From adam.trickett at iredale.net  Fri Oct  5 05:30:00 2007
From: adam.trickett at iredale.net (Adam Trickett)
Date: Fri, 5 Oct 2007 13:30:00 +0100
Subject: [Thamesvalley-pm] Fwd: London Perl Workshop 2007
Message-ID: <200710051330.07973.adam.trickett@iredale.net>

This came out today, and should be well worth attending.


----------  Forwarded Message  ----------

Please spread this far and wide to blogs, sites, mailing lists, print it and 
stick it up on a notice board. If you have an internal dev mailing list, spam 
that. Bonus points for people posting it to other london language lists. 

----

London Perl Workshop 2007

So it's that time of year again, the nights are growing long, the mornings are 
growing cold and people who should know better once again decide to organise 
a conference. That's right it's time for this year's London Perl Workshop. So 
let's get down to the details, because we haven't much time ...

When?

Saturday 1st of December (10:00->17:30)

Where?

Cavendish Campus, University of Westminster, London (Thanks again!)
(http://www.wmin.ac.uk/page-7679-smhp=4459)

What?

This year we will be encouraging people from other languages to participate so 
we can continue Perl's long tradition of stealing the best ideas from other 
languages (including future versions of Perl itself). We will also be doing a 
lot more lightning talks (because people like them).

What can you do?

Register to attend.

	http://conferences.yapceurope.org/lpw2007/

Propose a talk (remember lightning talks particularly welcome!)

Same as above. n.b. you need to be registered to submit a proposal. The 
deadline for proposals is if I remember, remember correctly, the 5th of 
November. But please don't leave it to the last minute.
(Talks should be 5 (lightning), 15 or 30 minutes long, unless it's something 
exceptional.)

-------------------------------------------------------

-- 
Adam Trickett
Overton, HANTS, UK

In the field of observation, chance favours only the
prepared minds.
	-- Louis Pasteur
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://mail.pm.org/pipermail/thamesvalley-pm/attachments/20071005/f85f71df/attachment.bin 

From gmatt at nerc.ac.uk  Tue Oct 16 06:45:31 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Tue, 16 Oct 2007 14:45:31 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
Message-ID: <4714C07B.3080506@nerc.ac.uk>

This question will probably show my naivity... I have a perl script that 
I want to make use of in my own perl scripts. I dont want to have to 
mangle the original script as I will loose the ability to easily 
incorporate updated versions. How can I call the first script from my 
own? Do I have to resort to a system()

GREG
-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From charlie at peopleandplanet.org  Tue Oct 16 06:49:29 2007
From: charlie at peopleandplanet.org (Charlie Harvey)
Date: Tue, 16 Oct 2007 14:49:29 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
In-Reply-To: <4714C07B.3080506@nerc.ac.uk>
References: <4714C07B.3080506@nerc.ac.uk>
Message-ID: <4714C169.8030801@peopleandplanet.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Greg Matthews wrote:
> This question will probably show my naivity... I have a perl script that 
> I want to make use of in my own perl scripts. I dont want to have to 
> mangle the original script as I will loose the ability to easily 
> incorporate updated versions. How can I call the first script from my 
> own? Do I have to resort to a system()
> 
> GREG

Hey Greg,

Sounds like you need to use do or eval.

do EXPR

Uses the value of EXPR as a filename and executes the contents of the
file as a Perl script.

    do 'stat.pl';

is just like

    eval `cat stat.pl`

http://perldoc.perl.org/functions/do.html

Cheers,

- --
Charlie Harvey
IT Officer
People & Planet

Printing causes pollution and wastes energy and paper.
Please consider not printing this email. Thanks.


Shared Planet 2007 | 16 - 18 November | Sheffield University

The UK's biggest student campaigning event. Designed to inform and
inspire you to take action on poverty, human rights and the environment.

