The 'include' statement that failed is,
.
The file name in the INCLUDE subdirectory was, 'asctyp.inc'.
I had thought already that the fault might lie in the case-sensitivity of Linux file names.
I had therefore tried already - or so I thought I had done - to re-BUILD, after, first, re-naming the file in the INCLUDE subdirectory to 'ASCTYP.INC', and I believed it to have made no difference. However, I have just tried it again, and it DID solve the problem.
I must, therefore, have made a mistake when I tried this correction the first time, and I apologise for wasting your time, Jeff.
I am wondering nonetheless, in view of the fact that FORTRAN is normally not case dependent, do you know of any setting one can use in GFortran that will over-ride the case-dependence of Linux for file names? Alternatively, could the Linux version of SF be fitted with a clickable tick-box option in Project Options-->File locations-->Search Directories, to make filenames case-independent?
If neither of these is on the cards then this won't matter all that much, because Linux users should know already that filenames and pathnames mentioned in the source codes have to be case-matched with the names of the files and paths to which they refer, but this will then detract just in a small way from the ability of SF in Linux to use SF projects developed in Windows seamlessly. Users will need to REMEMBER to check projects ported from Windows for INCLUDE filename case compatibility with source codes.
Just to re-confirm, when I changed about six or eight saved filenames in my INCLUDE sub-directory to match the cases of the filenames referred to by dozens of INCLUDE statements in my source codes in this project, SF for Linux did then BUILD it completely successfully, without further alterations of any kind.
---
John