Remove enforced --foreign in autoreconf

Hi,

Since 2004 we've been forcibly passing --foreign to automake in autoreconf. This is a change from upstream behaviour and can actually break packages that expect GNU behaviour (admittedly the failure case reported was "interesting"), and I'm on a bit of a mission to remove patches that don't have a solid reason behind them.

Forcing --foreign doesn't seem that important these days as the majority of packages are telling automake they're foreign, There are a few packages that need patches for this change, notably they're mostly packages that haven't seen a release for ten years.

This series is build tested against a world build in oe-core for NUC, so I'd call it a demonstration that the patch is viable and would appreciate others running build tests against other machines or layers.

Ross

The following changes since commit 9e99bcffec594cb5b4c3533200e8eb56d32e6a77:

insane: Ensure do_package_qa happens after do_packagedata (2014-07-11 23:20:04 +0100)

are available in the git repository at:

git://git.yoctoproject.org/poky-contrib ross/automake

for you to fetch changes up to 406153008dc78f4e3198c012f2db98fcac09b5c1:

  autoconf: remove automake patch enforcing --foreign (2014-07-15 17:05:29 +0100)

----------------------------------------------------------------
Ross Burton (7):
bzip2: use "foreign" automake strictness
libid3tag: use "foreign" automake strictness
which: use "foreign" automake strictness
libmad: use "foreign" automake strictness
oprofile: use "foreign" automake strictness
ltp: use "foreign" automake strictness
autoconf: remove automake patch enforcing --foreign

 .../autoconf/autoconf/autoreconf-foreign.patch     |   13 -------------
 meta/recipes-devtools/autoconf/autoconf_2.69.bb    |    1 -
 .../bzip2/bzip2-1.0.6/configure.ac                 |    2 +-
 .../ltp/ltp/automake-foreign.patch                 |   20 ++++++++++++++++++++
 meta/recipes-extended/ltp/ltp_20140422.bb          |    1 +
 .../which/which-2.20/automake.patch                |   10 ++++++++++
 meta/recipes-extended/which/which_2.20.bb          |    1 +
 meta/recipes-kernel/oprofile/oprofile.inc          |    1 +
 .../oprofile/oprofile/automake-foreign.patch       |    7 +++++++
 .../libid3tag/obsolete_automake_macros.patch       |    3 ++-
 .../libmad/libmad/automake-foreign.patch           |   12 ++++++++++++
 meta/recipes-multimedia/libmad/libmad_0.15.1b.bb   |    1 +
 12 files changed, 56 insertions(+), 16 deletions(-)
 delete mode 100644 meta/recipes-devtools/autoconf/autoconf/autoreconf-foreign.patch
 create mode 100644 meta/recipes-extended/ltp/ltp/automake-foreign.patch
 create mode 100644 meta/recipes-extended/which/which-2.20/automake.patch
 create mode 100644 meta/recipes-kernel/oprofile/oprofile/automake-foreign.patch
 create mode 100644 meta/recipes-multimedia/libmad/libmad/automake-foreign.patch

Ross Burton (7):
bzip2: use "foreign" automake strictness
libid3tag: use "foreign" automake strictness
which: use "foreign" automake strictness
libmad: use "foreign" automake strictness
oprofile: use "foreign" automake strictness
ltp: use "foreign" automake strictness
autoconf: remove automake patch enforcing --foreign