Tickets cost just ?16 before 26 October. Book now

http://peopleandplanet.org/sharedplanet
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHFMFpy48xIMwPXm0RAlBpAJ9ZGG+pkwd3d6AJrJ3HXbTyGLS+pACglc6Y
61XtKYBszpe67SU7U9BJaMc=
=1uy7
-----END PGP SIGNATURE-----

From gmatt at nerc.ac.uk  Tue Oct 16 07:26:21 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Tue, 16 Oct 2007 15:26:21 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
In-Reply-To: <4714C169.8030801@peopleandplanet.org>
References: <4714C07B.3080506@nerc.ac.uk>
	<4714C169.8030801@peopleandplanet.org>
Message-ID: <4714CA0D.2060400@nerc.ac.uk>

Charlie Harvey wrote:
> Hey Greg,
> 
> Sounds like you need to use do or eval.
> 
> do EXPR

hrmmm.. not used that before... at the moment I'm using this approach:

     open(PCA, "$pca_bin $pca_opts |") or die "No fork: $!\n";
     my @result = <PCA>;
     close PCA;

and set $pca_bin to be the path to the perl utility. Is this acceptable? 
I'm using a very similar approach in order to use ssh from perl. It 
works but I'm not clear if this is best practice.

its not clear from the docs whether you can capture output from "do" (as 
I do above with the pipe).

G

> 
> Uses the value of EXPR as a filename and executes the contents of the
> file as a Perl script.
> 
>     do 'stat.pl';
> 
> is just like
> 
>     eval `cat stat.pl`
> 
> http://perldoc.perl.org/functions/do.html
> 


-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From Rafiq.Ismail at MorganStanley.com  Wed Oct 17 01:57:47 2007
From: Rafiq.Ismail at MorganStanley.com (Ismail, Rafiq (IT))
Date: Wed, 17 Oct 2007 09:57:47 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
In-Reply-To: <4714CA0D.2060400@nerc.ac.uk>
References: <4714C07B.3080506@nerc.ac.uk><4714C169.8030801@peopleandplanet.org>
	<4714CA0D.2060400@nerc.ac.uk>
Message-ID: <0F354FDFECA12647B75F7B154EC382850855BDB4@LNWEXMB58.msad.ms.com>

I would probably suggest putting all your common/reused components/logic
in a module, where each of your scripts can share this. 

<morningRamble>
If, however, you're going to use do (perldoc -f do), which I'm not so
fond of, then I would recommend using such a pattern that insures you:

i) Minimise use/overlap of global variables between scripts
ii) Try and wrap your functionality in subroutines, as opposed to the
main body of the script 
Iii) If you can't go with (ii) then at least ensure that @ARGV is set to
the expected input of the included script.

When you do a 'do', you're basically evaluating the script into your
current execution thread.  So, if you have all your processing in the
main body of the script, dependent on the state of @ARGV - this might
cause confusion/collision when you evaluate the script from one which
has its own @ARGV and surrounding scope to deal with.  This is why it's
better to try and strip your processing out of the main body.

A way to get around this would be to have some main method which is
directly invoked when your script is executed - but not otherwise.
Further, if needed, try to return a result set which can be used by your
calling application.: 

fooMain(@ARGV) unless caller();

sub fooMain {
  my @args = @_;
  my @stuff;
#... Do stuff
  return @stuff; 
}


If your "do 'foo_script.pl'" succeeds, then you can directly call your
<script>Main() function and pass it your $pca_opts:

Unless (do 'fooScript.pl') {
	croak("could not eval script: $!")
}
My @pca_opts = split '\s', $pca_opts
My @results = fooMain( @pca_opts );


BTW, the reason I call it <script>Main is because you'd otherwise keep
redefining 'main' (assuming this is what you called your main methods)
with each new 'do <script>'.  

I've seen some horrid/bloated examples of reuse through do - and nasty
hard to find bugs from the side effects of global variables used across
multiple eval'ed scripts.  I still suggest re-use through a common
module.  

</morningRamble>
Raf





 

 

