Makefile: make it work for auto-completion

The latest autocompletion calls 'make -qp' to get all the possible
targets. This currently fails, as our makefile is calling itself
again and again ad-libitum.

Fix it by detecting that the recursion level is not 0, and bail out
if so. It works so well that it has the side effect of showing only
the 'public' target, and hide our internal ones! :-)

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

libc/glibc: do not try to download NPTL add-on

The NPTL add-on has always been internal, so there is no
reason to try downloading it, it will never succeed.
Add provision to skip other add-ons as well.

For consistency, do the same test in both glibc and eglibc.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: prepend CT-NG's version tag to PKGVERSION

"crosstool-NG-${CT_VERSION}" is currently the default for TOOLCHAIN_PKGVERSION,
and this options is passed as is to --with-pkgversion.

This patch prepends "crosstool-NG ${CT_VERSION}" to TOOLCHAIN_PKGVERSION before
passing it to --with-pkgversion.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

functions: save & restore the static companion libraries install dir

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

functions: fix downloading files

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

cc/gcc: CC_STATIC_LIBSTDCXX 'depends on' CONFIGURE_has_static_libstdcpp

Hide the staticaly linked libstdc++ option if the static libstdc++ is not
present, detected at configure time.

Add a blind option that says whether static linking is possible at all.
It defaults to 'y', but depends on the needed CONFIGURE_* options. For
now, it only depends on static libtdc++, but new dependencies can be
easily added.

Hide the global static toolchain option behind this new option.

Original patch by Bryan Hundven <bryanhundven@gmail.com>

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

configure: check for libstdc++.{so,dylib,a}

It's probably a good thing to check for at least one instance of
libstdc++.{so,dylib,a} at configure time. At least one is needed
by the companion libraries and gcc.

Original patch by Bryan Hundven <bryanhundven@gmail.com>

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

configure: add prefix to err messages in has_or_{abort,warn}

Make it easier to spot messages spewed out by has_or_{abort,warn}.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

configure: maintain single quotes in error messages

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

glibc: properly handle internal addons

Some addons are bundled with glibc/eglibc, so we should not try to
download and extract them.

This is done as thus:
- at download time:
- if the add-on download fails, keep going;
- at extract time:
- if the addon is present in the source tree, ignore it;
- if the addon is missing in the source tree:
- if the archive is present, extract it;
- if the archive is missing, bail out.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

libc/eglibc: use generic SVN functions

eglibc is only available from SVN. The script currently calls svn
in its own tortuous and convoluted way.

Use the egeneric SVN extract functions, and sinplify the eglibc
download function.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

scripts/functions: do not abort on failed svn

In case of eglibc, some add-ons that were previously external are
now internal (bundled with the main sources).

So we do not want to fail if an add-on can't be downloaded; we
want to post-pone the check until we can extract the main archive.

So:
- try to retrieve the add-on
- if it fails, print a warning instead of calling CT_Abort
- return 1

So, components that want to catch the error and want to handle it can,
while components that do not will gracefuly fail thanks to our catching
every errors.

Bonus: it works without changing any existing retrieval procedure! :-)

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

scripts/functions: do not abort on failed extract

For glibc/eglibc, if the add-on can not be extracted, we want to catch
the error to print a meaningful error message.

So:
- try to extract the tarball
- if it fails, print a waring instead of calling CT_Abort
- return 1

So, components that want to catch the error and want to handle it can,
while components that do not will gracefuly fail thanks to our catching
every errors.

Bonus: it works without changing any existing extract procedure! :-)

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

scripts/functions: do not abort on failed download

In case of glibc/eglibc, some add-ons that were previously external are
now internal (bundled with the main sources).

So we do not want to fail if an add-on tarball can't be downloaded; we
want to post-pone the check until we can extract the main archive.

So:
- try to download the tarball
- if it fails, print a warning instead of calling CT_Abort
- return 1

So, components that want to catch the error and want to handle it can,
while components that do not will gracefuly fail thanks to our catching
every errors.

