10:16 a.m.
> "Jeff" == Jeff Sparkes <jsparkes(a)gmail.com>
writes:
On Wed, Nov 13, 2013 at 5:44 AM, Uwe Brauer <oub(a)mat.ucm.es>
wrote:
> gdk_x11_window_lookup_for_display
>
Looks like you have a Gtk older than 2.24. I've checked in a
fix.
Here are the error messages:
,----
| event-gtk.c:395:34: error: use of undeclared identifier 'GDK_KEY_Shift_Lock'
| if (gd->lock_interpretation == GDK_KEY_Shift_Lock)
| ^
| event-gtk.c:399:14: error: use of undeclared identifier 'GDK_KEY_A'
| ((sym >= GDK_KEY_A) && (sym <= GDK_KEY_Z)) ||
| ^
| event-gtk.c:438:17: error: use of undeclared identifier 'GDK_KEY_exclam'
| if (keysym >= GDK_KEY_exclam && keysym <= GDK_KEY_asciitilde)
| ^
| event-gtk.c:448:10: error: use of undeclared identifier 'GDK_KEY_BackSpace'
| case GDK_KEY_BackSpace: return QKbackspace;
| ^
| event-gtk.c:449:10: error: use of undeclared identifier 'GDK_KEY_Tab'
| case GDK_KEY_Tab: return QKtab;
| ^
| event-gtk.c:450:10: error: use of undeclared identifier 'GDK_KEY_Linefeed'
| case GDK_KEY_Linefeed: return QKlinefeed;
| ^
| event-gtk.c:451:10: error: use of undeclared identifier 'GDK_KEY_Return'
| case GDK_KEY_Return: return QKreturn;
| ^
| event-gtk.c:452:10: error: use of undeclared identifier 'GDK_KEY_Escape'
| case GDK_KEY_Escape: return QKescape;
| ^
| event-gtk.c:453:10: error: use of undeclared identifier 'GDK_KEY_space'
| case GDK_KEY_space: return QKspace;
| ^
| event-gtk.c:454:10: error: use of undeclared identifier 'GDK_KEY_Delete'
| case GDK_KEY_Delete: return QKdelete;
| ^
| event-gtk.c:455:10: error: use of undeclared identifier 'GDK_KEY_Page_Up'
| case GDK_KEY_Page_Up: return KEYSYM("prior");
| ^
| event-gtk.c:456:10: error: use of undeclared identifier 'GDK_KEY_Page_Down'
| case GDK_KEY_Page_Down: return KEYSYM("next");
| ^
| event-gtk.c:498:7: error: use of undeclared identifier 'GDK_KEY_Shift_L'
| if (IS_MODIFIER_KEY (event->keyval) || (event->keyval == GDK_KEY_Mode_switch))
| ^
| event-gtk.c:102:18: note: instantiated from:
| ((((keysym) >= GDK_KEY_Shift_L) && ((keysym) <= GDK_KEY_Hyper_R)) \
| ^
| event-gtk.c:1273:14: warning: unused variable 'disp' [-Wunused-variable]
| GdkDisplay *disp = gdk_screen_get_display (screen);
| ^
| 14 diagnostics generated.
| make[1]: *** [event-gtk.o] Error 1
| make[1]: Leaving directory
`/home/oub/ALLES/Add-Import/xemacs-gtk/jsparkes-xemacs-gtk-456e7527e552/src'
| make: *** [src] Error 2
`----
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:08 p.m.
New subject: error when compiling xemacs-gtk
On Thu, Nov 14, 2013 at 5:16 AM, Uwe Brauer <oub(a)mat.ucm.es> wrote:
GDK_KEY_Tab
I haven't been able to determine when Gtk 2 added the GDK_KEY prefixes for
compatibility with Gtk 3, so I've ifdefed the cases for HAVE_GTK2.
Porting backwards is no fun.. try it again.
--
Jeff Sparkes
jsparkes(a)gmail.com
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:35 p.m.
New subject: error when compiling xemacs-gtk
> "Jeff" == Jeff Sparkes <jsparkes(a)gmail.com>
writes:
On Thu, Nov 14, 2013 at 5:16 AM, Uwe Brauer <oub(a)mat.ucm.es>
wrote:
> GDK_KEY_Tab
I haven't been able to determine when Gtk 2 added the GDK_KEY
prefixes for
compatibility with Gtk 3, so I've ifdefed the cases for HAVE_GTK2.
