On Fri, 21 Sep 2012 17:24:50 + Mike Svoboda wrote:
MS> Ted, this is awesome and really should make it into the main trunk.
MS> Having JSON support would rock. YAML would also rock. Both data formats
MS> are used all over the place, and we would have a ton of use cases for this.
MS> Please
On Fri, 21 Sep 2012 18:29:52 +0200 (CEST) neilhwatson wrote:
n> Still I'd like to see some more complex examples of what you had
n> mind.
The Design Center sketch parameters are good examples of the JSON data
I'd like to handle, and you can see in all the test.cf files how the
JSON parameters cu
You can find some experimental support for JSON in my fork of the core
CFEngine here:
https://github.com/tzz/core/tree/feature/json
This is rough undocumented code intended as a proof of concept, written
in my spare time. It's not officially supported by CFEngine!
It provides three things (the
On Thu, 13 Sep 2012 19:39:30 -0500 Diego Zamboni wrote:
DZ> I think this might be a better fit for the Design Center than to the
DZ> COPBL, given that it performs a specific, higher-level task (rather
DZ> than a generic, lower-level task, which is what we try to keep in the
DZ> stdlib).
DZ> We
On Tue, 18 Sep 2012 23:05:13 +0200 (CEST) neilhwatson wrote:
n> I could kick whoever decided to use version numbers that are not
n> numbers. This is a number 2.334. This is not a number 3.4.3.
Please note 2.334 means a different number in many locales! This is
*not* a simple problem!
(relate
On Thu, 13 Sep 2012 14:43:47 -0500 Nick Anderson wrote:
NA> On 09/13/2012 02:38 PM, Ted Zlatanov wrote:
>> As Nick mentioned, VCS::vcs_mirror in Design Center addresses this.
NA> There is a git_failsafe.cf or something similar in contrib that Ted has
NA> been working on. I h
On Fri, 31 Aug 2012 20:00:22 +0200 (CEST) neilhwatson wrote:
> AFAIK it is not possible to call out a single element in a list.
I don't endorse it, but I've heard you can use a file with
grep/head/tail to do this ;)
Ted
___
Help-cfengine mailing list
On Sat, 1 Sep 2012 11:08:13 +0200 (CEST) mpr wrote:
> Currently our policy is kept in git and distributed using
> cf-serverd. Since we already have public/private key authentication
> for git why not use git to distribute the files instead of cf-serverd
> (running "git pull" from update.cf)? Any
On Mon, 10 Sep 2012 22:22:30 +0200 (CEST) sauer wrote:
> as long as we're wishing for anonymous structures. :)
Design Center (through cf-sketch) attempts to alleviate this problem by
feeding a unique prefix to the sketch "entry point" bundle, and
prefixing every variable with it. Thus you can c
On Wed, 5 Sep 2012 20:28:49 +0200 (CEST) juriskrumins wrote:
j> mikesphar Wrote:
>> I believe that is a violation of the Linux
>> Standard Base specifications.
j> 100% agree. But it's obvious we're not in an ideal world. So violation/bugs
happens.
"In theory, theory and practice are the sam
On Thu, 13 Sep 2012 09:23:18 +0200 (CEST) bernhard.gl...@ecologic.eu wrote:
b> on my darwin systems I can't get rid of this annoying log entries each time
cf-execd is calling the cf-agent:
b> dnssd_clientstub write_all: SO_ISDEFUNCT failed 38 Socket operation on
non-socket
b> anybody an
On Sun, 12 Aug 2012 22:37:41 +0200 (CEST) sauer wrote:
n> Ted Zlatanov Wrote:
n> ---
>> On Tue, 7 Aug 2012 10:06:32 +0200 (CEST) sauer
>> wrote:
n> So, I'd like to "repeably randomize" an integer on a host.