Bonus: it works without changing any existing retrieval procedure! :-)

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

cc/gcc: fix a misleading FIXME

The FIXME about the static libstdc++ is misleading; it only deserves
being an INFO.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

gdb: use the PKGVERSION and BUGURL options

This patch makes gdb benefit from the TOOLCHAIN_PKGVERSION and
TOOLCHAIN_BUGURL options.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

eglibc: use the PKGVERSION and BUGURL options

This patch makes eglibc benefit from the TOOLCHAIN_PKGVERSION and
TOOLCHAIN_BUGURL options.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

binutils: use the PKGVERSION and BUGURL options

This patch makes binutils benefit from the TOOLCHAIN_PKGVERSION and
TOOLCHAIN_BUGURL options.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

gcc: promote PKGVERSION and BUGURL options to toolchain level

This patch promotes the PKGVERSION and BUGURL options to toolchain level so that
all toolchain components supporting them can benefit from them.

These options are passed to configure through --with-pkgversion and
--with-bugurl.

They are supported by binutils 2.18+, gcc 4.3+, eglibc 2.9+ and gdb 7.0+.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

gcc: add a CC_GCC_HAS_PKGVERSION_BUGURL option

This patch adds a blind option CC_GCC_HAS_PKGVERSION_BUGURL to test the support
of --with-pkgversion and --with-bugurl by GCC's configure.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

scripts/addToolVersion.sh: add a function to extract fields from versions

This patch adds a function to extract major, minor, revision, etc. from version
numbers.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

kconfig: remove useless 'default n'

kconfig bools are disabled by default, so specifying 'default n' is useless and
noisy. This patch removes all occurrences of 'default n'.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

libc/glibc: commonalise extraction btw glibc and eglibc

glibc and eglibc have a very similar extraction process, so it
makes sense to commonalise it.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

libc/glibc: fix {e,}glibc add-ons with short or long names

Fixes the issue with {e,}glibc addons having short and long names (such as
eglibc-ports-2_13 and ports), which caused configure scripts to run
through them twice and thus configuring incorrectly.

For instance, the mips64el-n32-linux-gnu toolchain would be recognized
correctly first, but then the second pass would change it to mips32,
building a mixed MIPS-III N32 and MIPS-I libc.

Signed-off-by: Nicolás Reynolds <fauno@kiwwwi.com.ar>
[yann.morin.1992@anciens.enib.fr: remove spurious trailing spaces]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

configure: add possibility to set arbitrary variable in check_for

If check_for is able to find the required prog/inc/lib, allow it to
set an arbitrary variable to 'y'. This variable is then pushed down
to the kconfig definition.

For example:
has_or_abort prog=foobar kconfig=has_foobar

If foobar is available, it yields a kconfig variable defaulting to y:
config CONFIGURE_has_foobar
bool
default y

If foobar is missing, it yields a kconfig variable defaulting to n:
config CONFIGURE_has_foobar
bool

Thus it is possible to depends on that variabel to show/hide options:
config SOME_FEATURE
bool
prompt "Some feature"
depends on CONFIGURE_has_foobar

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

configure: add has_or_warn

By popular request, add a function to only warn on missing stuff.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

configure: move error message down to has_or_abort

When has_or_warn will come, we do want to print the error message
only in has_or_abort, so move it down there.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

debug/gdb: bump Linaro revision to 2011.05-0

This patch bumps the Linaro GDB revision to 2011.05-0.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

cc/gcc: bump Linaro revisions to 2011.05-0

This patch bumps the Linaro GCC revisions to 2011.05-0 when applicable.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

debug/gdb: fix runtime sysroot

This patch sets the runtime sysroot to fix the following GDB warning:
"Unable to find dynamic linker breakpoint function.
GDB will be unable to debug shared library initializers
and track explicitly loaded dynamic code."

The sysroot can later be changed within gdb with the `set sysroot`
command if necessary.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

scripts: fix broken variable name

This patch fixes a config variable name missing its 'CT_' prefix.

Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

kconfig: fix forward dependencies

In case a set of options are conditional to two different paths,
the forward dependencies in these options is wrong, but in the
first path.

