Problems compiling Fortran code within IRAF 2.17.1

[msz59]

I am trying to compile old (F77) codes with IRAF 2.17.1 (Ubuntu 22.04), getting all kinds of problems.

Searching on this forum, I found this discussion
When I tried @jeromeorosz way, outside of IRAF, using @olebole hint to promote integer 4 to 8, namely

gfortran -std=legacy -fdefault-integer-8 correl.f spectrum.f libimfort.a libsys.a libvops.a libos.a libf2c.a

I got:

/usr/bin/ld: libf2c.a(main.o): in function `main':
(.text.startup+0x0): multiple definition of `main'; /tmp/ccP4uRco.o:correl.f:(.text+0x22e1): first defined here
/usr/bin/ld: libf2c.a(main.o): in function `main':
(.text.startup+0x9b): undefined reference to `MAIN__'
collect2: error: ld returned 1 exit status

Trying it inside IRAF, xc on Fortran sources fails weirdly on all OPEN/CLOSE statements (I'm adding the lines as comments to xc errrors:

ecl> xc -c correl.f
Error on line 97 of correl.f: p1_const:  bad constant type '1'      #       open(11,file=filen)   
Error on line 97 of correl.f: p1_const:  bad constant type '1'
Error on line 97 of correl.f: p1_const:  bad constant type '1'
Error on line 97 of correl.f: p1_const:  bad constant type '1'
Error on line 101 of correl.f: p1_const:  bad constant type '1'    #       close(11)
Error on line 144 of correl.f: p1_const:  bad constant type '1'    #       open(3,file='rv.cor')
Error on line 144 of correl.f: p1_const:  bad constant type '1'
Error on line 144 of correl.f: p1_const:  bad constant type '1'
Error on line 144 of correl.f: p1_const:  bad constant type '1'
Error on line 145 of correl.f: p1_const:  bad constant type '1'     #      open(7,file='rv.stat')
Error on line 145 of correl.f: p1_const:  bad constant type '1'
Error on line 145 of correl.f: p1_const:  bad constant type '1'
Error on line 145 of correl.f: p1_const:  bad constant type '1'
Error on line 146 of correl.f: p1_const:  bad constant type '1'      #     open(4,file=outim,status='unknown')
Error on line 146 of correl.f: p1_const:  bad constant type '1'
Error on line 146 of correl.f: p1_const:  bad constant type '1'
cc1: fatal error: correl.c: No such file or directory
compilation terminated.

And simalrly for the second file. Needless to say, Google search on "p1_const: bad constant type '1'" does not return any results. Is it really so that no I/O on external files (other than images opened via IRAF) is allowed?

Interestingly, when I converted both files with f2c outside IRAF, I was able to compile and link with xc:

ecl> xc -h correl.c spectrum.c -limfort -lsys -los -lvops
ecl> dir correl.e long+
xb-rwr-r- msz       357720 Jan 17 13:05 correl.e    

but I have yet to check if the executable works properly.

[olebole]

Could you provide a minimal code example here, i.e. a correl.f that is squeezed to a minimum size that shows the problem here?

[msz59]

Sure, below I'm attaching corr.f, stripped version showing the problem:

$ irafcl

  Community IRAF V2.17

  This product includes results achieved by the IRAF64 project in 2006-
  2009 directed by Chisato Yamauchi (C-SODA/ISAS/JAXA).

  Welcome to IRAF.  To list the available commands, type ? or ??.  To get
  detailed information about a command, type `help <command>'.  To run  a
  command  or  load  a  package,  type  its name.   Type  `bye' to exit a
  package, or `logout' to get out  of the CL.

  Visit http://iraf.net if you have questions.

  ***  Using global login file:  /home/msz/.iraf/login.cl
  The following commands or packages are currently defined:

      dataio.     language.   noao.       proto.      sptable.    wcstools.
      fitsutil.   lists.      obsolete.   rvsao.      system.     
      images.     mscred.     plot.       softools.   utilities.  

ecl> softools
      generic     memchk      mkpkg       rmbin       wtar        
      hdbexamine  mkapropos   mktags      rmfiles     xc          
      lroff       mkhelpdb    mkttydata   rtar        xyacc       

softools> xc -c corr.f
Error on line 60 of corr.f: p1_const:  bad constant type '1'
Error on line 60 of corr.f: p1_const:  bad constant type '1'
Error on line 60 of corr.f: p1_const:  bad constant type '1'
Error on line 60 of corr.f: p1_const:  bad constant type '1'
Error on line 64 of corr.f: p1_const:  bad constant type '1'
cc1: fatal error: corr.c: No such file or directory
compilation terminated.
softools> 

c sample of code to show compilation problems
c      
      implicit real*8(a-h,o-z)
      dimension phase(50)
      CHARACTER*20 filen
      parameter (ndim1=3800,nd2=36)
      real              data(ndim1,nd2)
      character*80      inpim, outim, errmsg
      integer           ncols, nlines
      integer           ier, axlen(7), naxis, pixtype, nargs

      nphas=nd2

c --- Get command line arguments.
        call clnarg (nargs)
        if (nargs .eq. 2) then
          call clargc (1, inpim, ier)
          if (ier .ne. 0) goto 91
          call clargc (2, outim, ier)
          if (ier .ne. 0) goto 91
        else
          write (*, '('' input image :'',$)')
          read (*,'(a)') inpim
          write (*, '('' output file:'',$)')
          read (*,'(a)') outim
        endif

c --- Open  images
        call imopen (inpim, 3, ninp, ier)
        if (ier .ne. 0) goto 91

c --- Determine the size and pixel type of the input image.
        call imgsiz (ninp, axlen, naxis, pixtype, ier)
        if (ier .ne. 0) goto 91
        ncols  = axlen(1)
        nlines = axlen(2)
        write(*,*)'Image size is ', ncols,nlines

c ---  load image:
        do 50 j=1,nlines
        call imgl2r(ninp, data(1,j),j, ier)
 50   continue
        goto 92

c     -- Error actions.
 91     call imemsg (ier, errmsg)
        write (*, '('' Error: '', a80)') errmsg
        stop 'at error'
c --- Clean up.
 99     call imclos (ninp, ier)
        if (ier .ne. 0) goto 91
        stop 'at end'

 92     continue

c ... etc.  ...

        write(*,*)"phases file:"
        read(*,*) filen
        open(11,file=filen)
        read(11,*,end=1000,err=1000)(phase(i),i=1,Nphas)
        goto 1001
 1000   Nphas=i-1
 1001   close(11)
        goto 99
        end

[olebole]

After some debugging, I found that the compilation problems with f2c are due to a bug in f2c used by recent IRAF versions; I created pull request #370 to fix it. So, this will be fixed in the next community IRAF version. Until then you are safe to use an external f2c to create C code.

For the compilation for gfortran, I don't have a solution. I would guess that linking the f2c main module to gfortran isn't really supported.