Fix color file reading when / is in the coloumn names

Unformatted reads in fortran with / stop rerading after the /
/ is quite common in color files for [fe/h]. So rewrite the line reading
to do the splt ourselves with a new split_line subroutine in utisl_lib.

Warn other people about this fortran "feature" in the code_style

Fixes MESAHub#379

colors/private/mod_colors.f90

@@ -42,7 +42,6 @@ subroutine do_colors_init(num_files,fnames,num_colors,ierr)
          character(len=*),dimension(:),intent(in) :: fnames 
          character(len=strlen) :: fname
          type (lgt_list), pointer :: thead =>null()
-         character(len=strlen),dimension(:),pointer :: col_names=>null()
 
          integer, intent(out) :: ierr
          integer :: i
@@ -81,10 +80,8 @@ subroutine do_colors_init(num_files,fnames,num_colors,ierr)
                ierr=-1
                return
             end if
-
-            col_names => thead_all(i)%color_names
 
-            call init_colors(fname,thead,col_names,thead_all(i)%n_colors,ierr)
+            call init_colors(fname,thead,thead_all(i)%color_names,thead_all(i)%n_colors,ierr)
 
             thead_all(i)%thead=>thead
             bc_total_num_colors=bc_total_num_colors+thead_all(i)%n_colors
@@ -100,7 +97,7 @@ subroutine init_colors(fname, thead, col_names, n_colors, ierr)
          integer, intent(out) :: ierr
          character(len=*),intent(in) :: fname
          type (lgt_list), pointer :: thead 
-         character(len=*),dimension(:),pointer :: col_names
+         character(len=*),dimension(:) :: col_names
          integer, intent(in) :: n_colors
          call Read_Colors_Data(fname, thead, col_names, n_colors, ierr) 
       end subroutine init_colors
@@ -217,10 +214,10 @@ subroutine Read_One_Colors_Data(fname, thead, n_colors, col_names, ierr)
          type (lgt_list), pointer :: tlist => null()
          type (lgz_list), pointer :: zlist => null()
          real(dp), dimension(max_num_bcs_per_file) :: colors
-         character(len=*),dimension(:),pointer,intent(inout) :: col_names
-
-         
-         character(len=strlen) :: tmp
+         character(len=*),dimension(:),intent(out) :: col_names
+         character(len=256) :: tmp_cols(3+n_colors)
+
+         character(len=4096) :: tmp
          integer :: num_entries, num_made, IO_UBV
 
          include 'formats'
@@ -243,16 +240,14 @@ subroutine Read_One_Colors_Data(fname, thead, n_colors, col_names, ierr)
          ierr = 0
          num_entries = 0
          cnt = 0
+         tmp = ''
          !First line should be a header and containing the name of the colours
          !#teff logg m_div_h col_1 col_2 etc
-         read(IO_UBV,fmt=*,iostat=ios) tmp,tmp,tmp,col_names(1:n_colors)
+         read(IO_UBV,'(a)') tmp
+
+         call split_line(tmp,3+n_colors,tmp_cols)
 
-         if(index(tmp,'[')/=0) then
-            write(*,*) 'Do not write metallicity as [Fe/H] use m_div_h in ',trim(fname)
-            ! https://github.com/MESAHub/mesa/issues/379
-            ! Some reason if we have [Fe/H] in tmp we start reading a completly different file
-            ierr = 1; return
-         end if
+         col_names(1:n_colors) = tmp_cols(4:n_colors+3)
 
          do while (.true.)
             read(IO_UBV,fmt=*,iostat=ios) lgt, lgg, lgz, colors(1:n_colors)
@@ -327,7 +322,7 @@ subroutine Read_Colors_Data(fname, thead, col_names, n_colors, ierr)
          integer, intent(out) :: ierr ! 0 means ok
          type (lgt_list), pointer,intent(inout) :: thead
          character (len=*),intent(in) :: fname
-         character(len=*),dimension(:),pointer :: col_names
+         character(len=*),dimension(:),intent(out) :: col_names
          integer, intent(in) :: n_colors
 
          Call Read_One_Colors_Data(fname, thead, n_colors, col_names, ierr)

docs/source/developing/code_style.rst

@@ -167,6 +167,20 @@ different compilers.
 
 Some helpful formats are provided in ``include/formats``.
 
+Unformatted reads
+-----------------
+
+.. code-block:: fortran
+
+    read(unit,*) x,y,z
+
+
+Should be avoided when the variables that are strings. This is becuase if the string contains a / (forward-slash) then when doing a unformatted 
+read frotran will stop reading the line.
+
+Either build a full format statement or read the line into one string and split on whitespace. There is also a function ``split_line`` in 
+``utils_lib.f90`` that can be used to split a line up based on whitespace.
+
 
 Constants
 ---------

utils/public/utils_lib.f90

@@ -1172,6 +1172,40 @@ character(len=strlen) function switch_str(str1,str2,flag)
 
     end function switch_str
 
+   subroutine split_line(line, num, out)
+      ! Given a string line, split on whitespace, into num sub-strings storing them in out
+      character(len=*),intent(in) :: line
+      integer, intent(in) :: num
+      character(len=*),dimension(:) :: out
+
+      integer :: i,kstart,k
+
+      if(size(out)<num) call mesa_error(__FILE__,__LINE__,'out array not large enough for num sub-strings')
+
+      out = ''
+
+      k = 1
+      kstart = 1
+      outer: do i=1, num
+         inner: do
+            if(line(k:k)==' ' .and. line(k+1:k+1)==' ' .and. k < len(line)) then
+               k = k+1
+               cycle inner
+            end if
+
+            if(line(k:k)==' ' .or. k > len(line))then
+               !write(*,*) '*',i,kstart,k,line(kstart:k-1)
+               out(i) = line(kstart:k-1)
+               k = k+1
+               kstart = k
+               cycle outer
+            end if
+            k = k + 1
+         end do inner
+      end do outer
+
+      end subroutine split_line
+
 
    ! backward compatibility so Bill can debug older versions of files without changing these calls
       logical function is_bad_num(x)