Bugzilla Bug 213382

Trying to compile pinentry fails:

In file included from /usr/include/gtk-2.0/gtk/gtkactiongroup.h:34,
                 from /usr/include/gtk-2.0/gtk/gtk.h:38,
                 from gtksecentry.c:40:
/usr/include/gtk-2.0/gtk/gtkitemfactory.h:50: warning: function declaration
isn't a prototype
gtksecentry.c:274: error: conflicting types for 'g_malloc'
/usr/include/glib-2.0/glib/gmem.h:47: error: previous declaration of 'g_malloc'
was here
gtksecentry.c:292: error: conflicting types for 'g_malloc0'
/usr/include/glib-2.0/glib/gmem.h:48: error: previous declaration of
'g_malloc0' was here
gtksecentry.c:312: error: conflicting types for 'g_realloc'
/usr/include/glib-2.0/glib/gmem.h:50: error: previous declaration of
'g_realloc' was here
make[2]: *** [gtksecentry.o] Error 1
make[2]: Leaving directory
`/var/tmp/portage/app-crypt/pinentry-0.7.4-r1/work/pinentry-0.7.4/gtk+-2'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/var/tmp/portage/app-crypt/pinentry-0.7.4-r1/work/pinentry-0.7.4'
make: *** [all] Error 2
 *
 * ERROR: app-crypt/pinentry-0.7.4-r1 failed.

I'm using dev-libs/glib-2.16.1.

I found a post on the gnupg devel list which has a commited patch attached.

I tried the patch from
http://www.gossamer-threads.com/lists/engine?do=post_attachment;postatt_id=2874;list=gnupg
(with both pinentry-0.7.3 and 0.7.4-r1) and I'm seeing a segfault before it
even starts up (but I don't know if this is because of the patch). The stack
trace looks like:

#0  0x0805929f in secmem_realloc ()
#1  0xb78f4678 in g_realloc () from /usr/lib/libglib-2.0.so.0
#2  0xb78f9b88 in g_option_group_add_entries () from /usr/lib/libglib-2.0.so.0
#3  0xb7b0a9fc in gdk_add_option_entries_libgtk_only ()
   from /usr/lib/libgdk-x11-2.0.so.0
#4  0xb7cbf6ab in gtk_get_option_group () from /usr/lib/libgtk-x11-2.0.so.0
#5  0xb7cbf762 in gtk_parse_args () from /usr/lib/libgtk-x11-2.0.so.0
#6  0xb7cbf804 in gtk_init_check () from /usr/lib/libgtk-x11-2.0.so.0
#7  0xb7cbf844 in gtk_init () from /usr/lib/libgtk-x11-2.0.so.0
#8  0x0804e800 in main ()
#9  0xb77463ec in __libc_start_main () from /lib/libc.so.6
#10 0x0804d461 in _start ()

Using pinentry 0.7.5 (cf. bug #213383):

test9 pinentry # pinentry-curses 
OK Your orders please

Didn't try anything else. I've only got the curses interface installed.

test9 pinentry # emerge -vpt pinentry

These are the packages that would be merged, in reverse order:

Calculating dependencies... done!
[ebuild   R   ] app-crypt/pinentry-0.7.5  USE="-caps -gtk ncurses -qt3" 0 kB
[1]

Total: 1 package (1 reinstall), Size of downloads: 0 kB
Portage tree and overlays:
 [0] /usr/portage
 [1] /usr/local/portage

(In reply to comment #1)
> I tried the patch from
> http://www.gossamer-threads.com/lists/engine?do=post_attachment;postatt_id=2874;list=gnupg
> (with both pinentry-0.7.3 and 0.7.4-r1) and I'm seeing a segfault before it
> even starts up

I now compiled pinentry with qt3 and gtk flags enabled, and I can startup
pinentry-gtk-2 or pinentry-qt just fine.

> (but I don't know if this is because of the patch). 

Well, either way, without the patch, pinentry doesn't even compile...

The failing pinentry seems to redefine g_malloc and co, but upstream changes
its argument from gulong to gsize, which seems to then be in conflict. Why was
it redefining those functions in the first place?

Oh and relevant glib upstream change here:

http://svn.gnome.org/viewvc/glib/branches/glib-2-16/glib/gmem.h?r1=4563&r2=6413

I have commited a workaround for this. The patches from the mailing-list cause
more breakage than they help to solve the problem in the short-term.