the delimiter for the end of the string. Put back this character
         read(unit, "(TL1)", iostat=iostat, iomsg=iomsg)
         return
      else
         dtv%raw = dtv%raw // ch
      endif
   enddo
   endsubroutine read_delimited

  subroutine read_undelimited_listdirected(dtv, unit, iostat, iomsg)
  !< Read an undelimited (no leading apostrophe or double quote) character value according to the rules for list directed input.
  !<
  !< A blank, comma/semicolon (depending on the decimal mode), slash or end of record terminates the string.
  !<
  !< If input is terminated by end of record, then this procedure returns an end-of-record condition.
  class(string),    intent(inout) :: dtv           !< The string.
  integer,          intent(in)    :: unit          !< Logical unit.
  integer,          intent(out)   :: iostat        !< IO status code.
  character(len=*), intent(inout) :: iomsg         !< IO status message.
  logical                         :: decimal_point !<True if DECIMAL=POINT in effect.

  call get_decimal_mode(unit=unit, decimal_point=decimal_point, iostat=iostat, iomsg=iomsg)
  if (iostat /= 0) return
  call dtv%read_undelimited(unit=unit, terminators=' '//'/'//merge(CK_',', CK_';', decimal_point), iostat=iostat, iomsg=iomsg)
  endsubroutine read_undelimited_listdirected

  subroutine read_undelimited(dtv, unit, terminators, iostat, iomsg)
  !< Read an undelimited string up until end of record or a character from a set of terminators is encountered.
  !<
  !< If a terminator is encountered, the file position will be at that terminating character. If end of record is encountered, the
  !< file remains at end of record.
  class(string),             intent(inout) :: dtv         !< The string.
  integer,                   intent(in)    :: unit        !< Logical unit.
  character(kind=CK, len=*), intent(in)    :: terminators !< Characters that are considered to terminate the string.
                                                          !< Blanks in this string are meaningful.
  integer,                   intent(out)   :: iostat      !< IO status code.
  character(len=*),          intent(inout) :: iomsg       !< IO status message.
  character(kind=CK, len=1)                :: ch          !< A character read.

  dtv%raw = ''
  do
    read(unit, "(A)", iostat=iostat, iomsg=iomsg) ch
    if (is_iostat_eor(iostat)) then
      ! end of record just means end of string. We pass on the condition
      return
    elseif (iostat /= 0) then
      ! something odd happened
      return
    endif
    if (scan(ch, terminators) /= 0) then
      ! change the file position so that the next read sees the terminator
      read(unit, "(TL1)", iostat=iostat, iomsg=iomsg)
      if (iostat /= 0) return
      iostat = 0
      return
    endif
    ! we got a character - append it
    dtv%raw = dtv%raw // ch
  enddo
  endsubroutine read_undelimited

   subroutine write_formatted(dtv, unit, iotype, v_list, iostat, iomsg)
   !< Formatted output.
   class(string),             intent(in)    :: dtv       !< The string.
   integer,                   intent(in)    :: unit      !< Logical unit.
   character(kind=CK, len=*), intent(in)    :: iotype    !< Edit descriptor.
   integer,                   intent(in)    :: v_list(:) !< Edit descriptor list.
   integer,                   intent(out)   :: iostat    !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg     !< IO status message.

   if (allocated(dtv%raw)) then
     write(unit, "(A)", iostat=iostat, iomsg=iomsg)dtv%raw
   else
     write(unit, "(A)", iostat=iostat, iomsg=iomsg)''
   endif
   endsubroutine write_formatted

   subroutine read_unformatted(dtv, unit, iostat, iomsg)
   !< Unformatted input.
   !<
   !< @bug Change temporary acks: find a more precise length of the input string and avoid the trimming!
   class(string),             intent(inout) :: dtv       !< The string.
   integer,                   intent(in)    :: unit      !< Logical unit.
   integer,                   intent(out)   :: iostat    !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg     !< IO status message.
   character(kind=CK, len=100)              :: temporary !< Temporary storage string.

   read(unit, iostat=iostat, iomsg=iomsg)temporary
   dtv%raw = trim(temporary)
   endsubroutine read_unformatted

   subroutine write_unformatted(dtv, unit, iostat, iomsg)
   !< Unformatted output.
   class(string),             intent(in)    :: dtv    !< The string.
   integer,                   intent(in)    :: unit   !< Logical unit.
   integer,                   intent(out)   :: iostat !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg  !< IO status message.

   if (allocated(dtv%raw)) then
     write(unit, iostat=iostat, iomsg=iomsg)dtv%raw
   else
     write(unit, iostat=iostat, iomsg=iomsg)''
   endif
   endsubroutine write_unformatted

   ! miscellanea
   elemental function replace_one_occurrence(self, old, new) result(replaced)
   !< Return a string with the first occurrence of substring old replaced by new.
   !<
   !< @note The doctest is not necessary, this being tested by [[string:replace]].
   class(string),             intent(in)  :: self      !< The string.
   character(kind=CK, len=*), intent(in)  :: old       !< Old substring.
   character(kind=CK, len=*), intent(in)  :: new       !< New substring.
   type(string)                           :: replaced  !< The string with old replaced by new.
   integer                                :: pos       !< Position from which replace old.

