Jesse,
you should use the fully qualified name $(tmpl.varA) for the variables in the
file, as
they get processed in a different bundle and are therefore not "local"
M
Jesse Becker wrote:
> I'm trying to get a basic templating system working for several files,
> and expand_scalars doesn't seem
I'm trying to get a basic templating system working for several files,
and expand_scalars doesn't seem to work. I've found serveral references
to other people having success with this, so I assume I've done
something wrong. I'm running the cfengine 3.0.3 community RPM build.
Here's the basic cod
* site.cf
I use this to set global classes that control what actions needs to be
taken in a particular environment/server. I also set global vars in here,
such as the slist which is passed to bundlesequence - that way all the
stuff that varies depending on what classes are defined is consoli
Hi all,
I'm curious about what good practices people have developed for
organizing their local configuration files, bundles, bodies, special
files, etc. as well as naming conventions for bundles, bodies,
variables, etc. to prevent conflicts with defaults. I checked
www.cfwiki.org but it doesn't se
I'd forgotten all about that syntax as my more intermediate-to-advanced Perl
has gotten a bit rusty. I'm glad to know that works with libpcre. Thanks for
the tip!
Justin
-Original Message-
From: Diego Zamboni [mailto:di...@zzamboni.org]
Sent: Wednesday, February 17, 2010 3:37 PM
To: Ju
Howdy,
PCRE allows you to set the /s flag using (?s) at the beginning of the regex:
PCRE_DOTALL
If this bit is set, a dot metacharater in the pattern matches all char-
acters, including those that indicate newline. Without it, a dot does
not match when the current
Theoretically speaking, this would probably only work with a two-line string,
correct? If so, is there a way to do it generically, i.e. where you don't know
how many newlines may be in the string nor on which line the desired match may
exist, similar to //s in Perl?
Justin
-Original Messag
Mark Burgess wrote on 2010-02-17 15:37:36:
>
> .* only matches up to the newline, so you'll have to say
>
> .*REGEX.*\n.*
That works. It's a bit painful compared with m// so I still wish for that
:) Still thanks for the clarification and your patience with my
stubbornness.
Sincerely,
--
.* only matches up to the newline, so you'll have to say
.*REGEX.*\n.*
nwat...@symcor.com wrote:
> .*OpenSSH.* still does not match
> "SSH-OpenSSH-4.3
> newline";
>
> I don't control the newline but I have to account for it somehow in my
> 'match'.
>
> Sincerely,
> --
> Neil Watson
>
BTW - try the Special topics Guide on Iteration for hints of a workaround
Lebel, Marco wrote:
> Mark,
>
> I appreciate and I am sure that you will find and fix the bug if a bug it is.
>
> That being said we are in a midst of deploying v3.0.2 on our farm of servers
> (half of them are now in hy
.*OpenSSH.* still does not match
"SSH-OpenSSH-4.3
newline";
I don't control the newline but I have to account for it somehow in my
'match'.
Sincerely,
--
Neil Watson
416-673-3465
CONFIDENTIALITY WARNING
This communication, including any attachments, is for the exclusive use of
addres
You can make it simply by writing .*REGEX.*
nwat...@symcor.com wrote:
> Mark Burgess wrote on 2010-02-17 15:25:23:
>
>> You misunderstand what is being matched against. That's documented.
>> It doesn't say that is
>> *searches* the string. It says it matches the string. You confused
>> // no
Mark Burgess wrote on 2010-02-17 15:25:23:
> You misunderstand what is being matched against. That's documented.
> It doesn't say that is
> *searches* the string. It says it matches the string. You confused
> // notation which is a
> search not a comparison.
Ack you are right. I assumed and
You misunderstand what is being matched against. That's documented. It doesn't
say that is
*searches* the string. It says it matches the string. You confused // notation
which is a
search not a comparison.
nwat...@symcor.com wrote:
> Mark Burgess wrote on 2010-02-17 15:09:29:
>
>> There woul
Mark Burgess wrote on 2010-02-17 15:09:29:
> There wouldn't be a line break in your line. This is standard regex
> stuff. Regexs don't go
> through line breaks.