On Tue, 7 Aug 2012 10:06:32 +0200 (CEST) sauer wrote:
n> So, I'd like to "repeably randomize" an integer on a host.
n> Specifically, I'd like to use essentially the logic from splayclass or
n> splaytime to pick a time to run a cron job. "Why not run the job from
n> CFEngine?" Well, because the
On Wed, 08 Aug 2012 00:31:26 -0500 Nick Anderson wrote:
NA> Has any thought been given to adding inputs from places other than body
common control?
I think it would make cf-promises unable to validate your policy, since
you'd need to actually run the policy to find out the full extent of it.
On Fri, 10 Aug 2012 10:07:34 +0200 (CEST) davidlee wrote:
n> Mark: Thanks. But why not do both reload and restart? They have different
semantics; they offer different appropriateness towards different behaviours.
I agree, reload and restart should both be provided. Here's the
standard Apach
On Aug 7, 2012, at 7:59, Abid Khwaja wrote:
> Even though I declared “dir_wl10domain" in a common bundle and thus the
> variable is available globally, I still must explicitly reference it as
> $(def.dir_wl10domain) when I reference it in a different bundle, as in the
> above. Is this correct
On Sat, 4 Aug 2012 15:34:01 +0200 (CEST) toddnni wrote:
n> In my opinion "services:" promises are an elegant way to abstract
n> multiple "processes:" and "commands:" promises in one promise. However
n> current "service_policy =>" options are start, stop or disable, and I
n> think that there shou
On Fri, 3 Aug 2012 19:33:43 +0200 (CEST) zzamboni wrote:
n> Ted, I think this has more to do with the bundle being called from a
n> methods: promise than with the fact that it has arguments - removing
n> the arguments results in the same problem, but adding "example2" to
n> the bundlesequence ma
On Fri, 3 Aug 2012 15:41:15 +0200 (CEST) neilhwatson wrote:
n> https://cfengine.com/bugtracker/view.php?id=1226
I think my bug is not quite the same, but it's probably related so I've
added my example to your bug report.
Thanks
Ted
___
Help-cfengine
On Thu, 02 Aug 2012 09:18:02 -0400 Ted Zlatanov wrote:
TZ> On Thu, 2 Aug 2012 00:22:35 -0500 Diego Zamboni wrote:
>>> Yup. Perhaps it could be returned by splitstring() instead of { "" }?
DZ> Bug report time, I guess :)
TZ> I reported it:
TZ> https://cfen
3.4.0 introduces meta: promise types. This example doesn't seem to
DTRT. Is this a bug or am I missing something? Maybe the metadata for
parameterized bundles has some other format?
Thanks
Ted
body common control
{
bundlesequence => { "example" };
}
bundle agent example
{
meta:
On Thu, 2 Aug 2012 00:22:35 -0500 Diego Zamboni wrote:
>> Yup. Perhaps it could be returned by splitstring() instead of { "" }?
DZ> Bug report time, I guess :)
I reported it:
https://cfengine.com/bugtracker/view.php?id=1273
Thanks!
Ted
___
Help-cf
On Thu, 2 Aug 2012 13:51:07 +0200 (CEST) loopx wrote:
n> This one is working fine :
n> processes:
n> redhat_6::
n> "rhsmcertd"
n> #classes=>
if_ok("rhsmcertd_running"),
n>
On Wed, 1 Aug 2012 17:28:59 -0500 Diego Zamboni wrote:
DZ> I believe that's what "cf_null" is supposed to be for:
http://cfengine.com/manuals/cf3-Reference#Special-list-value-cf_005fnull
Oh, I knew this!!! Thanks for pointing it out again, my brain is off
today I guess.