Porting backwards is no fun.. try it again.
I am sorry. I tried the version out and this time it is a different
error:
,----
| redisplay-gtk.c:72:3: error: use of undeclared identifier
'cairo_rectangle_int_t'; did you mean
| 'cairo_rectangle_list_t'?
| cairo_rectangle_int_t r;
| ^
| redisplay-gtk.c:86:3: error: use of undeclared identifier 'r'
| r.x = FRAME_LEFT_INTERNAL_BORDER_START (f);
| ^
| redisplay-gtk.c:87:3: error: use of undeclared identifier 'r'
| r.width = (FRAME_RIGHT_INTERNAL_BORDER_END (f) - r.x);
| ^
| redisplay-gtk.c:92:3: error: use of undeclared identifier 'r'
| r.y = FRAME_TOP_INTERNAL_BORDER_START (f) - 1;
| ^
| redisplay-gtk.c:93:3: error: use of undeclared identifier 'r'
| r.height = (FRAME_BOTTOM_INTERNAL_BORDER_END (f) - r.y);
| ^
| redisplay-gtk.c:103:29: error: use of undeclared identifier 'r'
| gdk_window_clear_area (w, r.x, r.y, r.width, r.height);
| ^
| 6 diagnostics generated.
| make[1]: *** [redisplay-gtk.o] Error 1
| make[1]: Leaving directory
`/home/oub/ALLES/Add-Import/xemacs-gtk/jsparkes-xemacs-gtk-143f9f4efe56/src'
| make: *** [src] Error 2
`----
Uwe
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:11 p.m.
New subject: error when compiling xemacs-gtk
Uwe Brauer writes:
>> "Jeff" == Jeff Sparkes <jsparkes(a)gmail.com>
writes:
> On Wed, Nov 13, 2013 at 5:44 AM, Uwe Brauer <oub(a)mat.ucm.es> wrote:
>> gdk_x11_window_lookup_for_display
>>
> Looks like you have a Gtk older than 2.24. I've checked in a fix.
Here are the error messages:
,----
| event-gtk.c:395:34: error: use of undeclared identifier 'GDK_KEY_Shift_Lock'
Uwe, please post your Installation file which is generated by
configure, and config.log. It looks to me like you're only getting
part of GTK+ somehow.
Also, post the result of "pkg-config --modversion gtk+-2.0", which
should tell us if you really are building with an older GTK+.
Jeff, I'm having trouble on the Mac. I do have GTK+ 2.24.22, and I get
a build and the automatic test suite runs OK. But as soon as I try to
run it with the GUI, it crashes:
(gdb) run -vanilla
Starting program: /Users/steve/src/XEmacs/21.5/xemacs-gtk/+build/src/xemacs -vanilla
Xlib: extension "RANDR" missing on display
"/tmp/launch-vyVDPM/org.x:0".
Program received signal SIGSEGV, Segmentation fault.
0x00000001018e68d6 in g_type_fundamental ()
from /opt/local/lib/libgobject-2.0.0.dylib
(gdb) bt 5
#0 0x00000001018e68d6 in g_type_fundamental ()
from /opt/local/lib/libgobject-2.0.0.dylib
#1 0x00000001018d88b4 in g_object_new ()
from /opt/local/lib/libgobject-2.0.0.dylib
#2 0x00000001001b4ba6 in gtk_xemacs_new (f=0x0)
at /Users/steve/src/XEmacs/21.5/xemacs-gtk/src/gtk-xemacs.c:123
#3 0x000000010018f4fa in gtk_init_device (d=0x1022359b0, unused_props=...)
at /Users/steve/src/XEmacs/21.5/xemacs-gtk/src/device-gtk.c:272
#4 0x0000000100077e7e in Fmake_device (type=..., connection=..., props=...)
at /Users/steve/src/XEmacs/21.5/xemacs-gtk/src/device.c:662
(More stack frames follow...)
(gdb) quit
The warning about the RANDR extension happens with most X programs (I
think I have an old Xquartz server or something). I doubt it's
relevant but I include it for completeness. If you haven't seen
anything like this crash, don't spend too much time on it. It looks
like it's deep in libgobject. I have checked the libraries, and all
of them except the basic libc come from MacPorts, which is up to date
so they should be consistent. However, my version of Mac OS X is old
(Snow Leopard) and I think my most recent updates to GTK and friends
were binary packages -- it's possible that the MacPorts maintainers
are assuming Lion or more recent.