> -----Original Message-----
> From: 
> thamesvalley-pm-bounces+rafiq.ismail=morganstanley.com at pm.org 
> [mailto:thamesvalley-pm-bounces+rafiq.ismail=morganstanley.com
@pm.org] On Behalf Of Greg Matthews
> Sent: 16 October 2007 15:26
> To: ThamesValley Perl Mongers
> Subject: Re: [Thamesvalley-pm] using third party perl scripts
> 
> Charlie Harvey wrote:
> > Hey Greg,
> > 
> > Sounds like you need to use do or eval.
> > 
> > do EXPR
> 
> hrmmm.. not used that before... at the moment I'm using this approach:
> 
>      open(PCA, "$pca_bin $pca_opts |") or die "No fork: $!\n";
>      my @result = <PCA>;
>      close PCA;
> 
> and set $pca_bin to be the path to the perl utility. Is this 
> acceptable? 
> I'm using a very similar approach in order to use ssh from 
> perl. It works but I'm not clear if this is best practice.
> 
> its not clear from the docs whether you can capture output 
> from "do" (as I do above with the pipe).
> 
> G
> 
> > 
> > Uses the value of EXPR as a filename and executes the 
> contents of the 
> > file as a Perl script.
> > 
> >     do 'stat.pl';
> > 
> > is just like
> > 
> >     eval `cat stat.pl`
> > 
> > http://perldoc.perl.org/functions/do.html
> > 
> 
> 
> -- 
> Greg Matthews           01491 692445
> Head of UNIX/Linux, iTSS Wallingford
> 
> --
> This message (and any attachments) is for the recipient only. 
> NERC is subject to the Freedom of Information Act 2000 and 
> the contents of this email and any reply you make may be 
> disclosed by NERC unless it is exempt from release under the 
> Act. Any material supplied to NERC may be stored in an 
> electronic records management system.
> 
> _______________________________________________
> Thamesvalley-pm mailing list
> Thamesvalley-pm at pm.org
> http://mail.pm.org/mailman/listinfo/thamesvalley-pm
>
--------------------------------------------------------

NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

From Rafiq.Ismail at MorganStanley.com  Wed Oct 17 03:04:05 2007
From: Rafiq.Ismail at MorganStanley.com (Ismail, Rafiq (IT))
Date: Wed, 17 Oct 2007 11:04:05 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
In-Reply-To: <4714CA0D.2060400@nerc.ac.uk>
References: <4714C07B.3080506@nerc.ac.uk><4714C169.8030801@peopleandplanet.org>
	<4714CA0D.2060400@nerc.ac.uk>
Message-ID: <0F354FDFECA12647B75F7B154EC382850855BE1D@LNWEXMB58.msad.ms.com>

Hi Charlie,
 

> -----Original Message-----
> From: 
> thamesvalley-pm-bounces+rafiq.ismail=morganstanley.com at pm.org 
> [mailto:thamesvalley-pm-bounces+rafiq.ismail=morganstanley.com
@pm.org] On Behalf Of Greg Matthews
> Sent: 16 October 2007 15:26
> To: ThamesValley Perl Mongers
> Subject: Re: [Thamesvalley-pm] using third party perl scripts
> 
> Charlie Harvey wrote:
> > Hey Greg,
> > 
> > Sounds like you need to use do or eval.
> > 
> > do EXPR
> its not clear from the docs whether you can capture output 
> from "do" (as I do above with the pipe).

Short answer no.  However you can capture the output from STDOUT:

Use FileHandle;
my $stdoutCapture;
open FAKE_OUT, ">", \$stdoutCapture;
select (FAKE_OUT);
do "script";
close FAKE_OUT;
select (STDOUT)

## Examine $stdoutCapture


Thanks,

Raf
--------------------------------------------------------

NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

From gmatt at nerc.ac.uk  Wed Oct 17 04:37:02 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Wed, 17 Oct 2007 12:37:02 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
In-Reply-To: <0F354FDFECA12647B75F7B154EC382850855BE1D@LNWEXMB58.msad.ms.com>
References: <4714C07B.3080506@nerc.ac.uk><4714C169.8030801@peopleandplanet.org>	<4714CA0D.2060400@nerc.ac.uk>
	<0F354FDFECA12647B75F7B154EC382850855BE1D@LNWEXMB58.msad.ms.com>
Message-ID: <4715F3DE.10904@nerc.ac.uk>

Ismail, Rafiq (IT) wrote:
>>> do EXPR
>> its not clear from the docs whether you can capture output 
>> from "do" (as I do above with the pipe).
> 
> Short answer no.  However you can capture the output from STDOUT:

thanks for the various advice. I must admit I was concerned with using 
"do". I was thinking about using it for config file slurping but was 
wondering what would happen if some malicious/ignorant user had write 
access to the config file... presumably they could execute any perl code 
they wanted as the userid that the main script is running as?

As for taking the code out of the 3rd party perl script - I am loathe to 
do this as the script is in active development and I would have to 
deconstruct new versions as they were published creating lots of work 
for something that is supposed to save me lots of work!

So far, using pipes looks the best option.

GREG
ps this particular project is the first time I'm using perl to produce 
html so I will probably have questions about that soon too!