DZ> But it's not trul
On Wed, 1 Aug 2012 22:48:35 +0200 (CEST) neilhwatson wrote:
n> Empty lists seemed to be allowed:
No, unfortunately, I don't think you can get an empty slist with regular
syntax. I give you... the war of the spaces!
body common control
{
bundlesequence => { "test" };
}
bundle agent test
It seems that splitstring("", " ", 100) returns a slist with an empty
string:
body common control
{
bundlesequence => { "test" };
}
bundle agent test
{
vars:
"spaces" slist => splitstring("", " ", "100");
reports:
cfengine::
"spaces '$(spaces)'";
}
=>
R: spaces ''
To
On Wed, 1 Aug 2012 12:00:00 +0200 (CEST) loopx wrote:
n> To make some promise, I need to define classes like this :
n> - XXX_running
n> - XXX_stopped
n> But, I've seen that it's possible to define classes with built-in
n> methods of Cfengine3. In the community lib, I've tried to use
n> "if_else
On Tue, 31 Jul 2012 08:59:08 -0500 Nick Anderson wrote:
NA> The file making the promise $(filename) gets set to different values
NA> based on a class. I thought due to normal ordering, on the first pass
NA> filename should get set to "/tmp/nodebug" since the debug class has not
NA> been raise
On Wed, 25 Jul 2012 16:35:11 +0200 (CEST) neilhwatson wrote:
n> It is not consistent but, I have seen the same think happend manually.
I would use strace and compare the output in a successful
vs. unsuccessful run. It sounds like a library issue or resource
exhaustion, but it's hard to tell.
T
On Tue, 24 Jul 2012 17:23:12 +0200 (CEST) linuxgurugamer wrote:
n> We have several different types of servers, which can be grouped into
several main groups. For this discussion, I'll refer to the following:
n> Group 1:web servers
n> Group 2:database server
n> I've been trying to impl
On Tue, 24 Jul 2012 19:02:41 +0200 (CEST) neilhwatson wrote:
n> Has anyone see this? Testing with 3.4.0b1.
-> Executing '/var/cfengine/bin/cf-serverd'
...(timeout=-678,owner=4294967295,group=4294967295)
-> Completed execution of /var/cfengine/bin/cf-serverd
-> Executing '/var/cfengine/bin/cf-ex
On Thu, 19 Jul 2012 19:07:20 +0200 Marco Marongiu
wrote:
MM> On 29/06/12 22:29, Marco Marongiu wrote:
>>> R: Interface: wlan0 (192.168.0.5)
>>> R: Octets 1-1: $(sys.ipv4_1[tun0])
>>> R: Octets 1-2: $(sys.ipv4_2[tun0])
>>> R: Octets 1-3: $(sys.ipv4_3[tun0])
>>
>> So, apparently, cfengine is abl
On Tue, 17 Jul 2012 20:51:17 +0200 (CEST) sauer wrote:
n> Speaking of the standard library, what are the chances of more
n> formally documenting the miniumum version requirement for using that
n> file? I updated it recently, and found that I had to comment out a
n> few chunks to retain compatabi
On Fri, 6 Jul 2012 16:21:11 +0200 (CEST) neilhwatson wrote:
n> Doesn't [namespacing] just push the concern to a different layer? Now my
bundles can be anything but, I have to worry if my namespace is unique.
Most of Computer Science and Engineering deals with pushing things to a
different lay
On Sat, 23 Jun 2012 20:07:25 +0200 Marco Marongiu
wrote:
MM> On 22/06/12 20:05, Ted Zlatanov wrote:
>> Marco and others, do you want to suggest some good blank templates for
>> the various bundles and bodies? I'll take care of integrating them with
>> cfengine-m
On Mon, 25 Jun 2012 12:40:37 -0400 Ted Zlatanov wrote:
TZ> In Design Center, cf-sketch feeds the bundles a unique prefix for every
TZ> activation, which lets you create a "canonical prefix" for each
TZ> top-level bundle call. grep for "canon_prefix", you'll
On Sat, 23 Jun 2012 19:34:36 +0200 (CEST) davidlee wrote:
n> As it was, the "$(key)_in_file" was OK for a single call, but could fail
when called on different files with the same key.
n> The revised version "$(key)_in_$(file)" allows multiple calls for the same
key on different files; it adds
On Fri, 15 Jun 2012 10:47:23 -0500 Ron Parker wrote:
RP> On Fri, Jun 15, 2012 at 9:57 AM, Marco Marongiu
wrote:
>> I found that having the promises laid out in normal order in the bundles
>> can help a lot in understanding how things actually go on when the
>> policy is run.
