Bugzilla Bug 198875

...
checking pkg-config is at least version 0.7... yes
checking for GLIB - version >= 2.8.0... yes (version 2.14.3)
checking for ffi.h... yes
...
libffi support: yes

This introduces a link dependency on -lffi, and can break the build (c.f. bug
194632).

libffi dependency should be USE-controlled or, failing that, disabled.

Upstreamed: http://bugs.gentoo.org/show_bug.cgi?id=198870

Created an attachment (id=135783) [edit]
libffi.patch

patch to configure.ac

Created an attachment (id=135798) [edit]
libffi.patch

oops

Created an attachment (id=135800) [edit]
libffi.patch

improve

Created an attachment (id=135801) [edit]
pygobject-2.14.0.ebuild

ebuild applying patch - pygobject sucks autotools-wise

After I suffered from this problem (compile died somewhere at something ffi),
applied the two patches above, and I'm happy to report they worked for me.

Sorry, upstream URL should be: http://bugzilla.gnome.org/show_bug.cgi?id=496006

just one question wrt to the choice of the use flag. What is the use of libffi
here ?

(In reply to comment #8)
> just one question wrt to the choice of the use flag. What is the use of libffi
> here ?

It's to allow connecting to signals on native python objects from C by using
libffi to provide C marshallers.  See
http://svn.gnome.org/viewvc/pygobject/trunk/configure.ac?view=log#rev651 and
http://bugzilla.gnome.org/show_bug.cgi?id=353816 .

Uh, so if you define a signal on a gobject.GObject-derived Python class (using
__gsignals__) and thn attach to the signal from C, you need ffi support or will
get:

GLib-GObject-CRITICAL **: g_closure_invoke: assertion `closure->marshal ||
closure->meta_marshal' failed
aborting...

I doubt anyone's using it yet, but it's possible packages will start to require
it in future.

Also looks like the package is not looking in the right place always for
libffi.
/bin/sh ../libtool --mode=link i686-pc-linux-gnu-gcc  -march=athlon-xp
-mcpu=athlon-xp -O3 -pipe -Wall -fno-strict-aliasing -std=c9x   -o _gobject.la
-rpath /usr/lib/python2.4/site-packages/gtk-2.0/gobject -module -avoid-version 
-export-symbols-regex init_gobject _gobject_la-gobjectmodule.lo
_gobject_la-pygboxed.lo _gobject_la-pygenum.lo _gobject_la-pygflags.lo
_gobject_la-pygobject.lo _gobject_la-pygmaincontext.lo
_gobject_la-pygmainloop.lo _gobject_la-pygparamspec.lo
_gobject_la-pygpointer.lo _gobject_la-pygiochannel.lo _gobject_la-pygsource.lo
_gobject_la-pygtype.lo _gobject_la-pygoptioncontext.lo
_gobject_la-pygoptiongroup.lo _gobject_la-ffi-marshaller.lo -pthread
-lgobject-2.0 -lgthread-2.0 -lrt -lglib-2.0 -lffi 
generating symbol list for `_gobject.la'
/usr/bin/nm -B  .libs/_gobject_la-gobjectmodule.o .libs/_gobject_la-pygboxed.o
.libs/_gobject_la-pygenum.o .libs/_gobject_la-pygflags.o
.libs/_gobject_la-pygobject.o .libs/_gobject_la-pygmaincontext.o
.libs/_gobject_la-pygmainloop.o .libs/_gobject_la-pygparamspec.o
.libs/_gobject_la-pygpointer.o .libs/_gobject_la-pygiochannel.o
.libs/_gobject_la-pygsource.o .libs/_gobject_la-pygtype.o
.libs/_gobject_la-pygoptioncontext.o .libs/_gobject_la-pygoptiongroup.o
.libs/_gobject_la-ffi-marshaller.o  | sed -n -e 's/^.*[    
]\([ABCDGIRSTW][ABCDGIRSTW]*\)[         ][     
]*\([_A-Za-z][_A-Za-z0-9]*\)$/\1 \2 \2/p' | /bin/sed 's/.* //' | sort | uniq >
.libs/_gobject.exp
/bin/grep -E -e "init_gobject" ".libs/_gobject.exp" > ".libs/_gobject.expT"
mv -f ".libs/_gobject.expT" ".libs/_gobject.exp"
echo "{ global:" > .libs/_gobject.ver
 cat .libs/_gobject.exp | sed -e "s/\(.*\)/\1;/" >> .libs/_gobject.ver
 echo "local: *; };" >> .libs/_gobject.ver
 i686-pc-linux-gnu-gcc -shared  .libs/_gobject_la-gobjectmodule.o
.libs/_gobject_la-pygboxed.o .libs/_gobject_la-pygenum.o
.libs/_gobject_la-pygflags.o .libs/_gobject_la-pygobject.o
.libs/_gobject_la-pygmaincontext.o .libs/_gobject_la-pygmainloop.o
.libs/_gobject_la-pygparamspec.o .libs/_gobject_la-pygpointer.o
.libs/_gobject_la-pygiochannel.o .libs/_gobject_la-pygsource.o
.libs/_gobject_la-pygtype.o .libs/_gobject_la-pygoptioncontext.o
.libs/_gobject_la-pygoptiongroup.o .libs/_gobject_la-ffi-marshaller.o 
/usr/lib/libgobject-2.0.so -L/usr/lib /usr/lib/libgthread-2.0.so -lrt
/usr/lib/libglib-2.0.so -lffi  -march=athlon-xp -mcpu=athlon-xp -pthread
-Wl,-soname -Wl,_gobject.so -Wl,-version-script -Wl,.libs/_gobject.ver -o
.libs/_gobject.so
/usr/lib/gcc/i686-pc-linux-gnu/3.4.6/../../../../i686-pc-linux-gnu/bin/ld:
cannot find -lffi
collect2: ld returned 1 exit status
make[2]: *** [_gobject.la] Error 1
make[2]: Leaving directory
`/var/tmp/portage/dev-python/pygobject-2.14.0/work/pygobject-2.14.0/gobject'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/var/tmp/portage/dev-python/pygobject-2.14.0/work/pygobject-2.14.0'
make: *** [all] Error 2
 * 
 * ERROR: dev-python/pygobject-2.14.0 failed.

