BlueZ upgrades

The following changes since commit 0aaa67a790ab211b35eed58173efa115ceb6e243:

oeqa/bbtests: Fix to ensure DL_DIR is set (2015-07-08 13:09:28 +0100)

are available in the git repository at:

git://git.yoctoproject.org/poky-contrib ciorga/PU2
http://git.yoctoproject.org/cgit.cgi/poky-contrib/log/?h=ciorga/PU2

Cristian Iorga (2):
bluez5: upgrade to 5.32
libical: upgrade to 1.0.1

 .../bluez5/{bluez5_5.31.bb => bluez5_5.32.bb}                    | 4 ++--
 .../libical/{libical_1.0.0.bb => libical_1.0.1.bb}               | 9 ++++++---
 2 files changed, 8 insertions(+), 5 deletions(-)
 rename meta/recipes-connectivity/bluez5/{bluez5_5.31.bb => bluez5_5.32.bb} (88%)
 rename meta/recipes-support/libical/{libical_1.0.0.bb => libical_1.0.1.bb} (67%)

2.1.4

Mostly a bug fix release.

Fixes in several different areas, including OPP, ATT and advertising (instance number handling in particular). Fix for handling a sudden disconnect when a connection setup process hasn’t yet completed.
New feature: ability to select between letting the stack handle ATT security elevation or doing the respective error handling in higher layers.

---
 meta/recipes-connectivity/bluez5/{bluez5_5.31.bb => bluez5_5.32.bb} | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
 rename meta/recipes-connectivity/bluez5/{bluez5_5.31.bb => bluez5_5.32.bb} (88%)

similarity index 88%
rename from meta/recipes-connectivity/bluez5/bluez5_5.31.bb
rename to meta/recipes-connectivity/bluez5/bluez5_5.32.bb
index 395c512..db02be4 100644
@@ -1,6 +1,6 @@
 require bluez5.inc
-SRC_URI[md5sum] = "cb1db487b65dd4d57a2dd6b8f063c393"
-SRC_URI[sha256sum] = "c4b661cc5866d9a9a9d608d700392afa278c3b6fe8426238bf6979be7c63818f"
+SRC_URI[md5sum] = "432b1f292c0b78771818c67140baf3d9"
+SRC_URI[sha256sum] = "fe4f35e701cbec04478818adf014ed9bbd1f7d11e2603370f1b70a78106f57f7"
 
 # noinst programs in Makefile.tools that are conditional on READLINE
 # support

2.1.4

Bugfix release mainly.
autotools build system removed, cmake is now used. perl is needed during the configure phase.

---
 .../libical/{libical_1.0.0.bb => libical_1.0.1.bb}               | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
 rename meta/recipes-support/libical/{libical_1.0.0.bb => libical_1.0.1.bb} (67%)

similarity index 67%
rename from meta/recipes-support/libical/libical_1.0.0.bb
rename to meta/recipes-support/libical/libical_1.0.1.bb
index be6b635..b50cf7e 100644
@@ -7,7 +7,10 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=d4fc58309d8ed46587ac63bb449d82f8 \
 SECTION = "libs"
 
 SRC_URI = "https://github.com/${BPN}/${BPN}/archive/v${PV}.tar.gz"
-SRC_URI[md5sum] = "f4b8e33ae5efb2f025eb43ce69682a36"
-SRC_URI[sha256sum] = "0072e83834092315772e6719b85fc8b11530b1ff53f4d108315fb38cddbce8c2"
+SRC_URI[md5sum] = "af91db06b22559f863869c5a382ad08a"
+SRC_URI[sha256sum] = "7d5f613454ec6c7d1bcfb441c919215be53292aa15cd1cb14249d1413d6c610c"
+
+inherit cmake perlnative
+
+FILES_${PN}-dev += "${libdir}/cmake"
 
-inherit autotools

2.1.4

Yes. I keep on meaning to write a QA test for this. :)

Cristi - if this is the best libical has as a tarball release then we should do a proper git clone. Amazingly, the download button on their website links to a tarball of master which is borderline reckless.

Ross

Hi all,

Hahaha. I am laughing because I assumed that the word “archive” means the file will stay there forever with the same checksum. Silly me. ☺ They have another link, but I am not sure if that stays up and that the checksum does not change.
Is this always true, that the checksum changes?

Thanks,
Cristian

From: Burton, Ross [mailto:[email protected]]
Sent: Thursday, July 9, 2015 6:10 PM
To: Martin Jansa
Cc: Iorga, Cristian; OE-core
Subject: Re: [OE-core] [PATCH 2/2] libical: upgrade to 1.0.1

On 9 July 2015 at 15:32, Martin Jansa <[email protected]<mailto:[email protected]>> wrote:
aren't github archives regenerated from time to time with different
checksums?

Yes. I keep on meaning to write a QA test for this. :)

Cristi - if this is the best libical has as a tarball release then we should do a proper git clone. Amazingly, the download button on their website links to a tarball of master which is borderline reckless.

Ross

Github automated tarballs are apparently dynamically generated and cached. They also offer hosting for release tarballs, but not many projects use those.

Ross

Would this one qualify as a release tarball download link?

https://github.com/libical/libical/releases/download/v1.0.1/libical-1.0.1.tar.gz

Thanks,
Cristian

From: Burton, Ross [mailto:[email protected]]
Sent: Thursday, July 9, 2015 7:05 PM
To: Iorga, Cristian
Cc: Martin Jansa; OE-core
Subject: Re: [OE-core] [PATCH 2/2] libical: upgrade to 1.0.1

On 9 July 2015 at 16:57, Iorga, Cristian <[email protected]<mailto:[email protected]>> wrote:
Hahaha. I am laughing because I assumed that the word “archive” means the file will stay there forever with the same checksum. Silly me. ☺ They have another link, but I am not sure if that stays up and that the checksum does not change.
Is this always true, that the checksum changes?

Github automated tarballs are apparently dynamically generated and cached. They also offer hosting for release tarballs, but not many projects use those.

Ross

Yes, wonders never cease. :)

Ross

FWIW: someone in systemd herd said that now the archives on github should be stable as long as git-archive is stable.
https://www.mail-archive.com/[email protected]/msg33014.html
https://www.mail-archive.com/[email protected]/msg33015.html
https://www.mail-archive.com/[email protected]/msg33017.html
https://www.mail-archive.com/[email protected]/msg33016.html

Now they are using them instead of proper tarballs:
https://www.mail-archive.com/[email protected]/msg33054.html

Martin 'JaMa' Jansa     jabber: [email protected]