>>
>> It would be
On Mon, 18 Jun 2012 18:26:48 +0200 (CEST) msvob...@linkedin.com wrote:
Hi Mike,
that was a very interesting and useful post.
I wanted to bring up a tool I wrote on my own recently, and it's not
CFEngine-specific (although it can produce a CFEngine config, it can
also produce JSON and eventually
On Thu, 21 Jun 2012 15:44:48 +0200 Bas van der Vlies wrote:
BvdV> On 06/21/2012 02:50 PM, Ted Zlatanov wrote:
>> On Thu, 21 Jun 2012 13:46:15 +0200 Bas van der Vlies wrote:
>> I have not updated CFEngine::stdlib in Design Center from the copbl
>> repository because the
On Thu, 21 Jun 2012 11:40:32 -0500 Diego Zamboni wrote:
DZ> Indeed, the version in github.com/cfengine/copbl/ is the master
DZ> copy, from which the others are (sometimes inconsistently)
DZ> updated. We are working on cleaning up some of this to reduce
DZ> confusion.
DZ> Namespace support will
On Thu, 21 Jun 2012 15:07:58 +0200 Riccardo Murri
wrote:
RM> On Thu, Jun 21, 2012 at 3:02 PM, Ted Zlatanov wrote:
>> On Thu, 21 Jun 2012 14:17:53 +0200 Riccardo Murri
>> wrote:
>>
RM> Would it be sensible to add a command-line switch to `cf-promises` to
RM> f
On Fri, 15 Jun 2012 21:39:46 +0200 (CEST) sauer wrote:
n> So, I have a couple of situations where I want to populate an exclude
n> list from an external file. For example, I want to remove the
n> suid/sgid bits from executable files which are group or
n> world-writeable, unless the file is in an
On Tue, 19 Jun 2012 15:59:36 -0500 Ron Parker wrote:
RP> It's not clear to me from the reference manual. If I want to set
RP> permissions on a directory and all of it's descendents that are not
RP> (NFS) mounts, do I want xdev to be true or false?
I think it's similar to the rsync -x flag: if
On Thu, 21 Jun 2012 14:17:53 +0200 Riccardo Murri
wrote:
RM> Would it be sensible to add a command-line switch to `cf-promises` to
RM> force update of `masterfiles/cf_promises_validated` ?
Would it work to just say "cf-promises ... && touch
masterfiles/cf_promises_validated"?
Ted
___
On Thu, 21 Jun 2012 13:46:15 +0200 Bas van der Vlies wrote:
BvdV> Question which cfengine_stdlib.cf to use? Is copbl repository the leading
one
BvdV> There is one in the copbl and one in the design-center repository and
they differ:
I have not updated CFEngine::stdlib in Design Center from t
Ted Z wrote:
>> `cpanm' doesn't support uninstalling, but neither does `cpan'
Because of this, and because CPAN doesn't have the notion of listing
installed modules, I wrote a sketch that simply runs the `cpanm' command
for every Perl module you specify. I just couldn't make it a reasonable
packa
On Tue, 12 Jun 2012 19:49:17 +0200 (CEST) bahamat wrote:
n> The default failsafe.cf includes promises that will start all three daemons.
If you really want it disabled you need to do it in both places. It's not lupus.
I'm really curious what you mean by "it's not lupus."
Ted
__
On Wed, 13 Jun 2012 10:18:45 -0400 Neil Watson
wrote:
NW> I can't recall if this is now possible with 3.3.0. Can we define a list
NW> in an array and then loop the list after passing it?
I do it in the db_install sketch in Design Center:
"base_packages" slist => { "@($(prefix)bycontext
On Thu, 14 Jun 2012 13:22:38 +0200 (CEST) bernhard.gl...@ecologic.eu wrote:
n> Didn't find any way to deliver a shadowed passwort/secret to the darwin auth
system.
n> Instead we can simply copy the users plist.
n> This works at least in OSX 10.7.4
Could you show the solution? That sounds ver
On Wed, 23 May 2012 20:21:06 +0200 (CEST) sauer wrote:
n> It's not git, but with subversion I just use a command promise (which
n> runs "svn update") in combination with a regular CFEngine file copy to
n> get the top-level directory's .svn directory in place.
