[Click] Re: IPRewriter "nochange" inputspec syntax broken?

[Eddie Kohler]

Hi Erik,

This was a bug bug bug.  It's fixed in CVS (as soon as it propagates); the patch 
is attached.  Thanks for reporting!  This was totally like a model bug report, 
you are my favorite person for the day.

Eddie


diff -u -u -r1.130 -r1.131
--- lib/confparse.cc    24 Jun 2004 00:04:50 -0000      1.130
+++ lib/confparse.cc    10 Jul 2004 04:56:08 -0000      1.131
@@ -63,7 +63,10 @@
  bool
  cp_eat_space(String &str)
  {
-  return str.substring(cp_skip_space(str.begin(), str.end()), str.end());
+  const char *begin = str.begin(), *end = str.end();
+  const char *space = cp_skip_space(begin, end);
+  str = str.substring(space, end);
+  return space != begin;
  }

  bool



Erik VandeKieft wrote:
> Is there something wrong with the IPRewriter element's "nochange"
> inputspec? I thought you're supposed to be able to specify an output port.
> When I specify one, I get an error. When I don't, it works. Here's a
> trivial example (not what I actually need to get to work, obviously, just
> meant to illustrate the syntax problem):
> 
> When you specify an output port:
> 
> # click -q -e 'InfiniteSource(\<0800>) -> IPRewriter(nochange 0) -> Discard'
> <expr>:1: While configuring
> 'IPRewriter at 2 :: IPRewriter':
>   input spec 0: syntax error; expected `nochange [OUTPUT]'
> Router could not be initialized!
> #
> 
> When you leave it blank:
> 
> # click -q -e 'InfiniteSource(\<0800>) -> IPRewriter(nochange) -> Discard'
> #
> 
> Am I using this element wrong?
> 
> Thanks,
> 
> --Erik Vandekieft