Re: config files substitution with awk

[Paolo Bonzini]

> +/@[a-zA-Z_][a-zA-Z_0-9]*@/ {
> +  nfields = split($ 0, field, "@")
> +  for (i = 1; i <= nfields; i++) {
> +    key = field[i]

The space after $ is useless.  Furthermore, this is wrong in cases like 
address@hidden@address@hidden@.

Rather, you need

  nfields = split($0, field, "@")
  for (i = 3; i <= nfields; ) {
    key = field[i-1]
    if (key ~ /[^a-zA-Z_0-9]/)
      i++;
    else
      i += 2;
    if (key in S)
      sub("@" key "@", S[key])
    else if (key in F) {
      while ((getline aline < F[key]) > 0)
         print(aline)
      close(F[key])
      next
    }

A smaller script to show the idea is this:

awk '{
  nfields = split ($0, field, "@");
  for (i = 3; i <= nfields; ) {
    key = field[i-1];
    if (key ~ /[^a-zA-Z_0-9]/) i++; else { print ">> " key;i+=2; }
  }
}'

with testcases such as

address@hidden@address@hidden
>> bar
address@hidden@@address@hidden
>> bar
>> cde
address@hidden@@address@hidden@
>> bar
>> cde
address@hidden @address@hidden
>> cde
address@hidden @address@hidden@
>> cde
address@hidden @cde@@ghi@
>> cde
>> ghi

Paolo