I do have libffi installed:
[ebuild   R   ] dev-libs/libffi-3.4.3  USE="nptl -nls (-uclibc)" 0 kB 

curie-int ~ # equery f libffi |grep libffi.so
/usr/lib/libffi/libffi.so
/usr/lib/libffi/libffi.so.3
/usr/lib/libffi/libffi.so.3.0.0

curie-int ~ # emerge --info
Portage 2.1.4_rc10 (default-linux/x86/2006.1/desktop, gcc-3.4.6, glibc-2.7-r1,
2.6.18-rc3-git7 i686)
=================================================================
System uname: 2.6.18-rc3-git7 i686 AMD Athlon(tm) XP 2500+
Timestamp of tree: Unknown
distcc 2.18.3 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632)
[disabled]
ccache version 2.4 [enabled]
app-shells/bash:     3.2_p17-r1
dev-java/java-config: 1.3.7, 2.1.2-r1
dev-lang/python:     2.4.4-r7
dev-python/pycrypto: 2.0.1-r6
dev-util/ccache:     2.4-r7
sys-apps/baselayout: 1.12.10-r5
sys-apps/sandbox:    1.2.18.1-r2
sys-devel/autoconf:  2.13, 2.61-r1
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10
sys-devel/binutils:  2.17
sys-devel/gcc-config: 1.4.0-r4
sys-devel/libtool:   1.5.24
virtual/os-headers:  2.6.23-r2
ACCEPT_KEYWORDS="x86 ~x86"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-march=athlon-xp -mcpu=athlon-xp -O3 -pipe"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /opt/glftpd/etc /opt/glftpd/ftp-data /opt/openjms/config
/usr/share/config /var/bind /var/qmail/alias /var/qmail/control
/var/vpopmail/domains /var/vpopmail/etc"
CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf
/etc/gconf /etc/php/apache2-php5/ext-active/ /etc/php/cgi-php5/ext-active/
/etc/php/cli-php5/ext-active/ /etc/revdep-rebuild /etc/terminfo
/etc/texmf/web2c /etc/udev/rules.d"
CXXFLAGS="-march=athlon-xp -mcpu=athlon-xp -O3 -pipe"
DISTDIR="/home/gentoo/distfiles"
FEATURES="autoaddcvs buildpkg ccache collision-protect cvs digest distlocks
metadata-transfer sandbox sfperms sign unmerge-orphans userfetch userpriv"
GENTOO_MIRRORS="http://distfiles.gentoo.org
http://distro.ibiblio.org/pub/linux/distributions/gentoo"
MAKEOPTS="-j4"
PKGDIR="/home/gentoo/packages/curie-int/"
PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress
--force --whole-file --delete --delete-after --stats --timeout=180
--exclude=/distfiles --exclude=/local --exclude=/packages
--filter=H_**/files/digest-*"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/gentoo-cvs/gentoo-x86"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC=""
USE="3dnow aalib acl acpi adns aio alsa amd apache2 apm avalon-framework
avalon-logkit berkdb bitmap-fonts cairo cdr cgi clearpasswd cli cracklib crypt
curl divx4linux dri dvd dvdr encode f77 fam fortran frxp gcj gd gdbm geoip gif
gpgme hpn iconv idn imap innodb ipalias ipv6 isdnlog jai java javamail jikes
jms jpeg junit libwww log4j logrotate mad maildir mcal md5sum midi mikmod mmx
mp3 mpeg mpm-prefork mudflap multicall multitarget mysql ncurses nptl objc
offensive ogg openmp pam pcap pcre pdflib perl pic plotutils png pnp ppds pppd
python qmail readline reflection scanner session slp smime snmp socks5 spell
spl sse ssl tetex tiff truetype truetype-fonts type1 type1-fonts udev ungif
unicode usb userlocales v4l v4l2 vhosts vim-syntax vorbis x86 xml xml2 xorg
xvid zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci
emu10k1 emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0
intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci"
ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file
hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route
share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias
authn_anon authn_dbm authn_default authn_file authz_dbm authz_default
authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs
dav_lock deflate dir disk_cache env expires ext_filter file_cache filter
headers include info log_config logio mem_cache mime mime_magic negotiation
rewrite setenvif speling status unique_id userdir usertrack vhost_alias"
ELIBC="glibc" INPUT_DEVICES="keyboard mouse evdev" KERNEL="linux"
LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses
text" USERLAND="GNU" VIDEO_CARDS="apm ark chips cirrus cyrix dummy fbdev glint
i128 i740 i810 imstt mach64 mga neomagic nsc nv r128 radeon rendition s3
s3virge savage siliconmotion sis sisusb tdfx tga trident tseng v4l vesa vga via
vmware voodoo"
Unset:  CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LC_ALL,
LDFLAGS, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS,
PORTAGE_RSYNC_EXTRA_OPTS