Fix by Arnaud Lacombe on linux-kbuild ML:
http://www.spinics.net/lists/linux-kbuild/msg04832.html

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kernel/linux: add latest versions

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

1.11: update version to 1.11.3+hg

Tagging release 1.11.3

1.11: update version to 1.11.3

config: document array-capable variables

Update help entries for thos variables that accept multiple
arguments with spaces (aka. array-capable variables).

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

scripts: munge .config to handle array variables

Transfrom array variables from pure strings (in the kconfig output)
into proper bash arrays declarations, for feeding to the build scripts.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
Signed-off-by: "Benoît THÉBAUDEAU" <benoit.thebaudeau@advansee.com>

config: rename variables that are arrays

Make it explicit that a variable is an array bu the name of the variable.
It will be used later when .config gets munged to allow both multiple
arguments and arguments with spaces at the same time to be passed from the
configuration down to the build scripts.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

scripts: interpret *_EXTRA_CONFIG config variables arrays

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

complibs/ppl: fix 0.11-0.11.2 to compile with --disable-shared

PPL 0.11 (through 0.11.2) had a small bug where it still tried to build
and test its Java interface even when shared libraries are disabled.
Since that's exactly what ct-ng does, it explodes.

This is the patch from the PPL authors (see final link below).

More information can be found in these messages/threads:

Anthony's initial report and analysis with Yann:
http://www.cygwin.com/ml/crossgcc/2011-05/msg00046.html

Ron Flory hit the same problem:
http://www.cygwin.com/ml/crossgcc/2011-05/msg00054.html

Anthony's report to the ppl-devel list:
http://www.cs.unipr.it/pipermail/ppl-devel/2011-May/017450.html

Roberto's reply with a link to the fix in the PPL git repo:
http://www.cs.unipr.it/pipermail/ppl-devel/2011-May/017455.html

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>
(transplanted from 139b85d70b623e51f7585e7c64ecb8c6a6cf4567)

complibs/ppl: build only C and C++ interfaces for PPL

By default, PPL wants to build interfaces for any of a variety of
langauges it finds on the local host (python, java, possibly perl, also
more esoteric languages such as ocaml and prolog).

These extra interfaces can double the compile time for the library. For
single-process builds, I found a savings of more than 40%:

default / j1: 716s total, 143.2s avg, 0.52s stdev
just_c / j1: 406s total, 81.2s avg, 0.33s stdev
just_c_cpp / j1: 413s total, 82.6s avg, 0.22s stdev

And for multi-process builds, it approached 50%:

default / j4: 625s total, 125.0s avg, 0.57s stdev
just_c / j4: 338s total, 67.6s avg, 1.25s stdev
just_c_cpp / j4: 327s total, 65.4s avg, 0.36s stdev

Since the PPL we build within ct-ng is only used by GCC, we only need to
build the C and C++ interfaces.

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>
(transplanted from ec30b191f0e3fe9bc73199f0bcb8d789db17aa92)

complibs/ppl: update GMP location configuration argument for PPL v0.11 and later

'configure' for PPL 0.11 (and later) needs "--with-gmp-prefix" to
provide the location of the GMP toolkit; the previous switches were
"--with-libgmp-prefix" and "--with-libgmpxx-prefix".

The upstream log message is:

commit 08dfb6fea094f8c5a533575a3ea2095edce99a6d
Author: Roberto Bagnara <bagnara@cs.unipr.it>
Date: Sun Jul 12 21:39:46 2009 +0200

New configure option --with-gmp-prefix supersedes the (now removed)
options --with-libgmp-prefix and --with-libgmpxx-prefix.

Link: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=08dfb6fea094f8c5a533575a3ea2095edce99a6d

Since PPL's 'configure' ignores unknown switches, we use all three so we
don't have to conditionalize the ppl.sh build script itself.

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>
(transplanted from 4f0c4fb572e2862c24b28e8d27ce7e9cb9adba65)

complibs/ppl: fix 0.11-0.11.2 to compile with --disable-shared

PPL 0.11 (through 0.11.2) had a small bug where it still tried to build
and test its Java interface even when shared libraries are disabled.
Since that's exactly what ct-ng does, it explodes.

This is the patch from the PPL authors (see final link below).

