John,

> I was wondering if you could help with the problem I mentioned recently when compiling Starlink on Fedora 28.

I was hoping somebody else with Fedora could help with practical
tips.

> I have found that one of my workarounds has resulted in a problem - I detailed it below in 1) between then asterisks

Please could you send me the applications/figaro/config.log file. 
Amongst a number of checks of features in the Fortran compiler, there 
should be a test for LOC or %LOC.  In mine on Mint the relevant part 
reads:

configure:18083: checking whether gfortran has the %LOC extension
configure:18115: result: no
configure:18128: checking whether gfortran supports intrinsic LOC
configure:18147: result: yes

indicating that compiler supports the LOC() syntax.

> In addition to the problem with plotting I mentioned in 1), when I run 
> FIGARO-isubset on and NDF file it doesn't actually work properly and 
> prints

KAPPA:NDFCOPY is better.  You can subset by pixels or world 
co-ordinates, it's not limited to two dimensions, and it preserves the 
WCS information.

> I think the # sections in the source code (See 1) below) are meant to 
> be place holders or substituted into the code if necessary - rather 
> than turned into print statements. It seems as though my build is 
> failing the HAVE_PERCENTLOC and HAVE_INTRINSIC_LOC for some reason - 
> maybe the line Do not have LOC or %LOC should never be reached. Any 
> idea why that might be or what additional tests I might perform? I 
> wonder if I am missing some devel file somewhere.

Yes, there is a preprocessor that edits the Fortran in these .F files.
The configure.ac has these lines to set the boolean HAVE_PERCENTLOC
and HAVE_INTRINSIC_LOC.

AC_FC_HAVE_PERCENTLOC()
AC_FC_CHECK_INTRINSICS(LOC)

If for some reason you have neither, then the error message you see 
appears.

I'd still like to know all the extra packages that you needed to 
install, so that I can update the main README in the source.

Malcolm

----
Starlink User Support list
For list configuration, including subscribing to and unsubscribing from the list, see
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=STARLINK