Created an attachment (id=139380) [edit]
pygobject-2.14.0.ebuild

pass -L/usr/lib/libffi for libffi support

Robin, please try the above.

Explanation: dev-libs/g-wrap installs libffi to /usr/lib/libffi.so;
dev-libs/libffi installs it to /usr/lib/libffi/libffi.so (see
dev-libs/libffi-3.4.3.ebuild).  Passing LDFLAGS=-L/usr/lib/libffi to configure
should ensure it finds the libffi libffi.so first.

The packages probably can't be installed together anyway; they probably still
collide on /usr/include/libffi/ffi.h.

adding hkBst to get input on the libffi matter. My problem is that libffi is
maintainer-needed and that g-wrap and libffi installs the same library but in
different places. Is adding a blocker on each other a good idea ?

Created an attachment (id=139388) [edit]
libffi.patch

Oops, the -lffi finds its way into gobject.pc, so it needs to have
-L/usr/lib/libffi there as well.  This should help.

Created an attachment (id=139389) [edit]
pygobject-2.14.0.ebuild

for the above

Created an attachment (id=139390) [edit]
libffi.patch

should be more upstream-friendly

(In reply to comment #14)
> adding hkBst to get input on the libffi matter. My problem is that libffi is
> maintainer-needed and that g-wrap and libffi installs the same library but in
> different places. Is adding a blocker on each other a good idea ?

libffi ebuild plain needs to die. Until 199850 is fixed, I'd recommend
hard-disabling this piece of junk, not adding any use flag.

I think until bug #199850 is resolved, it's better to desactivate this feature.

hum patch is inconsistent, it compares $with_ffi while the macro uses
$with_libffi, so if you pass --without-libffi which is what the macro will
match, it will give you [libffi requested, but ffi.h not found] because
$with_ffi is not set with the proper value.

ok, I've commited a modified version of the patch to the tree. I think the
latest if condition of the patch is wrong so I modified it. Currently it
respects the "force off" behavior so it's fine, but please review to make sure
it will behave correctly whenever libffi issue will be fixed.

I am not aware of package collisions for g-wrap but as a precaution it as been
blocking non-existing libffi-4. Having g-wrap install libffi is a workaround
against there not being any libffi-4.

(In reply to comment #22)
> I am not aware of package collisions for g-wrap but as a precaution it as been
> blocking non-existing libffi-4. Having g-wrap install libffi is a workaround
> against there not being any libffi-4.

It collides w/ gcc when compiled w/ USE=gcj (see Bug 200679). libffi should
definitely vanish from g-wrap, doesn't belong there.

If you really want support for this, gcc w/ USE=libffi provides it now.

this is now in 2.14.2.