To test, please unset USE libffi in your sys-devel/gcc. It's going to be p.use.masked to prevent wrong linking. Please test and mark stable. Leaving it up to the scheme herd to add arch's to CC.
Adding arch's as it's pretty major qa violation as is. To test, please check Comment #0 first.
I gather you also want [dev-libs, virtual]/libffi marked stable? Otherwise, this can't be. (What happens if you make [dev-libs, virtual]/libffi PDEPENDs in gcc if USE=libffi in gcc?) If this is also a request to mark the libffi's stable, then sparc is OK.
Yes this is a request to mark this version of g-wrap stable because it's using dev-libs or virtual/libffi, which is also a go. I'm fast pacing this because I want to package.use.mask sys-devel/gcc libffi USE flag soon as possible. And on top of that GCC bundles old copy of libffi, we have new virtual/libffi, believe it or not, old g-wrap versions also come with a copy! So please mark stable: latest dev-libs/libffi latest virtual/libffi =dev-libs/g-wrap-1.9.11 so we can move on to having a single, up to date, libffi in tree :)
Dunno about the PDEPEND idea, I'm hoping the USE libffi will eventually entirely die from gcc and this virtual/libffi version of ffi will be kept maintained.
Sparc stable; sparc also stable for dev-scheme/guile-lib-0.1.3 which is a [R]FEPEND for g-wrap-1.9.11. Note that repoman needs a bit of coaxing, because ~hppa, ~ppc64 are keyworded for g-wrap-1.9.11 but not for guile-lib.
x86 stable
Created attachment 194244 [details] build log
Created attachment 194245 [details] environment log
As this become stable I wqsn't able to compile it. Attached you can find build log and environment.
amd64 stable
(In reply to comment #9) > As this become stable I wqsn't able to compile it. > Attached you can find build log and environment. > Please open a new bug for separate issues.
Added 272673 as a blocker since dev-scheme/guile-lib-0.1.3 fails to compile on alpha
(In reply to comment #5) > Sparc stable; sparc also stable for dev-scheme/guile-lib-0.1.3 which is a > [R]FEPEND for g-wrap-1.9.11. Note that repoman needs a bit of coaxing, because > ~hppa, ~ppc64 are keyworded for g-wrap-1.9.11 but not for guile-lib. Marked ~hppa. Readding ppc64.
Stable for HPPA, forced because of: RDEPEND.bad 6 dev-libs/g-wrap/g-wrap-1.9.11.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit-use rland/developer) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit-use rland/desktop) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit-use rland) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11-r1.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit- userland/developer) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11-r1.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit- userland/desktop) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11-r1.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit- userland) ['dev-scheme/guile-lib'] DEPEND.bad 6 dev-libs/g-wrap/g-wrap-1.9.11.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit-use rland/developer) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit-use rland/desktop) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit-use rland) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11-r1.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit- userland/developer) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11-r1.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit- userland/desktop) ['dev-scheme/guile-lib'] dev-libs/g-wrap/g-wrap-1.9.11-r1.ebuild: ~ppc64(default/linux/powerpc/ppc64/2008.0/64bit- userland) ['dev-scheme/guile-lib']
Regarding alpha's dep on bug 274383: hkbst says guile-lib-0.1.6 would a sensible replacement for the broken-on-alpha 0.1.3. I suggest leaving it in ~alpha for 30 days, then stabilizing it and g-wrap-1.9.11. Any objections?
(In reply to comment #15) > Regarding alpha's dep on bug 274383: hkbst says guile-lib-0.1.6 would a > sensible replacement for the broken-on-alpha 0.1.3. I suggest leaving it in > ~alpha for 30 days, then stabilizing it and g-wrap-1.9.11. Any objections? > I don't have any objections, I suggest going with what hkbst says on scheme -issues since everytime I touch it; something breaks. May be luck, or not. ;-) But you might still do virtual/libffi, dev-libs/libffi sooner, since pygobject also deps on them, and it will be required for gnome's stable bug which they are planning.
Stabilized dev-libs/libffi and virtual/libffi.
And finally stabilized =dev-libs/g-wrap-1.9.11 on alpha
ppc stable
This bug is now major because it's blocking us from fixing dep. atoms in libffi to block gcc's libffi.
05 Sep 2009; Brent Baude <ranger@gentoo.org> g-wrap-1.9.11.ebuild: Marking g-wrap-1.9.11 ppc64 for bug 272675