> 
> Use FileHandle;
> my $stdoutCapture;
> open FAKE_OUT, ">", \$stdoutCapture;
> select (FAKE_OUT);
> do "script";
> close FAKE_OUT;
> select (STDOUT)
> 
> ## Examine $stdoutCapture
> 
> 
> Thanks,
> 
> Raf
> --------------------------------------------------------
> 
> NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.
> _______________________________________________
> Thamesvalley-pm mailing list
> Thamesvalley-pm at pm.org
> http://mail.pm.org/mailman/listinfo/thamesvalley-pm
> 


-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From Rafiq.Ismail at MorganStanley.com  Wed Oct 17 06:30:56 2007
From: Rafiq.Ismail at MorganStanley.com (Ismail, Rafiq (IT))
Date: Wed, 17 Oct 2007 14:30:56 +0100
Subject: [Thamesvalley-pm] using third party perl scripts
In-Reply-To: <4715F3DE.10904@nerc.ac.uk>
References: <4714C07B.3080506@nerc.ac.uk><4714C169.8030801@peopleandplanet.org>	<4714CA0D.2060400@nerc.ac.uk><0F354FDFECA12647B75F7B154EC382850855BE1D@LNWEXMB58.msad.ms.com>
	<4715F3DE.10904@nerc.ac.uk>
Message-ID: <0F354FDFECA12647B75F7B154EC382850855BF28@LNWEXMB58.msad.ms.com>

> -----Original Message-----
> From: thamesvalley-pm-bounces+rafiq.ismail=ms.com at pm.org 
> [mailto:thamesvalley-pm-bounces+rafiq.ismail=ms.com at pm.org] 
> On Behalf Of Greg Matthews
> Sent: 17 October 2007 12:37
> To: ThamesValley Perl Mongers
> Subject: Re: [Thamesvalley-pm] using third party perl scripts
> 
> Ismail, Rafiq (IT) wrote:
> >>> do EXPR
> >> its not clear from the docs whether you can capture output 
> from "do" 
> >> (as I do above with the pipe).
> > 
> > Short answer no.  However you can capture the output from STDOUT:
> 
> thanks for the various advice. I must admit I was concerned 
> with using "do". I was thinking about using it for config 
> file slurping but was wondering what would happen if some 
> malicious/ignorant user had write access to the config 
> file... presumably they could execute any perl code they 
> wanted as the userid that the main script is running as?

Not a good idea.  I tend to use XML::Simple for basic xml config files.
Have a look on cpan for 'Config'  there's a host of stuff out there.
--------------------------------------------------------

NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

From Rafiq.Ismail at MorganStanley.com  Wed Oct 17 10:50:03 2007
From: Rafiq.Ismail at MorganStanley.com (Ismail, Rafiq (IT))
Date: Wed, 17 Oct 2007 18:50:03 +0100
Subject: [Thamesvalley-pm] How about a social?
In-Reply-To: <200709241549.12583.adam.trickett@iredale.net>
References: <0F354FDFECA12647B75F7B154EC3828507F2FCE4@LNWEXMB58.msad.ms.com>
	<200709241549.12583.adam.trickett@iredale.net>
Message-ID: <0F354FDFECA12647B75F7B154EC382850855C047@LNWEXMB58.msad.ms.com>

Adam,

Reading about dead pm's on london.pm, I feel an enhanced sense of
preserving what looked like the start of a fairly valuable and
successful group.


> -----Original Message-----
> From: thamesvalley-pm-bounces+rafiq.ismail=ms.com at pm.org 
> [mailto:thamesvalley-pm-bounces+rafiq.ismail=ms.com at pm.org] 
> On Behalf Of Adam Trickett
> Sent: 24 September 2007 15:49
> To: ThamesValley Perl Mongers
> Subject: Re: [Thamesvalley-pm] How about a social?
> 
>
> The SC-LUG meetings are in Reading and usually on Wednesday 
> at the Back of Beyond Pub. The next one being 19:30 Wed 10th 
> Oct 2007, I can get that far and I need to say hello to 
> people there in person anyway.

Any chance that you could make this happen for the next meeting?  I
think that Reading is fairly central and perhaps if we could prepare
some talks and share air-time, this might be the way forward.  How would
the logistics of such a gate-crash work?

