2011-08-14scripts/functions: add xz support
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 17:51:38 +0200] rev 2609
scripts/functions: add xz support

Add support for the new XZ-compressed tarballs. At least glibc
uses that new format.

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

2011-05-04scripts/functions: test for decompressors before use
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Wed, 04 May 2011 00:04:23 +0200] rev 2608
scripts/functions: test for decompressors before use

./configure does check for the presence of gz and bzip2, so we can
safely use them in the build scripts.

On the other hand, more recent formats (eg. XZ) are not yet widely
available, and we do not want, and can't, force the user to install
them as a pre-requisite.

So, build up a list of allowed tarball formats based on the available
decompressors. For no, this is a static list, but the upcoming XZ
support will conditionnaly add to this list.

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

2011-08-10configure: add test for xz-utils
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Wed, 10 Aug 2011 23:13:46 +0200] rev 2607
configure: add test for xz-utils

Some upstream now release tarballs compressed with xz, along with the
more traditional bzip2 and/or gzip tarballs. xz (and its predecessor
lzma) achieves better compression, so is a good candidate to shorten
the time it takes to download the source tarballs.

Enable downloading xz-cmopressed tarballs if the user has xz-utils.

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

2011-08-14configure: do not require lzma
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 16:06:20 +0200] rev 2606
configure: do not require lzma

Nothing uses lzma now, so we can get rid of it.

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

2011-08-14scripts/functions: commonalise tar options in CT_EXtract
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 17:28:35 +0200] rev 2605
scripts/functions: commonalise tar options in CT_EXtract

This shortens command lines, so it's good! :-)

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

2011-08-14kernel/linux: remove check for lzma tarballs
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 16:07:03 +0200] rev 2604
kernel/linux: remove check for lzma tarballs

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

2011-08-14kernel/linux: simplify custom tarball handling
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 19:59:02 +0200] rev 2603
kernel/linux: simplify custom tarball handling

Simplify the way the custom tarball is handled:
- fake version="custom"
- at download, simply link the custom tarball to:
"linux-custom.${custom_extension}"
- at extract, the above allows to simply extract "linux-${LINUX_VERSION}"
where LINUX_VERISON is set to the fake version="custom"

Not that much convoluted, in fact... :-/

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

2011-08-14kernel/linux: deprecate copying headers from pre-instaleld headers
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 16:05:33 +0200] rev 2602
kernel/linux: deprecate copying headers from pre-instaleld headers

No-one ever inquired about this feature. Due to the way it works, it is
rather unlikely that every user used a properly-formatted tarball, or a
properly structured directory as imput.

Beside, it is not really easy to maintain: I just spent one full day
remembering how it worked... Sigh... :-(

Just deprecate this, until next release, or until someone complains.

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

2011-08-14samples: do not use the mirror by default
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sun, 14 Aug 2011 17:18:32 +0200] rev 2601
samples: do not use the mirror by default

As the mirror is always tried first, and the default is my server,
it means every download will first hit my server. As I do not have
much mirrored there, it means a lot of time is lost.

Disable use of the mirror for all samples, except the one that does
need it for the uClibc snapshot.

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

2011-08-13libc/uClibc: force use of ctor/dtor
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> [Sat, 13 Aug 2011 00:05:51 +0200] rev 2600
libc/uClibc: force use of ctor/dtor

Although ctor/dtor do not seem strictly required, missing them proves
rather inconvenient, as ld can't link binaries.

Reported-by: John Spencer <maillist-uclibc@barfooze.de> (sh4rm4 on IRC)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>