I also have a patch that fixes the pkg-config version detection
bogosity (I guess use of --version is left over from the days when
gtk-config was distributed with GTK+, now it refers to the version of
pkg-config d'oh :-). I did check for other uses of pkg-config
--version, there weren't any.
diff -r 456e7527e552 ChangeLog
--- a/ChangeLog Wed Nov 13 15:15:01 2013 -0500
+++ b/ChangeLog Thu Nov 14 22:34:51 2013 +0900
＠＠ -1,3 +1,9 ＠＠
+2013-11-14 Stephen J. Turnbull <stephen(a)xemacs.org>
+
+ * configure.ac (GTK+): Restore missing $ in variable reference.
+ Use pkg-config --modversion (not --version) to get version of package.
+ Adjust GTK+ version reporting (2.24.22 is known good).
+
2013-09-15 Jeff Sparkes <jsparkes(a)gmail.com>
* configure.ac: Force -lX11 for gtk to link XKeycodeToKeysym.
diff -r 456e7527e552 configure
--- a/configure Wed Nov 13 15:15:01 2013 -0500
+++ b/configure Thu Nov 14 22:34:51 2013 +0900
＠＠ -12246,7 +12246,7 ＠＠
$as_echo_n "checking for GTK configuration script... " >&6; }
for possible in gtk+-2.0
do
- possible_version=`pkg-config {possible} --version 2> /dev/null`
+ possible_version=`pkg-config ${possible} --modversion 2> /dev/null`
if test "x${possible_version}" != "x"; then
GTK_CONFIG="pkg-config ${possible}"
case "${possible_version}" in
＠＠ -12255,15 +12255,7 ＠＠
with_gtk=no
break
;;
- 0.22*)
- with_gtk=yes
- break
- ;;
- 0.25*)
- with_gtk=yes
- break
- ;;
- 0.26*)
+ 2.24.22)
with_gtk=yes
break
;;
＠＠ -12279,7 +12271,7 ＠＠
if test "${GTK_CONFIG}" != "no"; then
{ $as_echo "$as_me:${as_lineno-$LINENO}: checking gtk version"
>&5
$as_echo_n "checking gtk version... " >&6; }
- GTK_VERSION=`${GTK_CONFIG} --version`
+ GTK_VERSION=`${GTK_CONFIG} --modversion`
{ $as_echo "$as_me:${as_lineno-$LINENO}: result: ${GTK_VERSION}"
>&5
$as_echo "${GTK_VERSION}" >&6; }
diff -r 456e7527e552 configure.ac
--- a/configure.ac Wed Nov 13 15:15:01 2013 -0500
+++ b/configure.ac Thu Nov 14 22:34:51 2013 +0900
＠＠ -2860,7 +2860,7 ＠＠
AC_MSG_CHECKING(for GTK configuration script)
for possible in gtk+-2.0
do
- possible_version=`pkg-config {possible} --version 2> /dev/null`
+ possible_version=`pkg-config ${possible} --modversion 2> /dev/null`
if test "x${possible_version}" != "x"; then
GTK_CONFIG="pkg-config ${possible}"
case "${possible_version}" in
＠＠ -2868,15 +2868,7 ＠＠
with_gtk=no
break
;;
- 0.22*)
- with_gtk=yes
- break
- ;;
- 0.25*)
- with_gtk=yes
- break
- ;;
- 0.26*)
+ 2.24.22)
with_gtk=yes
break
;;
＠＠ -2889,7 +2881,7 ＠＠
if test "${GTK_CONFIG}" != "no"; then
AC_MSG_CHECKING(gtk version)
- GTK_VERSION=`${GTK_CONFIG} --version`
+ GTK_VERSION=`${GTK_CONFIG} --modversion`
AC_MSG_RESULT(${GTK_VERSION})
AC_MSG_CHECKING(gtk libs)
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
2:20 p.m.
New subject: error when compiling xemacs-gtk
> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org> writes:
>
> Here are the error messages:
>
> ,----
> | event-gtk.c:395:34: error: use of undeclared identifier
'GDK_KEY_Shift_Lock'
Uwe, please post your Installation file which is generated by
Do you mean the Makefile? I attach it below, together the config.log
and the config.status
config.cache
Installation
files
configure, and config.log. It looks to me like you're only getting
part of GTK+ somehow.