Raf
--------------------------------------------------------

NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

From gmatt at nerc.ac.uk  Thu Oct 18 08:48:20 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Thu, 18 Oct 2007 16:48:20 +0100
Subject: [Thamesvalley-pm] generating HTML
Message-ID: <47178044.7090102@nerc.ac.uk>

Well... I threatened this email so here goes...

Is there a neater way of generating html than this:

     print INDEX "    <tr>
       <td style=\"vertical-align: top;\">$host<br>
       </td>\n";
??

At the moment my code loops round generating table rows (of which the 
above is the first column).

Given that the table is of an arbitrary size (depending on the number of 
hosts defined) am I approaching this the right way? ie directly printing 
out the raw html?

The generated page is essentially static, it is just generated by a cron 
job and the table contains links to pages of generated data for each 
host (if available).

I'm used to making my perl look neat and readable but now that it 
contains huge dollops of html (such as the definition of my $header and 
my $footer) it looks a mess.

Let me know if anyone wants to see the whole thing (and have a good 
laugh)...

GREG
-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From abw at wardley.org  Thu Oct 18 11:13:32 2007
From: abw at wardley.org (Andy Wardley)
Date: Thu, 18 Oct 2007 19:13:32 +0100
Subject: [Thamesvalley-pm] generating HTML
In-Reply-To: <47178044.7090102@nerc.ac.uk>
References: <47178044.7090102@nerc.ac.uk>
Message-ID: <4717A24C.8010604@wardley.org>

Greg Matthews wrote:
> Is there a neater way of generating html than this:

You can go down the templating route, using either the Template
Toolkit (TT), HTML::Template, or one of the many other template modules.

Here it is using TT:

   use Template;

   # create a TT processor and tell it where your templates are
   my $tt   = Template->new( INCLUDE_PATH => '/path/to/templates' );

   # define some data
   my $data = {
      hosts => [ 'host1', 'host2', 'host3', ...etc... ],
   };

   $tt->process('index.html', $data) || die $tt->error();

Then in the index.html template (in the /path/to/templates directory) you
would write something like this:

   [% INCLUDE header %]

   <table>
   [% FOREACH host IN hosts %]
      <tr>
       <td style="vertical-align: top;">[% host %]</td>
      </tr>
   [% END %]
   </table>

   [% INCLUDE footer %]


Your header/footer templates would also need to be in the /path/to/templates
directory.

See http://tt2.org for more info on TT.  See the 'tpage' and 'ttree' scripts
(distributed with TT) for processing templates from the command line (or cron
job).

HTH
A

From dave at dave.org.uk  Thu Oct 18 12:32:41 2007
From: dave at dave.org.uk (Dave Cross)
Date: Thu, 18 Oct 2007 20:32:41 +0100
Subject: [Thamesvalley-pm] generating HTML
In-Reply-To: <47178044.7090102@nerc.ac.uk>
References: <47178044.7090102@nerc.ac.uk>
Message-ID: <4717B4D9.90309@dave.org.uk>

Greg Matthews wrote:
> Well... I threatened this email so here goes...
> 
> Is there a neater way of generating html than this:
> 
>      print INDEX "    <tr>
>        <td style=\"vertical-align: top;\">$host<br>
>        </td>\n";

As Andy says, you should really look at using a templating system (I
know a _great_ book about the Template Toolkit). But if you want to
stick with print statements you should, at least, use more appropriate
quoting operators to avoid having to escape internal quotation marks.

print INDEX qq(    <tr>
  <td style="vertical-align: top;">$host<br>
  </td>\n);

Another option is to use here-docs.

print INDEX <<END_OF_HTML
<tr>
  <td style="vertical-align: top;">$host<br>
  </td>
END_OF_HTML


Dave...

From ed-pm at s5h.net  Thu Oct 18 13:31:00 2007
From: ed-pm at s5h.net (ed)
Date: Thu, 18 Oct 2007 21:31:00 +0100
Subject: [Thamesvalley-pm] generating HTML
In-Reply-To: <47178044.7090102@nerc.ac.uk>
References: <47178044.7090102@nerc.ac.uk>
Message-ID: <20071018213100.0d39ca94@workstation>

On Thu, 18 Oct 2007 16:48:20 +0100
Greg Matthews <gmatt at nerc.ac.uk> wrote:

> Well... I threatened this email so here goes...
> 
> Is there a neater way of generating html than this:
> 
>      print INDEX "    <tr>
>        <td style=\"vertical-align: top;\">$host<br>
>        </td>\n";
> ??
> 
> At the moment my code loops round generating table rows (of which the 
> above is the first column).
> 
> Given that the table is of an arbitrary size (depending on the number
> of hosts defined) am I approaching this the right way? ie directly
> printing out the raw html?
> 
> The generated page is essentially static, it is just generated by a
> cron job and the table contains links to pages of generated data for
> each host (if available).
> 
> I'm used to making my perl look neat and readable but now that it 
> contains huge dollops of html (such as the definition of my $header
> and my $footer) it looks a mess.
> 
> Let me know if anyone wants to see the whole thing (and have a good 
> laugh)...