 .../autoconf/autoconf/autoreconf-foreign.patch     |   13 -------------
 meta/recipes-devtools/autoconf/autoconf_2.69.bb    |    1 -
 .../bzip2/bzip2-1.0.6/configure.ac                 |    2 +-
 .../ltp/ltp/automake-foreign.patch                 |   20 ++++++++++++++++++++
 meta/recipes-extended/ltp/ltp_20140422.bb          |    1 +
 .../which/which-2.20/automake.patch                |   10 ++++++++++
 meta/recipes-extended/which/which_2.20.bb          |    1 +
 meta/recipes-kernel/oprofile/oprofile.inc          |    1 +
 .../oprofile/oprofile/automake-foreign.patch       |    7 +++++++
 .../libid3tag/obsolete_automake_macros.patch       |    3 ++-
 .../libmad/libmad/automake-foreign.patch           |   12 ++++++++++++
 meta/recipes-multimedia/libmad/libmad_0.15.1b.bb   |    1 +
 12 files changed, 56 insertions(+), 16 deletions(-)
 delete mode 100644 meta/recipes-devtools/autoconf/autoconf/autoreconf-foreign.patch
 create mode 100644 meta/recipes-extended/ltp/ltp/automake-foreign.patch
 create mode 100644 meta/recipes-extended/which/which-2.20/automake.patch
 create mode 100644 meta/recipes-kernel/oprofile/oprofile/automake-foreign.patch
 create mode 100644 meta/recipes-multimedia/libmad/libmad/automake-foreign.patch

1.7.10.4

bzip2 doesn't want GNU-levels of automake strictness so tell it to be "foreign".

---
 meta/recipes-extended/bzip2/bzip2-1.0.6/configure.ac |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 14b1d18..47ee576 100644
@@ -1,7 +1,7 @@
 AC_PREREQ([2.57])
 
 AC_INIT(bzip2, 2.0.5, , libXrender)
-AM_INIT_AUTOMAKE()
+AM_INIT_AUTOMAKE(foreign)
 AM_MAINTAINER_MODE
 
 #AM_CONFIG_HEADER(config.h)

1.7.10.4

libid3tag doesn't want GNU-levels of automake strictness so tell it to be "foreign".

---
 .../libid3tag/libid3tag/obsolete_automake_macros.patch              |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index 8a04b54..2845fb1 100644
@@ -6,7 +6,8 @@ diff -Nurd libid3tag-0.15.1b/configure.ac libid3tag-0.15.1b/configure.ac
 +++ libid3tag-0.15.1b/configure.ac	2013-01-03 06:41:02.734835014 +0200
 @@ -28,7 +28,7 @@
 
- AM_INIT_AUTOMAKE
+-AM_INIT_AUTOMAKE
++AM_INIT_AUTOMAKE([foreign])
 
 -AM_CONFIG_HEADER([config.h])
 +AC_CONFIG_HEADERS([config.h])

1.7.10.4

which doesn't want GNU-levels of automake strictness so tell it to be "foreign".

---
 meta/recipes-extended/which/which-2.20/automake.patch |   10 ++++++++++
 meta/recipes-extended/which/which_2.20.bb             |    1 +
 2 files changed, 11 insertions(+)
 create mode 100644 meta/recipes-extended/which/which-2.20/automake.patch

new file mode 100644
index 0000000..92365dd
@@ -0,0 +1,10 @@
+diff --git a/configure.ac b/configure.ac
+index d974461..a20dfa8 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -2,2 +2,3 @@ dnl Process this file with autoconf to produce a configure script.
+-AC_INIT(which.c)
+-AM_INIT_AUTOMAKE(which, 2.20)
++AC_INIT([which],[2.20])
++AC_CONFIG_SRCDIR(which.c)
++AM_INIT_AUTOMAKE([foreign])

index f3c6037..7483554 100644
@@ -18,6 +18,7 @@ PR = "r3"
 EXTRA_OECONF = "--disable-iberty"
 
 SRC_URI = "${GNU_MIRROR}/which/which-${PV}.tar.gz \
+           file://automake.patch \
            file://remove-declaration.patch"
 
 SRC_URI[md5sum] = "95be0501a466e515422cde4af46b2744"

1.7.10.4

libmad doesn't want GNU-levels of automake strictness so tell it to be "foreign".