Also, post the result of "pkg-config --modversion
gtk+-2.0", which
should tell us if you really are building with an older GTK+.
Ok:
pkg-config --modversion gtk+-2.0
2.20.1
Uwe
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
1:01 p.m.
New subject: error when compiling xemacs-gtk
On Thu, Nov 14, 2013 at 9:11 AM, Stephen J. Turnbull <stephen(a)xemacs.org>wrote:
Jeff, I'm having trouble on the Mac. I do have GTK+ 2.24.22, and
I get
a build and the automatic test suite runs OK. But as soon as I try to
run it with the GUI, it crashes:
(gdb) run -vanilla
Starting program:
/Users/steve/src/XEmacs/21.5/xemacs-gtk/+build/src/xemacs -vanilla
Xlib: extension "RANDR" missing on display
"/tmp/launch-vyVDPM/org.x:0".
Program received signal SIGSEGV, Segmentation fault.
0x00000001018e68d6 in g_type_fundamental ()
from /opt/local/lib/libgobject-2.0.0.dylib
(gdb) bt 5
#0 0x00000001018e68d6 in g_type_fundamental ()
from /opt/local/lib/libgobject-2.0.0.dylib
#1 0x00000001018d88b4 in g_object_new ()
from /opt/local/lib/libgobject-2.0.0.dylib
#2 0x00000001001b4ba6 in gtk_xemacs_new (f=0x0)
at /Users/steve/src/XEmacs/21.5/xemacs-gtk/src/gtk-xemacs.c:123
#3 0x000000010018f4fa in gtk_init_device (d=0x1022359b0, unused_props=...)
at /Users/steve/src/XEmacs/21.5/xemacs-gtk/src/device-gtk.c:272
#4 0x0000000100077e7e in Fmake_device (type=..., connection=...,
props=...)
at /Users/steve/src/XEmacs/21.5/xemacs-gtk/src/device.c:662
(More stack frames follow...)
(gdb) quit
The warning about the RANDR extension happens with most X programs (I
think I have an old Xquartz server or something). I doubt it's
relevant but I include it for completeness. If you haven't seen
anything like this crash, don't spend too much time on it. It looks
like it's deep in libgobject. I have checked the libraries, and all
of them except the basic libc come from MacPorts, which is up to date
so they should be consistent. However, my version of Mac OS X is old
(Snow Leopard) and I think my most recent updates to GTK and friends
were binary packages -- it's possible that the MacPorts maintainers
are assuming Lion or more recent.
It certainly is deep in GObject, and pretty early in initialization. I'm
not sure how to debug that.
My Mac is a PowerBook G4 running Leopard. I haven't tried to compile on
that yet. :-)
I also have a patch that fixes the pkg-config version detection
bogosity (I guess use of --version is left over from the days when
gtk-config was distributed with GTK+, now it refers to the version of
pkg-config d'oh :-). I did check for other uses of pkg-config
--version, there weren't any.
That code dates back to Gtk 1.X. I've updated my configure.ac based on
your changes.
I've flipped the configure switch to pick GTK3 first, or fallback to GTK2.
I hope nobody wants to use GTK 2 even if they have 3..
--
Jeff Sparkes
jsparkes(a)gmail.com
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
6:24 p.m.
New subject: error when compiling xemacs-gtk
Stephen J. Turnbull writes:
Jeff, I'm having trouble on the Mac. I do have GTK+ 2.24.22, and
I get
a build and the automatic test suite runs OK. But as soon as I try to
run it with the GUI, it crashes:
(gdb) run -vanilla
Starting program: /Users/steve/src/XEmacs/21.5/xemacs-gtk/+build/src/xemacs -vanilla
Xlib: extension "RANDR" missing on display
"/tmp/launch-vyVDPM/org.x:0".
Program received signal SIGSEGV, Segmentation fault.
0x00000001018e68d6 in g_type_fundamental ()
from /opt/local/lib/libgobject-2.0.0.dylib
OK, I've managed to repair this by changing the type of GType from
"guint" to "GType" in a couple places. diff follows, but I don't
know
if it's appropriate to commit (GType might be GTK 3-only?)
N.B. contrary to the quote above, I am now building with GTK+ 3.10.1.
Unfortunately, now I get no toolbar (but space is reserved for it),
and the menubar is quite busted (no visible labels and the active
regions on the menubar for each submenu are about a dozen pixels wide.