As the others suggest, 

use templates;

but a simpler way might be to include a file from disk, and replace
bits using $var =~ s///g - if templates are too much work. I've not
personally been big on templates, but prefer to replace things in
files, just my way of being stuck in the mud I suppose.

-- 
The T1 to www.brokenmachine.com is rapping harder than Ice T because of
You have no chance to survive make your time. Tech Support is asleep.
:: http://www.s5h.net/ :: http://www.s5h.net/gpg
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mail.pm.org/pipermail/thamesvalley-pm/attachments/20071018/73e91425/attachment.bin 

From gmatt at nerc.ac.uk  Fri Oct 19 01:49:52 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Fri, 19 Oct 2007 09:49:52 +0100
Subject: [Thamesvalley-pm] generating HTML
In-Reply-To: <4717A24C.8010604@wardley.org>
References: <47178044.7090102@nerc.ac.uk> <4717A24C.8010604@wardley.org>
Message-ID: <47186FB0.9020108@nerc.ac.uk>

some great ideas, thanks.

Andy (and everyone else) suggested templates which having done a bit of 
digging looks to be an excellent suggestion (from a perl celebrity it 
seems!).

Dave showed me how to do quoting in a more sane way - thanks Dave, that 
was driving me nuts:

print INDEX qq(    <tr>
   <td style="vertical-align: top;">$host<br>
   </td>\n);

and also here documents - I started out using here docs but it plays 
havoc with readability. However, I might revisit them for assigning 
large chunks of html to variables if I dont go with templates.

ed suggested including a file from disk and substituting bits of it. 
This is kind of what I was thinking of and I'll have to consider it 
further but I think it may be too limiting if I cant be sure of the 
exact structure of my page in advance.

Lots to think about... thanks guys.

GREG

-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From Rafiq.Ismail at MorganStanley.com  Wed Oct 24 04:29:17 2007
From: Rafiq.Ismail at MorganStanley.com (Ismail, Rafiq (IT))
Date: Wed, 24 Oct 2007 12:29:17 +0100
Subject: [Thamesvalley-pm] generating HTML
In-Reply-To: <47186FB0.9020108@nerc.ac.uk>
References: <47178044.7090102@nerc.ac.uk> <4717A24C.8010604@wardley.org>
	<47186FB0.9020108@nerc.ac.uk>
Message-ID: <0F354FDFECA12647B75F7B154EC3828508798891@LNWEXMB58.msad.ms.com>

Greg,

> -----Original Message-----
> Sent: 19 October 2007 09:50
> To: ThamesValley Perl Mongers
> Subject: Re: [Thamesvalley-pm] generating HTML
> 
> some great ideas, thanks.
> 
> Andy (and everyone else) suggested templates which having 
> done a bit of digging looks to be an excellent suggestion 
> (from a perl celebrity it seems!).
> 
For another perspective.  I'm rather keen on HTML::Template these days,
which Andy also mentioned.  The templating language itself is very
restricted and simple.

yourtemplate.html:
...
<table>
<tr><td>host</td></tr>
<TMPL_LOOP NAME="HOSTS">
<tr><td><TMPL_VAR NAME="host"></td></tr>
</TMPL_LOOP>
</table>
...

Your script:
use HTML::Template;

my $template = HTML::Template->new(filename=>'yourtemplate.html');
my @templateData;

for my $host (@hosts) {
	push @templateData, { host=>$host }
}
 
$tempalte->param(HOSTS => \@templateData;

## print your http headers first and then

## Send rendered template to STDOUT
$template->output;


R.
--------------------------------------------------------

NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

From gmatt at nerc.ac.uk  Tue Oct 30 10:45:10 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Tue, 30 Oct 2007 17:45:10 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
Message-ID: <47276DA6.3030403@nerc.ac.uk>

Seems like I'm the only one to ask for advice here...

