svn commit: r272808 - head/include/rpcsvc

Benjamin Kaduk bjkfbsd at gmail.com
Fri Oct 10 15:06:05 UTC 2014


On Fri, Oct 10, 2014 at 10:50 AM, Baptiste Daroussin <bapt at freebsd.org>
wrote:

> On Fri, Oct 10, 2014 at 08:03:31AM -0600, Ian Lepore wrote:
> > On Thu, 2014-10-09 at 06:58 +0000, Baptiste Daroussin wrote:
> > > Author: bapt
> > > Date: Thu Oct  9 06:58:33 2014
> > > New Revision: 272808
> > > URL: https://svnweb.freebsd.org/changeset/base/272808
> > >
> > > -%#define ENTRY_VAL(obj, col) \
> > > -   (obj)->EN_data.en_cols.en_cols_val[col].ec_value.ec_value_val
> > > -%#define ENTRY_LEN(obj, col) \
> > > -   (obj)->EN_data.en_cols.en_cols_val[col].ec_value.ec_value_len
> > > +%#define ENTRY_VAL(obj, col)
> (obj)->EN_data.en_cols.en_cols_val[col].ec_value.ec_value_val
> > > +%#define ENTRY_LEN(obj, col)
> (obj)->EN_data.en_cols.en_cols_val[col].ec_value.ec_value_len
> > >  %
> > >  %#ifdef __cplusplus
> > >  %}
> > >
> >
> > How does unwrapping these lines so that they're longer than the style(9)
> > limits fix the build?
> >
> > (I'm not a fan of 80-column limits, it's a rule that needs to die die
> > die.  I'm just curious what the actual fix is here.)
> >
>
> to be honnest here I don't know how rpcgen passes lines to cpp but recent
> GNU cpp
> are failing because of that, looking at dragonfly I have seen they fixed
> the
> same way.
>

rpcgen accpets cpp syntax for including other snippets as rpcgen input, but
also needs to keep the ability to output cpp directives in the generated
code.  The '%' prefix is a sigil that the directive is to be skipped by
rpcgen and instead emitted to the generated code.  I assume that there was
a bug in the handling of continuation lines in this context which causes
things to be confused, but don't know the details of that, just what's
"supposed to be" going on.

-Ben


More information about the svn-src-head mailing list