Could you share the exact commands y
On Wed, 23 May 2012 09:52:13 -0500 Nick Anderson wrote:
NA> On 05/23/2012 09:44 AM, Ted Zlatanov wrote:
>> You may want to copy to an intermediate file (1) and use Augeas or
>> edit_line on a second copy (2) which then gets copied into the final
>> destination (3). That wa
On Wed, 23 May 2012 16:25:05 +0200 (CEST) daveseff wrote:
n> I am distributing a large XML config file to a group of servers for an
n> application. some parameters in the file need to be unique to the
n> machine it's on. Because of it's size I would normally use a copy_from
n> promise from a mast
On Tue, 22 May 2012 20:57:26 -0500 Nick Anderson wrote:
NA> On 05/22/2012 08:41 PM, Nick Anderson wrote:
>> https://cfengine.com/manuals/cf3-Reference#Function-getenv
NA> So env contains USER, and with that we can get gid.
NA> "user" string => getenv("USER","20");
NA> "no" int => get
On Tue, 22 May 2012 18:25:20 -0500 Nick Anderson wrote:
NA> On 05/22/2012 03:36 PM, Ted Zlatanov wrote:
>> How about vcs_mirror? That's pretty clear and the term is not used by
>> Git or Subversion. If you agree I'll make the change on my side.
NA> Works fo
On Tue, 22 May 2012 15:11:23 -0500 Nick Anderson wrote:
NA> So in general are you preferring prefix to passing an array name with
NA> new sketches? And also are you preferring a single usebundle call for
NA> each different configuration?
Yes, and yes. It makes passing data between JSON, Perl,
On Tue, 22 May 2012 12:50:39 -0500 Nick Anderson wrote:
NA> On 05/21/2012 02:51 PM, Ted Zlatanov wrote:
>> Nick, see if you can duplicate the bug on your side and I'll open the
>> ticket. No rush, I have the fix above in place.
NA> I tried to use it manually and none o
On Mon, 21 May 2012 15:51:11 -0400 Ted Zlatanov wrote:
TZ> On Mon, 21 May 2012 15:21:20 -0400 Ted Zlatanov wrote:
TZ> bundle agent vcs_freshclone(prefix)
TZ> {
TZ> vars:
TZ> "bundle_home" string => dirname("$(this.promise_filename)"),
TZ> policy =>
On Mon, 21 May 2012 15:21:20 -0400 Ted Zlatanov wrote:
TZ> But it seems like `this.promise_filename' refers to the top-level
TZ> (i.e. runme.cf) filename, even though the vcs_freshclone promises are
TZ> made in a different file. This makes it impossible to find templates
TZ>
One more improvement: the .git/config file is now enforced :)
This means that it's pretty much impossible for an attacker to make you
check out the wrong thing, even if they can modify the .git/config file
(they can always exploit a race condition, but it's still an improvement).
But it seems lik
On Fri, 18 May 2012 17:00:15 -0400 Ted Zlatanov wrote:
TZ> OK, pushed to https://github.com/tzz/design-center.git
I ran into one issue with the vcs_freshclone sketch, the vcs_configfile
variable is not defined in time on the first pass, so it always tries to
clone. I had to force just one
On Fri, 18 May 2012 15:17:12 -0400 Ted Zlatanov wrote:
TZ> cfsketch itself will make it possible to activate this multiple times,
TZ> so you don't have to deal with passing a 2-level array to describe
TZ> multiple repositories.
TZ> Although it will only support Git for now
On Thu, 17 May 2012 20:27:03 -0500 Nick Anderson wrote:
NA> Hey Analogue, I submitted a pull request to the design-center earlier
NA> today with a sketch that I think does what you were looking for.