The desired atom is before the line break. This I expect to match as it
does with pcregrep, sed, grep and perl.
[nwat...@unxxhd01
There wouldn't be a line break in your line. This is standard regex stuff.
Regexs don't go
through line breaks.
M
nwat...@symcor.com wrote:
> Mark Burgess wrote on 2010-02-17 15:01:05:
>
>> If there is a line break, use \n in the expression
>
> Assuming I know that there is a line break. Ty
Mark Burgess wrote on 2010-02-17 15:01:05:
>
> If there is a line break, use \n in the expression
Assuming I know that there is a line break. Typically one wants to match
a known string, described as regular expression, in an unknown or
unpredictable string. In this example the desire is to
If there is a line break, use \n in the expression
nwat...@symcor.com wrote:
> This
> "reg" string => "OpenSSH";
> "str" string => "SSH-OpenSSH-4.3
> newline"; # Continued from previous line
>
> classes:
>
> "port" expression => regcmp(".*${reg}.*","${str}");
>
> S
This
"reg" string => "OpenSSH";
"str" string => "SSH-OpenSSH-4.3
newline"; # Continued from previous line
classes:
"port" expression => regcmp(".*${reg}.*","${str}");
Should match. In my previous post the example die not match. I suspect
this is because of the new
Not sure what you mean i.e. did you expect a match or not?
I did try your example and it works for me or should I say it is a match!!!
Marco
-Original Message-
From: nwat...@symcor.com [mailto:nwat...@symcor.com]
Sent: Wednesday, February 17, 2010 2:32 PM
To: Lebel, Marco
Cc: help-cfeng
You are right the regex expects to look at the whole line. That is not as
I would expect from a regular expression. I think this leads to a
problem. Consider this new example where the string continues to another
line.
##
body common control {
bundlesequence => { "tes
Should it not be ".* OpenSSH.*" fro the regular expression?
Marco
-Original Message-
From: help-cfengine-boun...@cfengine.org
[mailto:help-cfengine-boun...@cfengine.org] On Behalf Of nwat...@symcor.com
Sent: Wednesday, February 17, 2010 2:17 PM
To: help-cfengine@cfengine.org
Subject: r
I think regcmp needs to match the whole string, not part of it.
On 18/02/2010 6:17 AM, wrote:
A regcmp I would expect to match does not. What have I done wrong?
[nwat...@unxxhd01 inputs]$ cat regcmp.cf
##
body common control {
bundlesequence => { "test" };
}
bundle com
A regcmp I would expect to match does not. What have I done wrong?
[nwat...@unxxhd01 inputs]$ cat regcmp.cf
##
body common control {
bundlesequence => { "test" };
}
bundle common agent {
classes:
"all" expression => "any";
}
bundle agent test {
classe
Sorry about that
Lebel, Marco wrote:
> Mark,
>
> I appreciate and I am sure that you will find and fix the bug if a bug it is.
>
> That being said we are in a midst of deploying v3.0.2 on our farm of servers
> (half of them are now in hybrid mode) and I do not know when the business
> will al
Mark,
I appreciate and I am sure that you will find and fix the bug if a bug it is.
That being said we are in a midst of deploying v3.0.2 on our farm of servers
(half of them are now in hybrid mode) and I do not know when the business will
allow us to upgrade to 3.0.4.
So, hopefully, other org
Just take a cold shower and give me a few hours. It seems to be a bug, so I'll
fix it.
nwat...@symcor.com wrote:
> There is probably another way to do it if you step back and take a look.
> Another option would be to try the lastest SVN to see if it's fixed and
> will be ready for 3.0.4 due at
There is probably another way to do it if you step back and take a look.
Another option would be to try the lastest SVN to see if it's fixed and
will be ready for 3.0.4 due at the end of the month.
