bug#14901: Trunk bootstrap broken

bug#14901: Trunk bootstrap broken

[Angelo Graziosi]

At least on Cygwin, the rev. 113450 fails to bootstrap:

[...]
gcc -std=gnu99 -c  -Demacs  -I. -I/work/emacs/src  -I../lib 
-I/work/emacs/src/../lib    -D_REENTRANT -I/usr/include/gtk-3.0 
-I/usr/include/atk-1.0 -I/usr/include/at-spi2-atk/2.0 
-I/usr/include/pango-1.0 -I/usr/include/gio-unix-2.0/ 
-I/usr/include/cairo -I/usr/include/gdk-pixbuf-2.0 
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include 
-I/usr/include/harfbuzz -I/usr/include/freetype2 -I/usr/include/pixman-1 
-I/usr/include/libpng15   -DGDK_DISABLE_DEPRECATION_WARNINGS 
-DGLIB_DISABLE_DEPRECATION_WARNINGS -I/usr/include/freetype2 
-D_REENTRANT -I/usr/include/librsvg-2.0 -I/usr/include/glib-2.0 
-I/usr/lib/glib-2.0/include -I/usr/include/gdk-pixbuf-2.0 
-I/usr/include/cairo -I/usr/include/libpng15 -I/usr/include/pixman-1 
-I/usr/include/freetype2   -fopenmp -I/usr/include/ImageMagick 
-I/usr/include/libxml2   -I/usr/include/dbus-1.0 
-I/usr/lib/dbus-1.0/include      -D_REENTRANT -I/usr/include/glib-2.0 
-I/usr/lib/glib-2.0/include   -D_REENTRANT -I/usr/include/gconf/2 
-I/usr/include/dbus-1.0 -I/usr/lib/dbus-1.0/include 
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include 
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include 
-I/usr/include/freetype2     -I/usr/include/freetype2    -MMD -MF 
deps/keyboard.d -MP  -I/usr/include/p11-kit-1   -D_REENTRANT 
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include      -g3 -O2 
/work/emacs/src/keyboard.c
/work/emacs/src/keyboard.c: In function ‘kbd_buffer_get_event’:
/work/emacs/src/keyboard.c:4079:59: error: ‘struct x_display_info’ has 
no member named ‘output_data’
/work/emacs/src/keyboard.c:4080:23: error: ‘struct x_display_info’ has 
no member named ‘output_data’
Makefile:356: recipe for target `keyboard.o' failed
make[2]: *** [keyboard.o] Error 1
make[2]: uscita dalla directory "/work/emacs/Work/src"
Makefile:381: recipe for target `src' failed
make[1]: *** [src] Error 2
make[1]: uscita dalla directory "/work/emacs/Work"
Makefile:992: recipe for target `bootstrap' failed
make: *** [bootstrap] Error 2

I guess these changes are the culprit:

=========================
2013-07-18  Juanma Barranquero

         * keyboard.c (kbd_buffer_get_event):
         * w32term.c (x_focus_changed): Port FOCUS_(IN|OUT)_EVENT 
changes to W32.
         Followup to revno:113431.
=========================


(notice that rev. 113449 bootstraps fine on Kubuntu...)


Ciao,
  Angelo.

bug#14901: Trunk bootstrap broken

[Juanma Barranquero]

On Thu, Jul 18, 2013 at 10:07 PM, Angelo Graziosi
<angelo.graziosi@alice.it> wrote:

> I guess these changes are the culprit:
>
> =========================
> 2013-07-18  Juanma Barranquero
>
>         * keyboard.c (kbd_buffer_get_event):
>         * w32term.c (x_focus_changed): Port FOCUS_(IN|OUT)_EVENT changes to
> W32.
>         Followup to revno:113431.
> =========================

Does it work with the following patch?


