I've just finished testing -n, -d, -p, --trace, and --warn-undefined-variables; they all seem to report correct line numbers.

I started to go down the path of changing job_next_command and new_job where allocated_variable_expand_for_file gets called, but quickly came to the conclusion that altering file->fileinfo.lineno wasn't the way to go at all.

The simplest approach I've found for fixing this is to change gmk_floc:

typedef struct
  {
    const char *filenm;
    unsigned long lineno;
+   unsigned long offset;
  } gmk_floc;

... then make the following modifications:

job.c, around line 1899:

      /* Finally, expand the line.  */
+     cmds->fileinfo.offset = i;
      lines[i] = allocated_variable_expand_for_file (cmds->command_lines[i], file );
    }

+ cmds->fileinfo.offset = 0;
  c->command_lines = lines;

job.c, around line 2099:

    }
+ child->file->cmds->fileinfo.offset = child->command_line;
  return 1;
}


Then at the very least change *some* of the following uses of (...)->lineno to add (...)->offset:

• job.c, child_error()
• job.c, new_job(), during tracing around line 2052
• commands.c, print_commands()
• file.c, line 273 in rehash_file()
• function.c, line 1712
• variable.c, line 1629

... and lastly (I think), in read.c and rule.c: change all initializations of "lineno" to also initialize "offset".

Hello


$ make -f makefile.mak
makefile.mak:2: One
makefile.mak:2: Two
makefile.mak:2: Three
Hello


makefile.mak (not sure if the TABs have been lost)
all:
    echo Hello
    $(warning One)
    $(warning Two)
    $(warning Three)


I can pay a bug bounty of $50 for a fix to the developer or FSF :)
Thank you
Jonny