   if (allocated(self%raw)) then
      replaced = self
      pos = index(string=self%raw, substring=old)
      if (pos>0) then
         if (pos==1) then
            replaced%raw = new//self%raw(len(old)+1:)
         else
            replaced%raw = self%raw(1:pos-1)//new//self%raw(pos+len(old):)
         endif
      endif
   endif
   endfunction replace_one_occurrence

   ! non type-bound-procedures
   subroutine get_delimiter_mode(unit, delim, iostat, iomsg)
   !< Get the DELIM changeable connection mode for the given unit.
   !<
   !< If the unit is connected to an internal file, then the default value of NONE is always returned.
   use, intrinsic :: iso_fortran_env, only : iostat_inquire_internal_unit
   integer,                   intent(in)    :: unit         !< The unit for the connection.
   character(len=1, kind=CK), intent(out)   :: delim        !< Represents the value of the DELIM mode.
   integer,                   intent(out)   :: iostat       !< IOSTAT error code, non-zero on error.
   character(*),              intent(inout) :: iomsg        !< IOMSG explanatory message - only defined if iostat is non-zero.
   character(10)                            :: delim_buffer !< Buffer for INQUIRE about DELIM, sized for APOSTROHPE.
   character(len(iomsg))                    :: local_iomsg  !< Local variant of iomsg, so it doesn't get inappropriately redefined.

   ! get the string representation of the changeable mode
   inquire(unit, delim=delim_buffer, iostat=iostat, iomsg=local_iomsg)
   if (iostat == iostat_inquire_internal_unit) then
      ! no way of determining the DELIM mode for an internal file
      iostat = 0
      delim = ''
      return
   elseif (iostat /= 0) then
      iomsg = local_iomsg
      return
   endif
   ! interpret the DELIM string
   if (delim_buffer == 'QUOTE') then
      delim = '"'
   elseif (delim_buffer == 'APOSTROPHE') then
      delim = ''''
   else
      delim = '"'
   endif
   endsubroutine get_delimiter_mode

   subroutine get_next_non_blank_character_this_record(unit, ch, iostat, iomsg)
   !< Get the next non-blank character in the current record.
   integer,                   intent(in)    :: unit   !< Logical unit.
   character(kind=CK, len=1), intent(out)   :: ch     !< The non-blank character read. Not valid if IOSTAT is non-zero.
   integer,                   intent(out)   :: iostat !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg  !< IO status message.

   do
      ! we spcify non-advancing, just in case we want this callable outside the context of a child input statement
      ! the PAD specifier simply saves the need for the READ statement to define ch if EOR is hit
      ! read(unit, "(A)", iostat=iostat, iomsg=iomsg, advance='NO') ch
      ! ...but that causes ifort to blow up at runtime
      read(unit, "(A)", iostat=iostat, iomsg=iomsg, pad='NO') ch
      if (iostat /= 0) return
      if (ch /= '') exit
   enddo
   endsubroutine get_next_non_blank_character_this_record

   subroutine get_next_non_blank_character_any_record(unit, ch, iostat, iomsg)
   !< Get the next non-blank character, advancing records if necessary.
   integer,                   intent(in)    :: unit        !< Logical unit.
   character(kind=CK, len=1), intent(out)   :: ch          !< The non-blank character read. Not valid if IOSTAT is non-zero.
   integer,                   intent(out)   :: iostat      !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg       !< IO status message.
   character(len(iomsg))                    :: local_iomsg !< Local variant of iomsg, so it doesn't get inappropriately redefined.

   do
      call get_next_non_blank_character_this_record(unit=unit, ch=ch, iostat=iostat, iomsg=local_iomsg)
      if (is_iostat_eor(iostat)) then
         ! try again on the next record
         read (unit, "(/)", iostat=iostat, iomsg=iomsg)
         if (iostat /= 0) return
      elseif (iostat /= 0) then
         ! some sort of problem
         iomsg = local_iomsg
         return
      else
         ! got it
         exit
      endif
   enddo
   endsubroutine get_next_non_blank_character_any_record

   subroutine get_decimal_mode(unit, decimal_point, iostat, iomsg)
   !< Get the DECIMAL changeable connection mode for the given unit.
   !<
   !< If the unit is connected to an internal file, then the default value of DECIMAL is always returned. This may not be the
   !< actual value in force at the time of the call to this procedure.
   use, intrinsic :: iso_fortran_env, only : iostat_inquire_internal_unit
   integer,                   intent(in)    :: unit           !< Logical unit.
   logical,                   intent(out)   :: decimal_point  !< True if the decimal mode is POINT, false otherwise.
   integer,                   intent(out)   :: iostat         !< IO status code.
   character(kind=CK, len=*), intent(inout) :: iomsg          !< IO status message.
   character(5)                             :: decimal_buffer !< Buffer for INQUIRE about DECIMAL, sized for POINT or COMMA.
   character(len(iomsg))                    :: local_iomsg    !< Local iomsg, so it doesn't get inappropriately redefined.

   inquire(unit, decimal=decimal_buffer, iostat=iostat, iomsg=local_iomsg)
   if (iostat == iostat_inquire_internal_unit) then
      ! no way of determining the decimal mode for an internal file
      iostat = 0
      decimal_point = .true.
      return
   else if (iostat /= 0) then
      iomsg = local_iomsg
      return
   endif
   decimal_point = decimal_buffer == 'POINT'
   endsubroutine get_decimal_mode
endmodule stringifor_string_t