=== modified file 'src/keyboard.c'
--- src/keyboard.c 2013-07-18 16:50:05 +0000
+++ src/keyboard.c 2013-07-18 20:22:48 +0000
@@ -4066,7 +4066,7 @@
         }
       else if (event->kind == FOCUS_OUT_EVENT)
         {
-#if defined (HAVE_NS) || defined (HAVE_X11) || defined (WINDOWSNT)
+#if defined (HAVE_NS) || defined (HAVE_X11) || defined (HAVE_NTGUI)

 #ifdef HAVE_NS
           struct ns_display_info *di;
@@ -4080,7 +4080,7 @@
             focused = FRAME_X_DISPLAY_INFO (di)->x_highlight_frame = 0;

           if (! focused) obj = make_lispy_focus_out (frame);
-#endif /* HAVE_NS || HAVE_X11 || WINDOWSNT */
+#endif /* HAVE_NS || HAVE_X11 || HAVE_NTGUI */

           kbd_fetch_ptr = event + 1;
         }

bug#14901: Trunk bootstrap broken

[Angelo Graziosi]

Il 18/07/2013 22.24, Juanma Barranquero ha scritto:
> On Thu, Jul 18, 2013 at 10:07 PM, Angelo Graziosi
> <angelo.graziosi@alice.it> wrote:
>
>> I guess these changes are the culprit:
>>
>> =========================
>> 2013-07-18  Juanma Barranquero
>>
>>          * keyboard.c (kbd_buffer_get_event):
>>          * w32term.c (x_focus_changed): Port FOCUS_(IN|OUT)_EVENT changes to
>> W32.
>>          Followup to revno:113431.
>> =========================
>
> Does it work with the following patch?
>
>
> === modified file 'src/keyboard.c'
> --- src/keyboard.c 2013-07-18 16:50:05 +0000
> +++ src/keyboard.c 2013-07-18 20:22:48 +0000
> @@ -4066,7 +4066,7 @@
>           }
>         else if (event->kind == FOCUS_OUT_EVENT)
>           {
> -#if defined (HAVE_NS) || defined (HAVE_X11) || defined (WINDOWSNT)
> +#if defined (HAVE_NS) || defined (HAVE_X11) || defined (HAVE_NTGUI)
>
>   #ifdef HAVE_NS
>             struct ns_display_info *di;
> @@ -4080,7 +4080,7 @@
>               focused = FRAME_X_DISPLAY_INFO (di)->x_highlight_frame = 0;
>
>             if (! focused) obj = make_lispy_focus_out (frame);
> -#endif /* HAVE_NS || HAVE_X11 || WINDOWSNT */
> +#endif /* HAVE_NS || HAVE_X11 || HAVE_NTGUI */
>
>             kbd_fetch_ptr = event + 1;
>           }
>

No, it fails...

Sincerely, I expected the failure. Why changing WINDOWSNT with 
HAVE_NTGUI should work? Cygwin, ideally, should work as on GNU/Linux. 
Here I do GTK builds.

BTW, rev. 113449 bootstraps fine on Cygwin too.

Ciao,
  Angelo.

bug#14901: Trunk bootstrap broken

[Jan Djärv]

Hello.

18 jul 2013 kl. 23:09 skrev Angelo Graziosi <angelo.graziosi@alice.it>:

