> > Is the BNF used directly "as-is" with YACC to generate a
> parser, or is
> > the BNF in pf.conf manpage an edited excerpt?
>
> The parser is not generated from the BNF syntax (never was),
> but rather the BNF is a manual excerpt (not all productions
> are listed), hence the possible discrepancies.
>
ok, so the syntax parser is "hand crafted"? nice :)
I was looking trough the source for the "original" BNF, but
could not find one... now I know why.
another suggestion to the BNF:
in the filteropt rule; the "fragment" keyword could be removed,
and the rule fragmentation could be changed from:
fragmentation = [ "fragment reassemble" | "fragment crop" |
"fragment drop-ovl" ]
to:
fragmentation = "fragment" [ "reassemble" | "crop" |
"drop-ovl" ]
and in the filteropt rule :
filteropt = ...
| "fragment" <-- drop this
...
| fragmentation
...
still not important, but the "fragment" keyword is redundant and
the fragmentation rule is not referenced anywhere else...
Hm, the fragmentation rule is used in scrub rules, but I have no
idea what the standalone "fragment" keyword does...
also, the "number" keyword in the bandwidth-spec rule should probably
not be quoted, as it refers to a number and not the string..
The hex-key and the string-key in the pooltype rule is never explained
or defined, nor are they defined as keywords. Of course one could
assume the complete obvious; it's a hexadecimal key and a string key.
my last comment on the BNF for now is the sticky-address option
also in pooltype... I'm assuming the author was tired, and nearing
the end of this extreeemly long BNF and forgot to define it as a
keyword: "sticky-address"?
I am a bit tired myself now, so I could of course be wrong...
Magne