---
 .../libmad/libmad/automake-foreign.patch                   |   12 ++++++++++++
 meta/recipes-multimedia/libmad/libmad_0.15.1b.bb           |    1 +
 2 files changed, 13 insertions(+)
 create mode 100644 meta/recipes-multimedia/libmad/libmad/automake-foreign.patch

new file mode 100644
index 0000000..3e54424
@@ -0,0 +1,12 @@
+Pass foreign to AM_INIT_AUTOMAKE so it doesn't enforce GNU strictness.
+
+Upstream-Status: Pending
+Signed-off-by: Ross Burton <[email protected]>
+
+diff --git a/configure.ac b/configure.ac
+index e602fd3..e075b86 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -29 +29 @@ AC_CONFIG_SRCDIR([decoder.h])
+-AM_INIT_AUTOMAKE
++AM_INIT_AUTOMAKE([foreign])

index 58d68a6..30937eb 100644
@@ -15,6 +15,7 @@ SRC_URI = "${SOURCEFORGE_MIRROR}/mad/libmad-${PV}.tar.gz \
            file://add-pkgconfig.patch \
            file://fix_for_mips_with_gcc-4.5.0.patch \
            file://obsolete_automake_macros.patch \
+           file://automake-foreign.patch \
 "
 
 SRC_URI[md5sum] = "1be543bc30c56fb6bea1d7bf6a64e66c"

1.7.10.4

oprofile doesn't want GNU-levels of automake strictness so tell it to be "foreign".

---
 meta/recipes-kernel/oprofile/oprofile.inc                    |    1 +
 meta/recipes-kernel/oprofile/oprofile/automake-foreign.patch |    7 +++++++
 2 files changed, 8 insertions(+)
 create mode 100644 meta/recipes-kernel/oprofile/oprofile/automake-foreign.patch

index 76d0b6c..6958203 100644
@@ -20,6 +20,7 @@ FILES_${PN}-staticdev += "${libdir}/${BPN}/lib*.a"
 
 SRC_URI = "file://opstart.patch \
            file://acinclude.m4 \
+           file://automake-foreign.patch \
            file://oprofile-cross-compile-tests.patch \
            file://run-ptest \
            file://root-home-dir.patch"

new file mode 100644
index 0000000..e82a381
@@ -0,0 +1,7 @@
+diff --git a/configure.ac b/configure.ac
+index 5740585..cf6c316 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -16 +16 @@ AC_CONFIG_SRCDIR([libop/op_config.h])
+-AM_INIT_AUTOMAKE
++AM_INIT_AUTOMAKE([foreign])

1.7.10.4

ltp doesn't want GNU-levels of automake strictness so tell it to be "foreign".

---
 .../ltp/ltp/automake-foreign.patch                 |   20 ++++++++++++++++++++
 meta/recipes-extended/ltp/ltp_20140422.bb          |    1 +
 2 files changed, 21 insertions(+)
 create mode 100644 meta/recipes-extended/ltp/ltp/automake-foreign.patch

new file mode 100644
index 0000000..c3dd891
@@ -0,0 +1,20 @@
+Use foreign strictness to avoid automake errors.
+
+Upstream-Status: Submitted (https://github.com/linux-test-project/ltp/issues/16)
+Signed-off-by: Ross Burton <[email protected]>
+
+diff --git a/configure.ac b/configure.ac
+index 9f397e7..fc57957 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -4 +4 @@
+-AM_INIT_AUTOMAKE
++AM_INIT_AUTOMAKE([foreign])
+--- a/utils/ffsb-6.0-rc2/configure.in
++++ b/utils/ffsb-6.0-rc2/configure.in
+@@ -2,2 +2,3 @@ dnl Process this file with autoconf to produce a configure script.
+-AC_INIT(main.c)
+-AM_INIT_AUTOMAKE(ffsb, 6.0-RC2)
++AC_INIT([ffsb], [6.0-RC2])
++AC_CONFIG_SRCDIR([main.c])
++AM_INIT_AUTOMAKE([foreign])

index 995d617..30761fd 100644
@@ -25,6 +25,7 @@ SRCREV = "f4c3bfe1eab51eb72caeb0f3336d2790c9a8bd1b"
 SRC_URI = "git://github.com/linux-test-project/ltp.git \
     file://0001-Rename-runtests_noltp.sh-script-so-have-unique-name.patch \
     file://ltp-Do-not-link-against-libfl.patch \
+    file://automake-foreign.patch \
 "
 
 S = "${WORKDIR}/git"

1.7.10.4

Since 2004 we've been carrying a patch[1] make autoreconf pass --foreign to automake. Presumably at the time this was due to many upstreams using hand-coded bootstrap scripts that passed --foreign manually, but we were using autoreconf. These days many projects have added foreign to AM_INIT_AUTOMAKE and use autoreconf directly, so this patch isn't as critical as it used to be.