> Il 18/07/2013 22.24, Juanma Barranquero ha scritto:
>> On Thu, Jul 18, 2013 at 10:07 PM, Angelo Graziosi
>> <angelo.graziosi@alice.it> wrote:
>> 
>>> I guess these changes are the culprit:
>>> 
>>> =========================
>>> 2013-07-18  Juanma Barranquero
>>> 
>>>         * keyboard.c (kbd_buffer_get_event):
>>>         * w32term.c (x_focus_changed): Port FOCUS_(IN|OUT)_EVENT changes to
>>> W32.
>>>         Followup to revno:113431.
>>> =========================
>> 
>> Does it work with the following patch?
>> 
>> 
>> === modified file 'src/keyboard.c'
>> --- src/keyboard.c 2013-07-18 16:50:05 +0000
>> +++ src/keyboard.c 2013-07-18 20:22:48 +0000
>> @@ -4066,7 +4066,7 @@
>>          }
>>        else if (event->kind == FOCUS_OUT_EVENT)
>>          {
>> -#if defined (HAVE_NS) || defined (HAVE_X11) || defined (WINDOWSNT)
>> +#if defined (HAVE_NS) || defined (HAVE_X11) || defined (HAVE_NTGUI)
>> 
>>  #ifdef HAVE_NS
>>            struct ns_display_info *di;
>> @@ -4080,7 +4080,7 @@
>>              focused = FRAME_X_DISPLAY_INFO (di)->x_highlight_frame = 0;
>> 
>>            if (! focused) obj = make_lispy_focus_out (frame);
>> -#endif /* HAVE_NS || HAVE_X11 || WINDOWSNT */
>> +#endif /* HAVE_NS || HAVE_X11 || HAVE_NTGUI */
>> 
>>            kbd_fetch_ptr = event + 1;
>>          }
>> 
> 
> No, it fails...
> 
> Sincerely, I expected the failure. Why changing WINDOWSNT with HAVE_NTGUI should work? Cygwin, ideally, should work as on GNU/Linux. Here I do GTK builds.
> 
> BTW, rev. 113449 bootstraps fine on Cygwin too.

FWIW, trunk compiles fine on GNU/Linux (with Gtk+ 3 on Fedora 18).

	Jan D.

bug#14901: Trunk bootstrap broken

[Juanma Barranquero]

On Thu, Jul 18, 2013 at 11:09 PM, Angelo Graziosi
<angelo.graziosi@alice.it> wrote:

> Sincerely, I expected the failure. Why changing WINDOWSNT with HAVE_NTGUI
> should work?

Because the point of using HAVE_NTGUI was *excluding* Cygwin. If
Cygwin already satisfies HAVE_X11, why would it fail?

   J

bug#14901: Trunk bootstrap broken

[Juanma Barranquero]

On Thu, Jul 18, 2013 at 11:09 PM, Angelo Graziosi
<angelo.graziosi@alice.it> wrote:

> BTW, rev. 113449 bootstraps fine on Cygwin too.

The only part of revno:113450 relevant to this bug is changing di to
FRAME_X_DISPLAY_INFO (di), and that should work on Cygwin. Please, try
to find why it doesn't.

Thanks,

   Juanma

bug#14901: Trunk bootstrap broken

[Angelo Graziosi]

After Paul's changes, rev. 113451 bootstraps... Thanks.

Ciao,
  Angelo.

bug#14901: Trunk bootstrap broken

[Juanma Barranquero]

> After Paul's changes, rev. 113451 bootstraps... Thanks.

Yes, I just saw. Wonderful.

But in ay case, could you please test that it works by changing WINDOWSNT
to HAVE_NTGUI in that #ifdef? I think it's the right test anyway.

Thanks,

   Juanma

bug#14901: Trunk bootstrap broken

[Angelo Graziosi]

Il 19/07/2013 1.17, Juanma Barranquero ha scritto:
>> After Paul's changes, rev. 113451 bootstraps... Thanks.
>
> Yes, I just saw. Wonderful.
>
> But in ay case, could you please test that it works by changing WINDOWSNT
> to HAVE_NTGUI in that #ifdef? I think it's the right test anyway.
>

Do you mean this patch?

--- work/emacs/src/keyboard.c~	2013-07-18 23:32:06.000000000 +0200
+++ work/emacs/src/keyboard.c	2013-07-19 11:32:26.125000000 +0200
@@ -4068,7 +4068,7 @@
          {
  #if defined HAVE_X11 || defined HAVE_NS
  # define DISPLAY_LIST_INFO(di) (di)
-#elif defined WINDOWSNT
+#elif defined HAVE_NTGUI
  # define DISPLAY_LIST_INFO(di) FRAME_X_DISPLAY_INFO (di)
  #endif
  #ifdef DISPLAY_LIST_INFO


It works.. or at least it passes the critical point (I haven't completed 
the build,, here it take > 30 min.)


