You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@apr.apache.org by "Roy T. Fielding" <fi...@ebuilt.com> on 2001/03/06 20:47:05 UTC

Re: cvs commit: apr CHANGES configure.in

On Tue, Mar 06, 2001 at 04:54:49AM -0000, rbb@apache.org wrote:
>   +++ configure.in	2001/03/06 04:54:49	1.256
>   @@ -350,6 +350,13 @@
>    dnl threaded poll() and we don't want to use sendfile on early FreeBSD 
>    dnl systems if we are also using threads.
>    
>   +AC_ARG_WITH(sendfile, [  --with-sendfile  Force sendfile to be on or off],
>   +  [ if test "$withval" = "yes"; then
>   +        sendfile="1"
>   +    else
>   +        sendfile="0"
>   +    fi ] )
>   +

This seems to say that --with-sendfile means --without-sendfile.  I think
that --with options should always default to "yes" if no =value is given.
Otherwise, the option should be --disable-sendfile.

....Roy


Re: cvs commit: apr CHANGES configure.in

Posted by Greg Stein <gs...@lyra.org>.
Boolean switches should use AC_ARG_ENABLE(). AC_ARG_WITH is to allow you to
specify "configure with <this> package". The code below should fall out
simply as:

AC_ARG_ENABLE(sendfile, [ --enable-sendfile  Enable or disable sendfile support]
  [ if test "$enable_sendfile" = "yes"; then
       sendfile="1"
    else
       sendfile="0"
    fi
  ])

No switch leaves sendfile unset (e.g. default to the platform availability).
--enable-sendfile sets it to 1, and --disable-sendfile sets it to 0.

Users expect AC_ARG_ENABLE type switches for these things.

(--with-debug and --with-maintainer-mode need to switch, too, for that
 matter)

Cheers,
-g

On Tue, Mar 06, 2001 at 12:10:42PM -0800, rbb@covalent.net wrote:
> On Tue, 6 Mar 2001, Roy T. Fielding wrote:
> 
> > On Tue, Mar 06, 2001 at 04:54:49AM -0000, rbb@apache.org wrote:
> > >   +++ configure.in	2001/03/06 04:54:49	1.256
> > >   @@ -350,6 +350,13 @@
> > >    dnl threaded poll() and we don't want to use sendfile on early FreeBSD
> > >    dnl systems if we are also using threads.
> > >
> > >   +AC_ARG_WITH(sendfile, [  --with-sendfile  Force sendfile to be on or off],
> > >   +  [ if test "$withval" = "yes"; then
> > >   +        sendfile="1"
> > >   +    else
> > >   +        sendfile="0"
> > >   +    fi ] )
> > >   +
> >
> > This seems to say that --with-sendfile means --without-sendfile.  I think
> > that --with options should always default to "yes" if no =value is given.
> > Otherwise, the option should be --disable-sendfile.
> 
> The way that autoconf works, by using AC_ARG_WITH, we get multiple
> options:
> 
> 	--with-sendfile=yes
> 	--with-sendfile
> 	--with-sendfile=no
> 	--without-sendfile
> 
> The first two are the same, as are the last two.
> 
> The logic in that file makes --with-sendfile force sendfile to be on, and
> --without-sendfile to force it to be off.
> 
> Ryan
> 
> _______________________________________________________________________________
> Ryan Bloom                        	rbb@apache.org
> 406 29th St.
> San Francisco, CA 94131
> -------------------------------------------------------------------------------

-- 
Greg Stein, http://www.lyra.org/

Re: cvs commit: apr CHANGES configure.in

Posted by "Roy T. Fielding" <fi...@ebuilt.com>.
> The way that autoconf works, by using AC_ARG_WITH, we get multiple
> options:
> 
> 	--with-sendfile=yes
> 	--with-sendfile
> 	--with-sendfile=no
> 	--without-sendfile

Oh, I see it now -- the option parser for configure adds a value of "yes"
if the with value is not given, and adds a value of "no" if a without
value is not given, so the macro doesn't need to worry about it at all.
Never mind.

....Roy


Re: cvs commit: apr CHANGES configure.in

Posted by rb...@covalent.net.
On Tue, 6 Mar 2001, Roy T. Fielding wrote:

> On Tue, Mar 06, 2001 at 04:54:49AM -0000, rbb@apache.org wrote:
> >   +++ configure.in	2001/03/06 04:54:49	1.256
> >   @@ -350,6 +350,13 @@
> >    dnl threaded poll() and we don't want to use sendfile on early FreeBSD
> >    dnl systems if we are also using threads.
> >
> >   +AC_ARG_WITH(sendfile, [  --with-sendfile  Force sendfile to be on or off],
> >   +  [ if test "$withval" = "yes"; then
> >   +        sendfile="1"
> >   +    else
> >   +        sendfile="0"
> >   +    fi ] )
> >   +
>
> This seems to say that --with-sendfile means --without-sendfile.  I think
> that --with options should always default to "yes" if no =value is given.
> Otherwise, the option should be --disable-sendfile.

The way that autoconf works, by using AC_ARG_WITH, we get multiple
options:

	--with-sendfile=yes
	--with-sendfile
	--with-sendfile=no
	--without-sendfile

The first two are the same, as are the last two.

The logic in that file makes --with-sendfile force sendfile to be on, and
--without-sendfile to force it to be off.

Ryan

_______________________________________________________________________________
Ryan Bloom                        	rbb@apache.org
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------