This addresses the following security issues:
* Ansible: Splunk and Sumologic callback plugins leak sensitive data in logs (CVE-2019-14864)
* CVE-2019-14846 - Several Ansible plugins could disclose aws
credentials in log files. inventory/aws_ec2.py, inventory/aws_rds.py,
lookup/aws_account_attribute.py, and lookup/aws_secret.py,
lookup/aws_ssm.py use the boto3 library from the Ansible process. The
boto3 library logs credentials at log level DEBUG. If Ansible's
logging was enabled (by setting LOG_PATH to a value) Ansible would set
the global log level to DEBUG. This was inherited by boto and would
then log boto credentials to the file specified by LOG_PATH. This did
not affect aws ansible modules as those are executed in a separate
process. This has been fixed by switching to log level INFO
* Convert CLI provided passwords to text initially, to prevent unsafe
context being lost when converting from bytes->text during post
processing of PlayContext. This prevents CLI provided passwords from
being incorrectly templated (CVE-2019-14856)
* properly hide parameters marked with no_log in suboptions when
invalid parameters are passed to the module (CVE-2019-14858)
Changelog: 24220a618a/changelogs/CHANGELOG-v2.8.rst
This fixes the following security issues:
* Ansible: Splunk and Sumologic callback plugins leak sensitive data
in logs (CVE-2019-14864)
* CVE-2019-14846 - Several Ansible plugins could disclose aws
credentials in log files. inventory/aws_ec2.py, inventory/aws_rds.py,
lookup/aws_account_attribute.py, and lookup/aws_secret.py,
lookup/aws_ssm.py use the boto3 library from the Ansible process. The
boto3 library logs credentials at log level DEBUG. If Ansible's
logging was enabled (by setting LOG_PATH to a value) Ansible would set
the global log level to DEBUG. This was inherited by boto and would
then log boto credentials to the file specified by LOG_PATH. This did
not affect aws ansible modules as those are executed in a separate
process. This has been fixed by switching to log level INFO
* Convert CLI provided passwords to text initially, to prevent unsafe
context being lost when converting from bytes->text during post
processing of PlayContext. This prevents CLI provided passwords from
being incorrectly templated (CVE-2019-14856)
* properly hide parameters marked with no_log in suboptions when invalid
parameters are passed to the module (CVE-2019-14858)
* resolves CVE-2019-10206, by avoiding templating passwords from
prompt as it is probable they have special characters.
* Handle improper variable substitution that was happening in
safe_eval, it was always meant to just do 'type enforcement' and have
Jinja2 deal with all variable interpolation. Also see CVE-2019-10156
Changelog: 0623dedf2d/changelogs/CHANGELOG-v2.7.rst (v2-7-15)
Update to latest version & updated the patch file to match with the
lastest verison.
Fixes the following security issue:
* CVE-2019-19722: Mails with group addresses in From or To fields
caused crash in push notification drivers.
This makes ~2.5x speed up of an empty container instantiate, hence reduces
rebuild time of system with many declarative containers.
Note that this doesn't affect production systems much, becaseu those most
likely already include `minimal.nix` profile.
Communication between spamc and spamd would fail with messages about
addresses being too long:
error: Bad arg length for Socket::unpack_sockaddr_in, length is 28,
should be 16
By adding Socket6 as a buildInput, spamd is now able to process emails
without choking on IPv6 addresses.