Sincerely,
--
Neil Watson
416-673-3465
CONFIDENTIALITY WARNING
This communication, including an
Neil,
Thanks for your effort, really appreciated...
So I guess my take away is that it should work... but, helas, I came across a
bug. Oh well back to hard coding the iterations.
Marco
-Original Message-
From: nwat...@symcor.com [mailto:nwat...@symcor.com]
Sent: Wednesday, February 1
Seems to work the same on Community 3.0.2 and 3.0.3.
[r...@unxxhd01 inputs]# /home/nwatson/src/cfengine-3.0.3/src/cf-agent -f
./test.cf
R: Server = g1 Client_list = p1
R: Server = g1 Client_list = p2
R: Server = g1 Client_list = p3
List parameter not found while constructing "do_gather" - use
@(
Thanks, Mark! I was beating my head against the wall on this one. :)
I'll update the ticket I opened and let support know this is a known
issue so they can close the ticket.
Justin
-Original Message-
From: Mark Burgess [mailto:mark.burg...@iu.hio.no]
Sent: Wednesday, February 17, 2010 1
What version are you running? If old, could this be related to a bug.
Lebel, Marco wrote:
> Neil,
>
> Yeap I figured you would suggest this, so I did that one too and tried it...
> and sorry to say no success. It does not even iterate over the first entry
> properly.
>
> So I am basically st
Neil,
Yeap I figured you would suggest this, so I did that one too and tried it...
and sorry to say no success. It does not even iterate over the first entry
properly.
So I am basically stuck to hard code the iteration n'est-ce pas?
Marco
-Original Message-
From: nwat...@symcor.com
Perhaps fully qualify info_gathers also.
Sincerely,
--
Neil Watson
416-673-3465
CONFIDENTIALITY WARNING
This communication, including any attachments, is for the exclusive use of
addressee and may contain proprietary and/or confidential information. If you
are not the intended recipient, any
Here is the result:
R: The client list for server gatherer1 are provider1)
R: The client list for server gatherer1 are provider2)
R: The client list for server gatherer1 are provider3)
List parameter not found while constructing "do_gather" - use @(scope.variable)
in calling reference
R: The clie
Greetings Japanese Cfengine Users,
I shall be visiting Japan at the start of April and am gauging interest
to arrange a community meeting or "update" presentation about Cfengine.
Please reply to me privately if you are interested in such a meeting.
Mark
--
Mark Burgess
--
Try using the full qualification @(master.info_)
Lebel, Marco wrote:
> Hello all,
>
>
>
> I have the following :
>
>
>
> bundle agent master
>
> {
>
>vars:
>
>"info_provider[gatherer1]" slist => { "provider1", "provider2",
> "provider3" };
>
>"info_provider[gatherer2
Do you get any warnings? Try passing a fully qualified reference such as
@(master.info_provider[$(info_gatherers)])
Sincerely,
--
Neil Watson
416-673-3465
CONFIDENTIALITY WARNING
This communication, including any attachments, is for the exclusive use of
addressee and may contain proprietary
Hello all,
I have the following :
bundle agent master
{
vars:
"info_provider[gatherer1]" slist => { "provider1", "provider2",
"provider3" };
"info_provider[gatherer2]" slist => { "provider4", "provider5",
"provider6" };
"info_gatherers" slist => getindices("info_provider");
Hi Colin,
I would suggest a search and replace. For some examples see the cf2
reference or my cookbook (
http://watson-wilson.ca/blog/cfcookbook.html#SECTION00015
). Get something going and post back here with what you've tried.
Sincerely,
--
Neil Watson
416-673-3465
CONFIDE
Justin,
have you broken our software again ;-) We found this problem last week -
remarkable that
it's the simplest things that go awry. Anyway, this will be fixed in the
updates that will
be available at the end of the month. Sorry for the inconvenience.
cheers,
Mark
Justin Lloyd wrote:
> I'
41 matches
Mail list logo