NA> https://github.com/cfengine/design-center/pull/49
NA>
https://github.com/nickanderson/desig
On Thu, 17 May 2012 12:51:39 -0500 Nick Anderson wrote:
NA> Any new status updates on cfsketch to come along with that?
Real Soon Now ;)
Ted
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine
On Thu, 17 May 2012 12:34:10 -0500 Nick Anderson wrote:
NA> On 05/17/2012 12:31 PM, Ted Zlatanov wrote:
>> On Thu, 17 May 2012 12:25:53 -0500 Nick Anderson wrote:
NA> On 05/16/2012 03:44 PM, Ted Zlatanov wrote:
>>>> I'd love to see native pluggable CFEngine s
On Thu, 17 May 2012 12:25:53 -0500 Nick Anderson wrote:
NA> On 05/16/2012 03:44 PM, Ted Zlatanov wrote:
>> I'd love to see native pluggable CFEngine support for VCS pulls, like it
>> has for package management.
NA> That would be cool. It seems to be a pattern I do a lot
On Wed, 16 May 2012 14:33:03 -0500 Nick Anderson wrote:
NA> On 05/16/2012 02:28 PM, no-re...@cfengine.com wrote:
>> bundle agent git_master
>> {
>> commands:
>> "/usr/bin/git pull origin"
>> comment => "Update /repo from git repository",
>> contain => silent_in_dir("/repo");
>> }
NA> What hap
On Mon, 30 Apr 2012 10:09:37 -0500 Nick Anderson wrote:
NA> Well I think if we approach this with the expectation that we will stop
NA> someone with root access from doing anything then we just performing an
NA> exercise in futility.
NA> You could argue that security is only the the inverse of
On Mon, 30 Apr 2012 16:24:52 +0200 (CEST) msvob...@linkedin.com wrote:
n> Giving developers root access to development machines is a known evil.
n> I would rather not give root access to people who aren't
n> administrators, but in reality, this doesn't happen.
n> Folks that run QA, performance e
On Thu, 26 Apr 2012 22:08:56 +0200 (CEST) msvob...@linkedin.com wrote:
ms> Most exploits / data loss happen from _within_ the organization. If I
ms> give a developer / fellow co-worker root access to a machine inside
ms> our network, he can snoop around the Cfengine area and see things he
ms> re
On Fri, 27 Apr 2012 23:56:49 +0200 Mikhail Gusarov
wrote:
MG> On 27.04.2012 23:22, no-re...@cfengine.com wrote:
>> Is there a way to tell cfengine not to start if there is no networking up
>> without hacking startup script?
MG> It does not hang, but it tries to resolve IP address into DNS nam
On Wed, 18 Apr 2012 11:35:01 +0200 Bas van der Vlies wrote:
BvdV> I am using slist because configuration parameters can contain more then
one value. For most configuration files
BvdV> it is nice that a 'list' variable is repeated multiple times, eg:
BvdV> [%CFEngine BEGIN %]
BvdV> search $(te
On Thu, 5 Apr 2012 05:59:35 -0700 Aleksey Tsalolikhin
wrote:
AT> Hi. Is there a tidy or pretty-print program for CFEngine 3?
AT> I am just looking for something to make sure there is uniform
AT> indentation style in my examples, I've got like 200 of them by now and
AT> I don't want to correct t
On Tue, 3 Apr 2012 20:45:08 +0200 (CEST) sauer wrote:
n> I've been working on this for sudoers, where later entries override
n> earlier entries. The way I'm currently doing it is to have a
n> directory full of small templates which start with a number. It looks
n> like the new lsdir function w
On Tue, 20 Mar 2012 21:53:37 -0500 Nick Anderson wrote:
NA> I'm just curious. At what point do you decide a module is better than doing
things natively?
At JOB^1 I wrote a Perl script to look up company inventory metadata and
feed it to cfengine through the module interface. The script was m
On Tue, 13 Mar 2012 09:30:04 -0500 Nick Anderson wrote:
NA> On 03/13/2012 09:22 AM, Mikhail Gusarov wrote:
>> Something like "git clean -fxd"?