[1] oe-classic 2ab2a92eadaf2f80410d8746099f8a9b1b81ff91

---
 .../autoconf/autoconf/autoreconf-foreign.patch            |   13 -------------
 meta/recipes-devtools/autoconf/autoconf_2.69.bb           |    1 -
 2 files changed, 14 deletions(-)
 delete mode 100644 meta/recipes-devtools/autoconf/autoconf/autoreconf-foreign.patch

deleted file mode 100644
index 88f7501..0000000
@@ -1,13 +0,0 @@
-Upstream-Status: Pending
-
---- autoconf-2.59/bin/autoreconf.in~autoreconf-foreign	2004-05-09 20:55:06.000000000 -0400
-+++ autoconf-2.59/bin/autoreconf.in	2004-05-09 20:55:55.000000000 -0400
-@@ -184,6 +184,8 @@
- 
-   $aclocal_supports_force = `$aclocal --help` =~ /--force/;
- 
-+  $automake   .= ' --foreign';
-+
-   # Dispatch autoreconf's option to the tools.
-   # --include;
-   $autoconf   .= join (' --include=', '', @include);

index 5c652d4..6614a7b 100644
@@ -10,7 +10,6 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=751419260aa954499f7abaabaa882bbe \
 SRC_URI += "file://autoreconf-include.patch \
 	    file://check-automake-cross-warning.patch \
 	    file://autoreconf-exclude.patch \
-	    file://autoreconf-foreign.patch \
 	    file://autoreconf-gnuconfigize.patch \
             file://config_site.patch \
             file://remove-usr-local-lib-from-m4.patch \

1.7.10.4

to

using

AM_INIT_AUTOMAKE and

can it be somehow made such that it can be passed via option flags so we dont have to patch the packages not using above automake macro

-------------

meta/recipes-devtools/autoconf/autoconf/autoreconf-foreign.patch

a/meta/recipes-devtools/autoconf/autoconf/autoreconf-foreign.patch b/meta/recipes-devtools/autoconf/autoconf/autoreconf-foreign.patch

Packages that don't use the right automake macro are generally very old and we're already carrying patches to fix the autoconf/automake files, so I'm not sure this is actually worth it. I'm trying to drop the patches we're carrying against core build tools, not increase it. :)

Ross

While I understand that, I am also worried about breakage/work it will cause for layers additional to OE-Core
thats why providing a least impact migration path is desirable.

world builds from
http://www.openembedded.org/wiki/Bitbake_World_Status
already have around ~50 failed recipes per architecture, I don't think we need more now :/

Do we have some volunteers to fix-up meta-oe/* layers?

I'm going to use PNBLACKLIST more aggressively as agreed on OEDAM, but with this speed next meta-oe release will have half of recipes which were available in daisy.

Martin 'JaMa' Jansa     jabber: [email protected]

I made a quick test and the failures I spotted from meta-oe likely due to this patch were:

re2c
libsdl-iomage
claws-plugin-gtkhtml2-viewer
claws-plugin-rssyl
libsdl-tiff
libol
ode
dejagnu
memcached
rp-pppoe
ipsec-tools
mtr
hplip
onig

Its probably not a complete list as I have a lot more failures than your world builds but it gives some idea of number and its not a massive list.

Cheers,

Richard

There were more efl ones once I sorted a fetch issue.

I've run out of time/energy today but I have about 20 fixes for various build issues in meta-oe in:

http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/metaoe&id=bfb0ddc8c3d3f7c963c3529efc970b9ba8d8d788

Mentioning so people don't duplicate work. I'll try and split into patches as/when I get time if nobody beats me to it.

Cheers,

Richard