[xdebug-general] Re: 1.3.1: debugclient compile failure

On Sun, 9 May 2004, Roman Neuhauser wrote:

> # derick[@]xdebug.org / 2004-05-09 14:08:41 +0200:
> > On Sun, 9 May 2004, Roman Neuhauser wrote:
> >
> > > # derick[@]xdebug.org / 2004-05-09 12:46:24 +0200:
> > > > On Sun, 9 May 2004, Roman Neuhauser wrote:
> > > >
> > > > > I had to make changes to main.c to get it to compile under FreeBSD 4.9
> > > > > with the system libedit. The code is correct for FreeBSD 5.2 which
> > > > > presumably has newer libedit imported. Could you please introduce some
> > > > > checks into configure?
> > > >
> > > > I can't as libedit doesn't provide a version string.
> > >
> > > I must be missing something, but the autoconf checks are based on
> > > whether a feature works, not on version strings, no?
> >
> > Sometimes, if you want to add it, great... but I'm not going to
> > spend time on it :)
>
> The attached patch solves half the problem: checking for the number
> of arguments el_init() will take; the check for H_SETSIZE/H_EVENT
> is broken, it returns H_EVENT even if histedit.h defines H_SETSIZE.
>
> Any ideas? I don't know any of this stuff.

I finally committed a modified patch to solve the H_SETSIZE thing. If
you use the HEADER_EGREP macro, then it checks cpp's output of "include
histedit.h", in which ofcourse H_SETSIZE has been replaced by nothing
(as this is what CPP is supposed to do). This macro does do it
correctly:

      AC_EGREP_CPP(yes, [
#include <histedit.h>
#ifdef H_SETSIZE
yes
#endif
], [
  AC_DEFINE([XDC_H_SETSIZE], H_SETSIZE, [H_SETSIZE or H_EVENT])
  AC_MSG_RESULT([H_SETSIZE])
], [
  AC_DEFINE([XDC_H_SETSIZE], H_EVENT, [H_SETSIZE or H_EVENT])
  AC_MSG_RESULT([H_EVENT])
])

This fix will be in 2.0.0beta4, which I'll release shortly.

Derick

-- 
Xdebug | http://xdebug.org | xdebug-general@lists.xdebug.org