Some efficiency improvements when parsing strings.

.gitignore

@@ -27,7 +27,8 @@ visual_studio/jsonfortranlib/
 visual_studio/jsonfortrantest/Debug/
 visual_studio/jsonfortrantest/Release/
 visual_studio/jsonfortrantest/x64
-
+visual_studio/config
+My Advisor Results*
 My Amplifier*
 My Inspector*
 x64/

files/inputs/invalid.json

@@ -10,4 +10,4 @@
     "..\\path\\to\\files\\file1.txt",
     "..\\path\\to\\files\\file2.txt",
     "..\\path\\to\\files\\file3.txt"
-}
+}

files/inputs/invalid3.json

@@ -0,0 +1,4 @@
+{
+"a": "blah\",
+"b": 2
+}

files/inputs/test1.json

@@ -12,7 +12,8 @@
 		"..\\path\\to\\files\\file3.txt",
 		"test \u2FA4 \uABCD \uABCD\uABCDtest",
 		"   test  \\u \" blah\\\" test test",
-		"..\\path\\to\\files\\"
+		"..\\path\\to\\files\\",
+		"\\"
 	],
     "a": {
         "b": 1

src/json_file_module.F90

@@ -64,7 +64,8 @@ module json_file_module
 
         private
 
-        type(json_core) :: core !! The instance of the [[json_core(type)]] factory used for this file.
+        type(json_core)          :: core         !! The instance of the [[json_core(type)]]
+                                                 !! factory used for this file.
         type(json_value),pointer :: p => null()  !! the JSON structure read from the file
 
     contains
@@ -319,10 +320,17 @@ subroutine json_file_check_for_errors(me,status_ok,error_msg)
     implicit none
 
     class(json_file),intent(inout) :: me
-    logical(LK),intent(out) :: status_ok !! true if there were no errors
-    character(kind=CK,len=:),allocatable,intent(out) :: error_msg !! the error message (if there were errors)
-
+    logical(LK),intent(out),optional :: status_ok !! true if there were no errors
+    character(kind=CK,len=:),allocatable,intent(out),optional :: error_msg !! the error message
+                                                                           !! (if there were errors)
+
+#if defined __GFORTRAN__
+    character(kind=CK,len=:),allocatable :: tmp  !! workaround for gfortran bugs
+    call me%core%check_for_errors(status_ok,tmp)
+    error_msg = tmp
+#else
     call me%core%check_for_errors(status_ok,error_msg)
+#endif
 
     end subroutine json_file_check_for_errors
 !*****************************************************************************************

src/json_string_utilities.F90

@@ -433,68 +433,74 @@ end subroutine escape_string
 !  * `\t`        - horizontal tab
 !  * `\uXXXX`    - 4 hexadecimal digits
 
-    subroutine unescape_string(str_in, str_out, error_message)
+    subroutine unescape_string(str, error_message)
 
     implicit none
 
-    character(kind=CK,len=*),intent(in)              :: str_in  !! string as stored in a [[json_value]]
-    character(kind=CK,len=:),allocatable,intent(out) :: str_out !! decoded string
-    character(kind=CK,len=:),allocatable,intent(out) :: error_message !! will be allocated if there was an error
+    character(kind=CK,len=:),allocatable,intent(inout) :: str           !! in: string as stored
+                                                                        !! in a [[json_value]].
+                                                                        !! out: decoded string.
+    character(kind=CK,len=:),allocatable,intent(out)   :: error_message !! will be allocated if
+                                                                        !! there was an error
 
     integer :: i   !! counter
-    integer :: n   !! length of str_in
-    integer :: m   !! length of str_out
+    integer :: n   !! length of `str`
+    integer :: m   !! length of `str_tmp`
     character(kind=CK,len=1) :: c  !! for scanning each character in string
+    character(kind=CK,len=:),allocatable :: str_tmp !! temp decoded string (if the input
+                                                    !! string contains an escape character
+                                                    !! and needs to be decoded).
 
-#if defined __GFORTRAN__
-    character(kind=CK,len=:),allocatable :: tmp  !! for GFortran bug workaround
-#endif
-
-    if (scan(str_in,backslash)>0) then
+    if (scan(str,backslash)>0) then
 
         !there is at least one escape character, so process this string:
 