More information can be found in these messages/threads:

Anthony's initial report and analysis with Yann:
http://www.cygwin.com/ml/crossgcc/2011-05/msg00046.html

Ron Flory hit the same problem:
http://www.cygwin.com/ml/crossgcc/2011-05/msg00054.html

Anthony's report to the ppl-devel list:
http://www.cs.unipr.it/pipermail/ppl-devel/2011-May/017450.html

Roberto's reply with a link to the fix in the PPL git repo:
http://www.cs.unipr.it/pipermail/ppl-devel/2011-May/017455.html

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>

complibs/ppl: build only C and C++ interfaces for PPL

By default, PPL wants to build interfaces for any of a variety of
langauges it finds on the local host (python, java, possibly perl, also
more esoteric languages such as ocaml and prolog).

These extra interfaces can double the compile time for the library. For
single-process builds, I found a savings of more than 40%:

default / j1: 716s total, 143.2s avg, 0.52s stdev
just_c / j1: 406s total, 81.2s avg, 0.33s stdev
just_c_cpp / j1: 413s total, 82.6s avg, 0.22s stdev

And for multi-process builds, it approached 50%:

default / j4: 625s total, 125.0s avg, 0.57s stdev
just_c / j4: 338s total, 67.6s avg, 1.25s stdev
just_c_cpp / j4: 327s total, 65.4s avg, 0.36s stdev

Since the PPL we build within ct-ng is only used by GCC, we only need to
build the C and C++ interfaces.

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>

complibs/ppl: remove unnecessary (and confusing) --enable-cxx configure switch.

PPL does not use the "--enable-cxx" configure switch at all; it's
possibly a cut-and-paste leftover from 'gmp.sh'. (PPL is written in C++
natively, so it doesn't make much sense to have to enable C++; GMP, on
the other hand, is written in C with an optional C++ wrapper.)

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>

complibs/ppl: update GMP location configuration argument for PPL v0.11 and later

'configure' for PPL 0.11 (and later) needs "--with-gmp-prefix" to
provide the location of the GMP toolkit; the previous switches were
"--with-libgmp-prefix" and "--with-libgmpxx-prefix".

The upstream log message is:

commit 08dfb6fea094f8c5a533575a3ea2095edce99a6d
Author: Roberto Bagnara <bagnara@cs.unipr.it>
Date: Sun Jul 12 21:39:46 2009 +0200

New configure option --with-gmp-prefix supersedes the (now removed)
options --with-libgmp-prefix and --with-libgmpxx-prefix.

Link: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=08dfb6fea094f8c5a533575a3ea2095edce99a6d

Since PPL's 'configure' ignores unknown switches, we use all three so we
don't have to conditionalize the ppl.sh build script itself.

Signed-Off-By: Anthony Foiani <anthony.foiani@gmail.com>

1.11: update version to 1.11.2+hg

Tagging release 1.11.2

1.11: update version to 1.11.2

libc/glibc: source common parts instead of symlinking

Symlinks are a nightmare when one is grepping...

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: add the nconf frontend

The nconf frontend is the new hot topic in the kconfig land! :-)

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

libc/glibc: use the multi-inclusion capability of the new kconfig

Now that kconfig can include a file multiple times, link the glibc/eglibc
common stuff to two .in.2 files.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: remove mis-leading messages

Unlinke the Linux kernel, crosstool-NG does not (directly) use make
as a frontend, so remove any occurence of reference to running make
after the configuration.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: fix choice multi-display

In case a choices has different combinations, it can be displayed
more than once.

See the discussion thread on the linux-kbuild ML:
http://www.spinics.net/lists/linux-kbuild/msg04709.html

The fix applied here was proposed by Arnaud LACOMBE.
That fix, or a similar one, will probably be pushed upstream soon.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: don't trim spaces with leading pipe

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: do not warn on missing env variable

In crosstool-NG, we use env variables for the backend mode.
So it is perfectly legit that these variables are not set.
So do not warn about it.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kconfig: update from linux-next

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

config: add main menu title

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

kernel/linux: reorder stuff

Yet another reordering...

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>