NA> Something like that. When I do it with svn I like to track which files
NA> had which state so I can send a report to the person who cared about the
On Mon, 12 Mar 2012 18:43:52 -0500 Nick Anderson wrote:
NA> I'm kind of surprised that rsync is so much faster than svn update. But
NA> I haven't tried to use subversion to distribute lots of files, only used
NA> for small trees. But i seem to use subversion for distributing small
NA> trees a lo
On Fri, 24 Feb 2012 07:35:34 -0800 Aleksey Tsalolikhin
wrote:
AT> This site has been using CFEngine 2 for 5 years and they have a very
AT> large policy set. We're talking thousands or more of little inputs files.
AT> They do EVERYTHING in CFEngine. Their system architect teaches all
AT> their
On Tue, 21 Feb 2012 12:54:44 +0100 (CET) rsdejo...@gmail.com wrote:
n> I have a question ,
n> We want to copy the repository files to rhel 5 + 6 and sles 9 / 10 / 11
n> I know how to divide into OS ,but how do you divide further in the tree ?
n> example => /etc/yum.repos.d/baseline.repo
n> It
On Wed, 4 Jan 2012 04:25:56 +0100 (CET) bughari wrote:
n> Is there command available in cfengine3 to do health check of all cf-agent
is active and talking to cf-serverd which we can run from cfengine master
server.
You could use loggly or Splunk or plain syslog or something similar to
collect
On Wed, 28 Dec 2011 13:13:13 + Michael Gliwinski
wrote:
MG> Another option which I'm considering now is to let managed nodes
MG> pull from VCS directly into their own masterfiles and change policy
MG> in update.cf to just copy locally into inputs/ modules/ etc. after
MG> doing some local pr
On Mon, 19 Dec 2011 13:42:03 -0600 Nick Anderson wrote:
NA> What would the best way to capture a list of the files detected changed
NA> into a list?
Use etckeeper+Git, it will do a much better job, especially if you
snapshot often. Then you just use Git commands from CFEngine or outside
(which
On Fri, 16 Dec 2011 15:53:10 +0100 (CET) neilhwatson wrote:
n> One other point comes to mind if you'll indulge me again. I think this was
n> mentioned before. Passing parameters to bundles. I'd like to see this done
n> better. The current positional approach is very limited. It might be
hel
On Fri, 16 Dec 2011 08:06:32 +0100 Mark Burgess wrote:
MB> Please forget you ever saw this.
Saw what?!?!?
MB> This might look superficially attractive to a programmer but it
MB> suffers from many problems, not the least of which would be a major
MB> increase in the size and complexity of the p
On Fri, 16 Dec 2011 11:13:29 +0100 (CET) lauwersw wrote:
n> But your comment made me think: maybe "local" classes should remain, but be
transformed in "scoped" classes, so they can be accessed from other bundles as
well.
I think local classes should remain as they are, but the user should
have
On Tue, 13 Dec 2011 15:50:33 -0600 Ted Zlatanov wrote:
TZ> On Wed, 07 Dec 2011 14:43:00 +0100 Marco Marongiu
wrote:
MM> The bad ones: as soon as I did it, font-locking broke. The following
MM> error appears in the *Messages* window:
>>> Error during redisplay: (void-
On Wed, 14 Dec 2011 15:50:13 + (GMT) Mark Dixon
wrote:
MD> OK, so this works because of rule 5 of agent normal ordering ("The
MD> order of promises within one of the above types follows their top-down
MD> ordering within the bundle itself") from the reference manual?