However, I can activate the submenus and was able to quit using File |
Exit XEmacs. The tabs are somewhat broken, which is probably related
to the many thousands (literally) of repetitions of the warning:
(xemacs:50428): Gtk-WARNING **: GtkXEmacs 0x102bf5830 is mapped but visible child
GtkNotebook 0x1034c9510 is not mapped
Ah, note that G_TYPE_INVALID is actually ((GType) 0) according to the
web docs.
diff -r d7342b18c0c6 src/gtk-xemacs.c
--- a/src/gtk-xemacs.c Wed Nov 20 00:06:22 2013 -0500
+++ b/src/gtk-xemacs.c Thu Nov 21 02:53:11 2013 +0900
＠＠ -60,12 +60,12 ＠＠
static gint gtk_xemacs_expose (GtkWidget *widget, GdkEventExpose *event);
#endif
-guint
+GType
gtk_xemacs_get_type (void)
{
- static guint xemacs_type = 0;
+ static GType xemacs_type = G_TYPE_INVALID;
- if (!xemacs_type)
+ if (xemacs_type == G_TYPE_INVALID)
{
xemacs_type =
g_type_register_static_simple (GTK_TYPE_FIXED,
＠＠ -77,6 +77,7 ＠＠
(GTypeFlags) 0);
}
+ assert(xemacs_type != G_TYPE_INVALID);
return xemacs_type;
}
diff -r d7342b18c0c6 src/gtk-xemacs.h
--- a/src/gtk-xemacs.h Wed Nov 20 00:06:22 2013 -0500
+++ b/src/gtk-xemacs.h Thu Nov 21 02:53:11 2013 +0900
＠＠ -47,7 +47,7 ＠＠
GtkFixedClass parent_class;
};
-guint gtk_xemacs_get_type (void);
+GType gtk_xemacs_get_type (void);
GtkWidget *gtk_xemacs_new (struct frame *f);
END_C_DECLS
Regards,
Steve
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
6:48 p.m.
New subject: error when compiling xemacs-gtk
On Wed, Nov 20, 2013 at 1:24 PM, Stephen J. Turnbull <stephen(a)xemacs.org>wrote:
Stephen J. Turnbull writes:
> Jeff, I'm having trouble on the Mac. I do have GTK+ 2.24.22, and I get
> a build and the automatic test suite runs OK. But as soon as I try to
> run it with the GUI, it crashes:
>
> (gdb) run -vanilla
> Starting program:
/Users/steve/src/XEmacs/21.5/xemacs-gtk/+build/src/xemacs -vanilla
> Xlib: extension "RANDR" missing on display
"/tmp/launch-vyVDPM/org.x:0".
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x00000001018e68d6 in g_type_fundamental ()
> from /opt/local/lib/libgobject-2.0.0.dylib
OK, I've managed to repair this by changing the type of GType from
"guint" to "GType" in a couple places. diff follows, but I don't
know
if it's appropriate to commit (GType might be GTK 3-only?)
N.B. contrary to the quote above, I am now building with GTK+ 3.10.1.
Its fine for GTK 2. I've applied the patch, and found one other case.
Unfortunately, now I get no toolbar (but space is reserved for it),
I've wasted countless hours trying to get that toolbar to appear. I just
can't figure it out.
and the menubar is quite busted (no visible labels and the active
regions on the menubar for each submenu are about a dozen pixels wide.
Yes, there's code do display emacs keys in the menu labels. It currently
works for GTK 2 only, I should have it fixed soon.
However, I can activate the submenus and was able to quit using File
|
Exit XEmacs. The tabs are somewhat broken, which is probably related
to the many thousands (literally) of repetitions of the warning:
(xemacs:50428): Gtk-WARNING **: GtkXEmacs 0x102bf5830 is mapped but
visible child GtkNotebook 0x1034c9510 is not mapped
Yesterday I checked in a partial fix that fixes the tab appearance, but not
that warning. I haven't worked how the instantiator gets the notebook tabs
to work.
I also bought a copy of Snow Leopard on ebay. If I can get it to run in a
virtual machine, I'll be able to test for OS X.
--
Jeff Sparkes
jsparkes(a)gmail.com
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta
4009
days inactive
4009
days old
7 comments
3 participants
participants (3)
-
Jeff Sparkes -
Stephen J. Turnbull -
Uwe Brauer