-        n = len(str_in)
-        str_out = repeat(space,n) !size the output string (will be trimmed later)
-        m = 0  !counter in str_out
-        i = 0  !counter in str_in
+        n = len(str)
+        str_tmp = repeat(space,n) !size the output string (will be trimmed later)
+        m = 0  !counter in str_tmp
+        i = 0  !counter in str
 
         do
 
             i = i + 1
             if (i>n) exit ! finished
-            c = str_in(i:i) ! get next character in the string
+            c = str(i:i) ! get next character in the string
 
             if (c == backslash) then
 
                 if (i<n) then
 
                     i = i + 1
-                    c = str_in(i:i) !character after the escape
-
-                    if (any(c == [quotation_mark,backslash,slash, &
-                         to_unicode(['b','f','n','r','t'])])) then
-
-                        select case(c)
-                        case (quotation_mark,backslash,slash)
-                            !use d as is
-                        case (CK_'b')
-                             c = bspace
-                        case (CK_'f')
-                             c = formfeed
-                        case (CK_'n')
-                             c = newline
-                        case (CK_'r')
-                             c = carriage_return
-                        case (CK_'t')
-                             c = horizontal_tab
-                        end select
+                    c = str(i:i) !character after the escape
 
+                    select case(c)
+                    case (quotation_mark,backslash,slash)
+                        !use d as is
+                        m = m + 1
+                        str_tmp(m:m) = c
+                    case (CK_'b')
+                        c = bspace
+                        m = m + 1
+                        str_tmp(m:m) = c
+                    case (CK_'f')
+                        c = formfeed
                         m = m + 1
-                        str_out(m:m) = c
+                        str_tmp(m:m) = c
+                    case (CK_'n')
+                        c = newline
+                        m = m + 1
+                        str_tmp(m:m) = c
+                    case (CK_'r')
+                        c = carriage_return
+                        m = m + 1
+                        str_tmp(m:m) = c
+                    case (CK_'t')
+                        c = horizontal_tab
+                        m = m + 1
+                        str_tmp(m:m) = c
 
-                    else if (c == 'u') then !expecting 4 hexadecimal digits after
-                                            !the escape character    [\uXXXX]
+                    case (CK_'u') ! expecting 4 hexadecimal digits after
+                                  ! the escape character    [\uXXXX]
 
                         !for now, we are just returning them as is
                         ![not checking to see if it is a valid hex value]
@@ -504,54 +510,59 @@ subroutine unescape_string(str_in, str_out, error_message)
                         !   \uXXXX
 
                         if (i+4<=n) then
-                            m = m + 1
-                            str_out(m:m+5) = str_in(i-1:i+4)
-                            i = i + 4
-                            m = m + 5
+
+                            ! validate the hex string:
+                            if (valid_json_hex(str(i+1:i+4))) then
+                                m = m + 1
+                                str_tmp(m:m+5) = str(i-1:i+4)
+                                i = i + 4
+                                m = m + 5
+                            else
+                                error_message = 'Error in unescape_string:'//&
+                                                ' Invalid hexadecimal sequence in string "'//&
+                                                trim(str)//'" ['//str(i-1:i+4)//']'
+                                if (allocated(str_tmp)) deallocate(str_tmp)
+                                return
+                            end if
                         else
                             error_message = 'Error in unescape_string:'//&
-                                                 ' Invalid hexadecimal sequence'//&
-                                                 ' in string: '//str_in(i-1:)
-                            if (allocated(str_out)) deallocate(str_out)
+                                            ' Invalid hexadecimal sequence in string "'//&
+                                            trim(str)//'" ['//str(i-1:)//']'
+                            if (allocated(str_tmp)) deallocate(str_tmp)
                             return
                         end if
 
-                    else
+                    case default
+
                         !unknown escape character
                         error_message = 'Error in unescape_string:'//&
-                                             ' unknown escape sequence in string "'//&
-                                             trim(str_in)//'" ['//backslash//c//']'
-                        if (allocated(str_out)) deallocate(str_out)
+                                        ' unknown escape sequence in string "'//&
+                                        trim(str)//'" ['//backslash//c//']'
+                        if (allocated(str_tmp)) deallocate(str_tmp)
                         return
-                    end if
+
+                    end select
 
                 else
-                    !an escape character is the last character in
-                    ! the string [this may not be valid syntax,
-                    ! but just keep it]
-                    m = m + 1
-                    str_out(m:m) = c
+                    ! an escape character is the last character in
+                    ! the string. This is an error.
+                    error_message = 'Error in unescape_string:'//&
+                                    ' invalid escape character in string "'//&
+                                    trim(str)//'"'
+                    if (allocated(str_tmp)) deallocate(str_tmp)
+                    return
                 end if
 
             else
                 m = m + 1
-                str_out(m:m) = c
+                str_tmp(m:m) = c
             end if
 
         end do
 
         !trim trailing space:
-#if defined __GFORTRAN__
-        ! workaround for Gfortran 6.1.0 bug
-        tmp = str_out(1:m)
-        str_out = tmp
-#else
-        str_out = str_out(1:m)
-#endif
+        str = str_tmp(1:m)
 
-    else
-        !there are no escape characters, so return as is:
-        str_out = str_in
     end if
 
     end subroutine unescape_string