Ciao,
  Angelo.

bug#14901: Trunk bootstrap broken

[Juanma Barranquero]

On Fri, Jul 19, 2013 at 11:54 AM, Angelo Graziosi
<angelo.graziosi@alice.it> wrote:

> Do you mean this patch?
>
> -#elif defined WINDOWSNT
> +#elif defined HAVE_NTGUI

Yes.

> It works.. or at least it passes the critical point (I haven't completed the
> build,, here it take > 30 min.)

OK, I'll change it later if you don't report back any trouble.

Thanks,

   Juanma

bug#14901: Trunk bootstrap broken

[Eli Zaretskii]

> From: Juanma Barranquero <lekktu@gmail.com>
> Date: Fri, 19 Jul 2013 01:17:13 +0200
> Cc: 14901@debbugs.gnu.org
> 
> > After Paul's changes, rev. 113451 bootstraps... Thanks.
> 
> Yes, I just saw. Wonderful.

Not so wonderful.  (Doesn't anyone watch compiler warnings these
days?)  Using this:

    struct x_display_info *di;

on anything but X11 is a no-no, since that struct is not portable.
The portable way is to use Display_Info, which a type defined on any
platform we support (including TTYs and even MS-DOS).  No need for the
DISPLAY_INFO macro, either.

> But in ay case, could you please test that it works by changing WINDOWSNT
> to HAVE_NTGUI in that #ifdef? I think it's the right test anyway.

This whole rigmarole with multiple HAVE_* is not needed at all, since
HAVE_WINDOW_SYSTEM captures all of them.

I fixed both of the above in trunk revision 113456.  I've verified
this to compile on GNU/Linux and on MS-Windows, I expect the NS port
will build without any problems, too.

bug#14901: Trunk bootstrap broken

[Juanma Barranquero]

On Fri, Jul 19, 2013 at 1:02 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> Not so wonderful.

I mean't "wonderful that the Cygwin port is not broken".

>  (Doesn't anyone watch compiler warnings these days?)

I compile with full warnings, so unfortunately that means that there
are almost too many warnings. So I check the warnings mostly on
bootstrap (where I compare the previous log with the latest one to
detect any new warning).

> Using this:
>
>     struct x_display_info *di;
>
> on anything but X11 is a no-no, since that struct is not portable.
> The portable way is to use Display_Info, which a type defined on any
> platform we support (including TTYs and even MS-DOS).  No need for the
> DISPLAY_INFO macro, either.

Cool, thanks.

> This whole rigmarole with multiple HAVE_* is not needed at all, since
> HAVE_WINDOW_SYSTEM captures all of them.

We have tons of macros, and only a few of them are described in
admin/CPP-DEFINES.

> I fixed both of the above in trunk revision 113456.  I've verified
> this to compile on GNU/Linux and on MS-Windows, I expect the NS port
> will build without any problems, too.

Thanks.

bug#14901: Trunk bootstrap broken

[Angelo Graziosi]

Il 19/07/2013 13.02, Eli Zaretskii ha scritto:
>
> I fixed both of the above in trunk revision 113456.  I've verified
> this to compile on GNU/Linux and on MS-Windows, I expect the NS port
> will build without any problems, too.
>

On Cygwin, it bootstraps fine too... :-)

Ciao,
Angelo.

bug#14901: Trunk bootstrap broken

[Eli Zaretskii]

> Date: Fri, 19 Jul 2013 14:18:48 +0200
> From: Angelo Graziosi <angelo.graziosi@alice.it>
> CC: Juanma Barranquero <lekktu@gmail.com>, 
>  bug-emacs <bug-gnu-emacs@gnu.org>
> 
> Il 19/07/2013 13.02, Eli Zaretskii ha scritto:
> >
> > I fixed both of the above in trunk revision 113456.  I've verified
> > this to compile on GNU/Linux and on MS-Windows, I expect the NS port
> > will build without any problems, too.
> >
> 
> On Cygwin, it bootstraps fine too... :-)

Thanks for testing.