| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
| |
it conflicts with mkisofs from cdrkit
|
|
|
|
|
|
| |
fixes #6127
(cherry picked from commit 69bc162319b12e9b6c6d3ea345dbf7c218753594)
|
|
|
|
|
|
| |
fixes #6116
(cherry picked from commit ecfc04f3961ec4ffa2c972bd72253ba1a03a3c1e)
|
|
|
|
|
|
|
|
| |
This also updates /etc/resolv.conf atomically to avoid transient
failures during updates.
Signed-off-by: David Scott <dave.scott@docker.com>
(cherry picked from commit 9223390957114a01f06a7d402f52462917ed34aa)
|
|
|
|
|
|
|
| |
the make will return success even if it fails. so we also check if brctl
was successfully built.
(cherry picked from commit cec59c9a31b98de61666f03481fd4862e44d023a)
|
|
|
|
| |
apparently the package is empty. we trigger rebuild
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since commit 71eb72d62425082850604f526dbcbfdcf2808c31 (2016-03-13,
pre-v3.4) openssh is build with pid dir explicitly set to /run.
The change was not reflected in sshd.confd or sshd.initd, though,
and sadly not even in the commit message.
(Before it was set implicitly to /var/run.)
/var/run and /run semantics are the same, but AL does not truly
guarantee (at least yet) that the first is symlinked to the latter
(which is a common practice among Linux distributions nowadays, where
/run is tmpfs mounted very early - in AL openrc's init.sh does that).
alpine-baselayout package simply has run and var/run directories and
they are not related in any way from the package point of view.
Unless you create such symlink yourself or it is created via openrc's
boot service bootmisc (performing /var/run -> /run migration and some
other stuff), you cannot use /var/run/ and /run/ paths interchangeably.
The patch should be applied to 3.4-stable branch too (without changing
pkgver used there and with proper pkgrel increment, of course).
I was seeing false crashed state next to sshd in rc-status after
upgrading AL from 3.3 to 3.4 on machine where bootmisc is not used.
(I don't think it's a grave enough lack to warrant patch rejection.)
(cherry picked from commit 528334a1c482176781261a8a8a1c7a5aec582c1b)
|
| |
|
|
|
|
| |
CVE-2016-6316
|
|
|
|
| |
fixes #6045
|
|
|
|
| |
fixes #6040
|
| |
|
| |
|
|
|
|
|
|
| |
Fixes #4488
(cherry picked from commit 4eb826279c131257c3bc37d42f394208022ef07e)
|
|
|
|
|
|
| |
fixes #5960
(cherry picked from commit dc7fda197a3664b715eb74c816fea8496d05db6e)
|
|
|
|
|
|
|
| |
the directory is so packages can ship config. User may want override
those in the config file so we let the config file win.
This is also consistent with sysctl.
|
|
|
|
| |
(cherry picked from commit 02ddd15e9c42044d8fecd41aba173d8f95a989e4)
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
synchronize with linux-grsec config
(cherry picked from commit 0c74584afabbe2383306124e1c318a4fd5420327)
|
|
|
|
| |
(cherry picked from commit 3d378c2353116571eb24c51d019521e0e295e73e)
|
|
|
|
| |
(cherry picked from commit 62e75f7dec8a8115f4c68c9d5c65bfcb89059260)
|
|
|
|
| |
(cherry picked from commit a6a0d8600cb0f4a54215e3c7e4b78c7efb76fc41)
|
| |
|
| |
|
|
|
|
|
|
|
| |
subversion needs to convert to and from utf8 and it will fail to do so
if we disable iconv.
(cherry picked from commit 29f770bc53e9ff5a6c405f9b4ec36df0ca87516c)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fixes #4802
Seems like iscsiadm relies on checking optopt for checking if there's an
error, instead of checking whether opterr is set or if getopt returns
'?' and then using optopt to retrieve the option that it didn't
recognize.
For some reason musl sets optopt to the last option parsed even if it
was valid. Is musl supposed to do that? glibc doesn't behave this way.
The posix standard for getopt also reads like optopt is only supposed to
be set when there's an error.
|
|
|
|
| |
fixes #6017
|
|
|
|
| |
fixes #5995
|
|
|
|
| |
fixes #6024
|
|
|
|
|
|
| |
so we can edit sysfs variables during boot
(cherry picked from commit 4c9c24b11fa49edd270288882fbb81380c5dd3ac)
|
|
|
|
| |
(cherry picked from commit ac94d4b9a3edac9db7aa1481b4866cb39d032843)
|
|
|
|
|
|
|
|
|
| |
CVE-2015-8934: out of bounds heap read in RAR parser
CVE-2016-4300: Heap buffer overflow vulnerability in the 7zip read_SubStreamsInfo
CVE-2016-4302: Heap buffer overflow in the Rar decompression functionality
CVE-2016-4809: Memory allocate error with symbolic links in cpio archives
CVE-2016-5844: undefined behaviour (integer overflow) in iso parser
CVE-2016-6250: Buffer overflow when writing large iso9660 containers
|
|
|
|
|
|
| |
(CVE-2016-6263, CVE-2015-8948, CVE-2016-6262, CVE-2016-6261)
(cherry picked from commit 87698baa9ec19d0554e5233954b6f266efe8b5cd)
|
|
|
|
| |
fixes #5850
|
| |
|
|
|
|
|
|
|
|
| |
(CVE-2016-5419,CVE-2016-5420,CVE-2016-5421)
fixes #6004
(cherry picked from commit da2c76f8a5b41b865d6af208d9bc32764eec75a6)
|
|
|
|
| |
(cherry picked from commit 981244f6c35a6cf7851ef23ebcaf8f27ceb13b9d)
|
|
|
|
| |
(cherry picked from commit 7a470e83787c8c68cdc022a37b7626ded436d43b)
|
|
|
|
|
|
|
|
|
| |
version bump to 3.4.2
Removed *.lax-noc.com mirrors as a file dated 2 Jun 2016 reads:
DISTRO_NO_LONGER_SUPPORTED
(cherry picked from commit 74e83c00d524411a362cf8e1391e05b94edf3c83)
|
|
|
|
|
|
| |
ref #5674
(cherry picked from commit c08456b886621eec011b4b679d553db436bca693)
|
|
|
|
|
|
| |
reported upstream: https://bugs.busybox.net/show_bug.cgi?id=9131
(cherry picked from commit de80fb9f36b7f7a4a0d6d7cdd92fa70c9ed4608c)
|
| |
|
|
|
|
| |
fixes #5958
|
|
|
|
| |
fixes #5952
|