I wrote: >> .fl.h: >> d=`dirname [EMAIL PROTECTED] ; \ >> f=`basename $@ .h`.fl ; \ >> cp $< $(top_builddir)/$$d/$$f ; \ >> cd $(top_builddir)/$$d ; \ >> fluid -c $$f >> >> BUILT_SOURCES += src/foo.h
Ralf replied: > dirname and basename are not portable to ancient hosts. In practice > they will work fine, though. I'd write > > cp $< $(top_builddir)/$$d/$$f && \ > cd $(top_builddir)/$$d && \ > fluid -c $$f > > ... > Can't the `fluid' program be taught to accept input files with a path > and still output in the current directory? I raised that as a possible feature request on the fltk mailing list and the wizard there provided a lateral thinking solution instead: "Well, if you actually need to generate the files in the output directory (vs. treating them as source files with the object files sent to the output directory), then using: fluid -c $(srcdir)/input.fl is your friend - FLUID will put the output files in the current directory, not the directory of the input file." I hope it's just a question of putting all of the different strands together now. I'll post the result once I've worked it out. Cheers Duncan