I am trying to convert my program (http://cuetools.berlios.de) to use automake and autoconf (my first attempt at using these tools). I am having some difficulty because the code uses two sets of scanners and parsers.
Previously, my hand hacked Makefile had included the following lines: toc_scan.c: LFLAGS+= -Ptoc_yy cue_scan.c: LFLAGS+= -Pcue_yy toc_parse.c: YFLAGS+= -p toc_yy cue_parse.c: YFLAGS+= -p cue_yy (these only work with gmake). I have read the automake manual regarding Yacc and Lex support. So far, I have used the list of recommended defines, and that has allowed me to remove the Yacc command line prefixes in the Makefile. I am unsure what to do now regarding the Lex prefixes. The automake manual seems to indicate that only Yacc requires the variable prefixing hack, but I am getting variable redefined errors when I don't use the Lex command line prefix. I have looked at the gdb source, which is recommeded by the automake manual. However, gdb only seems to have one Lex file, so it only helped me with the Yacc part. In addition, gdb seems to only use autoconf, not automake. Again, this is my first attempt at using the automake/autoconf tools, so I could be wrong about this. Do I have to find a list of symbol names used by Lex, and redefine them to have a prefix? Is there any examples of packages that have multiple Yacc _and_ Lex files that use autoconf/automake? -- Svend