Re: CVS commit: pkgsrc/editors/emacs20

[Ryo ONODERA <ryo_on%yk.rim.or.jp@localhost>]

Hi,

From: David Holland <dholland-pkgchanges%netbsd.org@localhost>, Date: Sat, 6 
Jul 2013 16:39:12 +0000

> On Sat, Jul 06, 2013 at 01:36:26PM +0900, Ryo ONODERA wrote:
>  > > On Fri, Jul 05, 2013 at 01:46:13PM +0000, Ryo ONODERA wrote:
>  > >  > Modified Files:
>  > >  >       pkgsrc/editors/emacs20: distinfo
>  > >  >       pkgsrc/editors/emacs20/patches: patch-ab
>  > >  > 
>  > >  > Log Message:
>  > >  > Fix x11 option build.
>  > > 
>  > > @@ -72,8 +72,7 @@
>  > >   #include "confdefs.h"
>  > >   #include <assert.h>
>  > >   Syntax Error
>  > > -+Should \\
>  > > -+be one line.
>  > > ++Should be one line.
>  > >   EOF
>  > > 
>  > > I don't think you wanted that... what were you trying to do?
>  > 
>  > This test case is for gcc -E -traditional-cpp detection.
>  > I want to disable this detection, so it is not the best way,
>  > it is ugly, but making the test one line is easiest way of disabling
>  > -traditional-cpp detection.
> 
> The logic there was intentionally patched in; if you think it should
> go away, it should be removed entirely rather than being neutered.
> 
> However, I don't think it should go away: probing for a cpp that will
> work fixed the build on a number of platforms and until the tools
> wrappers are prepared to provide transparent support for "gcc -E
> -traditional-cpp", which they aren't, this logic is still needed.

I see. But I have no motivation for supporting such old emacsen.
Please improve it freely.

> What's the build problem? I build it routinely with X support (and one
> of these days I'm going to bring back emacs20-nox11 to reduce the
> number of special cases in the emacs infrastructure) and I haven't had
> trouble, but I haven't pulled the absolute latest X updates yet.
> 
> There isn't anywhere I can see in the makefiles that it uses the cpp
> it probes on source files, except for one case involving alloca that
> shouldn't affect netbsd (and that might or might not use the probed
> cpp anyway...) so I wonder if you're running into some other problem.


With emacs 21 and emacs20 with x11 option,
gcc -E -traditional-cpp and 6.99.23's native X generates
the following error to config.log.
This error results disabling x11 support.

configure:3588: checking for X
configure:3655: gcc -E -traditional-cpp -I/usr/X11R6/include -I/usr/pkg/include 
-I/usr/local/include -L/usr/pkg/lib -L/usr/local/lib  -DTERMINFO 
-I/usr/pkg/include -I/usr/include -I/usr/X11R7/include conftest.c >/dev/null 
2>conftest.out
In file included from 
/usr/tmp/pkgsrc/editors/emacs21/work/.buildlink/include/X11/Xlib.h:48:0,
                 from 
/usr/tmp/pkgsrc/editors/emacs21/work/.buildlink/include/X11/Intrinsic.h:54,
                 from configure:3652:
/usr/tmp/pkgsrc/editors/emacs21/work/.buildlink/include/X11/Xfuncproto.h:135:0: 
error: syntax error in macro parameter list
configure: failed program was:
#line 3650 "configure"
#include "confdefs.h"
#include <X11/Intrinsic.h>

Of course, you can bypass this error with another X11 file inclusion,
but in build stage, you will see same syntax error message again.

You can reproduce on pkgsrc-2013Q2 branch with 6.99.23 of NetBSD.

--
Ryo ONODERA // ryo_on%yk.rim.or.jp@localhost
PGP fingerprint = 82A2 DC91 76E0 A10A 8ABB  FD1B F404 27FA C7D1 15F3