MD> http://cfengine.co
On Wed, 14 Dec 2011 14:54:02 + (GMT) Mark Dixon
wrote:
MD> How would you cleanly do an if..elif..else ? The example you give
MD> appears to be an if..else.
cfengine doesn't have that. It has a switch statement, really, but it
will spin through the switch until it's settled (up to 3 passes
While we're discussing fundamentals... :) Some small ideas I think may
be useful to the cfengine users:
1) allow passing body parameters as JSON, not just positional
parameters. Take the standard file permissions promise:
"$(crond)/cfrun"
perms => mo("644","root");
it instead become
On Wed, 14 Dec 2011 10:08:22 + (GMT) Mark Dixon
wrote:
MD> However, it wasn't the "if" bit I was fuzzy about, it was how people use
MD> the language to do the job of an "if..elif..else" construct in the
MD> clearest and least fragile way.
MD> Extending the example on the devops page naiv
On Wed, 07 Dec 2011 14:43:00 +0100 Marco Marongiu
wrote:
MM> The bad ones: as soon as I did it, font-locking broke. The following
MM> error appears in the *Messages* window:
>> Error during redisplay: (void-variable font-lock-constant-name-face)
MM> Now class names are black as normal text (t
On Sat, 3 Dec 2011 00:58:10 +0100 (CET) theaccount wrote:
n> Is there a tool that you would recommend for remote script execution?
I know some people that use func (Python tool) and are happy with it.
Rundeck is also interesting.
Ted
___
Help-cfengin
On Wed, 07 Dec 2011 23:56:19 +0100 Nicolas Charles
wrote:
>> Check it out here http://www.packtpub.com/cfengine-3-beginners-guide/book
NC> I've ordered this book a couple of weeks ago and start to review it, to
NC> see if my team would learn from it. I really have mixed feelings about
NC> th
On Fri, 02 Dec 2011 10:38:14 -0500 Ted Zlatanov wrote:
TZ> For the attribute 0 indent, it's a feature but I can change it to give
TZ> an optional extra indent to continued lines. I was trying to make the
TZ> appearance uniform and consistent but I see how it can be useful
On Fri, 02 Dec 2011 16:48:38 +0100 Mikhail Gusarov
wrote:
MG> On 12/02/2011 04:44 PM, Ted Zlatanov wrote:
MG> Could you ask Emacs guys (or make a patch yourself) to rename
MG> cfengine-mode to cfengine2-mode, or something similar?
>>
>> That would break backwards comp
On Fri, 02 Dec 2011 15:59:49 +0100 Mikhail Gusarov
wrote:
MG> Ted,
MG> On 11/25/2011 02:26 PM, Ted Zlatanov wrote:
>> cfengine SVN maintainers: please review the attached file and especially
>> the copyright notice. I believe it's no different from before as far as
>
On Fri, 02 Dec 2011 15:47:06 +0100 Marco Marongiu
wrote:
MM> I confirm the following:
MM> My platform: Ubuntu 10.04.3 LTS (as appears in motd)
MM> My Emacs:
MM> This is GNU Emacs 23.1.1 (i486-pc-linux-gnu, GTK+ Version 2.20.1)
MM> of 2011-03-04 on palmer, modified by Debian
...
MM> I confirm
On Tue, 29 Nov 2011 08:55:46 -0500 Ted Zlatanov wrote:
TZ> On Mon, 28 Nov 2011 17:21:06 +0100 Marco Marongiu
wrote:
MM> I've just done a quick test, and things seem to work smoothly (GNU Emacs
MM> 23.1.1 running on Ubuntu Linux 10.04 LTS)
TZ> On Mon, 28 Nov 2011 17:
On Wed, 30 Nov 2011 20:15:56 -0800 (PST) "Carl E. Ma"
wrote:
CEM> I am new to Cfengine3. Is it possible to use cfengine to clone one
server's configuration to another server?
CEM> My intention is using cfengine to deploy system/application
CEM> configuration change on VM host. If this host c
On Wed, 23 Nov 2011 15:45:24 +0100 (CET) no-re...@cfengine.com wrote:
n> Forum: CFEngine Help
n> Subject: Re: putting cfengine mailing list on Gmane?
n> Author: zzamboni
n> Link to topic: https://cfengine.com/forum/read.php?3,23979,24049#msg-24049
n> List archiving has resumed under
http://dir.
1 - 100 of 146 matches
Mail list logo
