| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
For auto/any packages, currently only auto/any is stored in the
statistics as the new installed version. This patch changes this
behavior, so that the exact version string is saved additionaly.
|
|
|
|
|
| |
Actions that are executed always, does not set the host statistics
to modified. (One always action would lead to always modified hosts)
|
|
|
|
|
|
|
| |
This is used to first update the apt repositories and than install
all current versions in one bcfg2-run. Also the installation of
software from custom repositories (that are configured on the client
via bcfg2) should be possible on the first run.
|
| |
|
|
|
|
|
|
| |
getparent is not supported by xml.etree so try to emulate it
with a generated parent_map. This is the only possibility, because
xml.etree does not store references to the parent elements.
|
|
|
|
| |
This reverts commit 38f55b6257ce51b1cf65ecb3d980d7528c1b53ff.
|
|
|
|
|
|
|
| |
The get_bootstatus method has the correct logic to get the desired
result when bootstatus is unset.
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
|\
| |
| | |
Make the "latest" value more usable in VCS.py
|
| |
| |
| |
| | |
copy only when the working copy is not updated to the latest revision.
|
| |
| |
| |
| |
| |
| |
| | |
When status='ignore', set svcstatus to True so that the entry verifies
properly.
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
You can specify a refname (like refs/heads/master) as revision and you
will always get the current tree of this refname. During verify it is
checked if the ref had changed in the remote repo and if an "update"
is necessary.
|
| |
| |
| |
| |
| |
| |
| | |
Older version of dulwich will not create the directory (even
with mkdir=True) for the repository. So we cannot rely on it.
Because we clean existing folders before, we have to create
the directory in every case.
|
| |
| |
| |
| |
| | |
Add an option to specify whether a command should be executed within
a shell to enable flow control and other shell-specific syntax.
|
| | |
|
| |
| |
| |
| | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| |
| |
| |
| | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
Checking the status for services that don't support it leads to ugly
output in the interactive client on some platforms.
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This new attribute allows the specification of a boot-time status
separately from the current 'status' attribute. This allows for more
fine-grained control over e.g. services that are really just boot
scripts.
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| |
| |
| |
| | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
|
| |
| |
| |
| | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
|
| |
| |
| |
| | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
|
| | |
|
| | |
|
|\ \
| | |
| | |
| | | |
https://github.com/fennm/bcfg2 into maint
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The bcfg2 man page states that -s disabled should stop bcfg2 from
attempting to modify any services, but the Chkconfig driver (at least)
does start the service during the Install phase even with -s disabled.
This patch adds support to the Chkconfig driver for the servicemode
config parameter. It still does chkconfig --add, which I think makes
sense to happen as part of configuration, but it does not attempt
to actually start the service.
|
| | | |
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* Fixed Path entries with secontext='__default__' where no fcontext
rule applied to the path.
* Permitted setting secontext='' when no SELinux context should be
applied to a Path entry
|
| | |
| | |
| | |
| | | |
the fact that there are local changes may be retained). cope with this bug.
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | |
| | |
| | |
| | | |
not set
|
| | |
| | |
| | |
| | | |
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|
| | | |
|
| | | |
|
| | | |
|
|/ /
| |
| |
| |
| |
| |
| |
| | |
There is no good way of sending unicode probes to older clients which do
not have support for them. This change will cause unicode probes to be
skipped for unsupported clients and handled properly for new clients.
Signed-off-by: Sol Jerome <sol.jerome@gmail.com>
|