I'm using ssh to connect to a series of machines like so:

open( SSH, "$ssh_bin $ssh_opts $host $command |") or die...
while (<SSH>) {
	....
}
close SSH;

I'm opening a new ssh shell for each command.

I want to capture the output of various commands to a set of lists or 
files. For instance, the output of "uname -a" needs to be recorded in a 
file uname.out.

Is it possible to send a series of commands through the same SSH 
session? and capture the output for each separately?

That would save repetitive long timeouts when hosts are down.

So far I have figured out that I can open ssh like this:

open(SSH,"| /usr/bin/ssh $options $host");
print SSH "$command1\n";
print SSH "$command2\n";
close SSH;

but I haven't figured out how to capture the results of each command 
separately.

hope this makes sense and someone can point me the right direction

GREG

-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From ed-pm at s5h.net  Tue Oct 30 11:20:28 2007
From: ed-pm at s5h.net (ed)
Date: Tue, 30 Oct 2007 18:20:28 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
In-Reply-To: <47276DA6.3030403@nerc.ac.uk>
References: <47276DA6.3030403@nerc.ac.uk>
Message-ID: <20071030182028.3de17fd1@workstation>

On Tue, 30 Oct 2007 17:45:10 +0000
Greg Matthews <gmatt at nerc.ac.uk> wrote:

> Seems like I'm the only one to ask for advice here...
> 
> I'm using ssh to connect to a series of machines like so:
> 
> open( SSH, "$ssh_bin $ssh_opts $host $command |") or die...
> while (<SSH>) {
> 	....
> }
> close SSH;
> 
> I'm opening a new ssh shell for each command.
> 
> I want to capture the output of various commands to a set of lists or 
> files. For instance, the output of "uname -a" needs to be recorded in
> a file uname.out.
> 
> Is it possible to send a series of commands through the same SSH 
> session? and capture the output for each separately?
> 
> That would save repetitive long timeouts when hosts are down.
> 
> So far I have figured out that I can open ssh like this:
> 
> open(SSH,"| /usr/bin/ssh $options $host");
> print SSH "$command1\n";
> print SSH "$command2\n";
> close SSH;
> 
> but I haven't figured out how to capture the results of each command 
> separately.

It's not possible to do that, unfortunately.
 
> hope this makes sense and someone can point me the right direction

There is a nice library called IPC that includes Open2 and Open3. Open2
does what you would expect from open( "| dd |" ). Open3 includes a file
handle for stderr.

If you would like to communicate with ssh, and access stdin, stdout and
stderr, then open3 is the function for you.

  perldoc IPC::Open3

Will provide you with some useful information, using it isn't as
complicated as the examples look though, simply use:

  my $pid = open3( my $stdin, my $stdout, my $stderr, 
    "/usr/bin/ssh somehost" );

  while( <$stdout> ) {
    chomp;
    print( "Data from stdout: $_\n" );
  }

  kill( 9, $pid );

-- 
The 5 1/4 Floppy Drive to Alderan is reporting a faulty control module
because of Nutty drinking a bottle of peach schnapps. Covad is ignoring
the fact that they have one network admin left. ::
http://www.s5h.net/ :: http://www.s5h.net/gpg
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mail.pm.org/pipermail/thamesvalley-pm/attachments/20071030/9c0a5dc5/attachment.bin 

From adam.trickett at iredale.net  Tue Oct 30 11:36:18 2007
From: adam.trickett at iredale.net (Adam Trickett)
Date: Tue, 30 Oct 2007 18:36:18 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
In-Reply-To: <47276DA6.3030403@nerc.ac.uk>
References: <47276DA6.3030403@nerc.ac.uk>
Message-ID: <200710301836.18804.adam.trickett@iredale.net>

On Tuesday 30 Oct 2007, Greg Matthews wrote:
> Seems like I'm the only one to ask for advice here...
>
> I'm using ssh to connect to a series of machines like so:
>
> open( SSH, "$ssh_bin $ssh_opts $host $command |") or die...
> while (<SSH>) {
> 	....
> }
> close SSH;
>

Did you consider the Net::SSH::Perl or Net::SSH module, or do you really need 
to do it this way?


-- 
Adam Trickett
Overton, HANTS, UK

Capitalism is the extraordinary belief that the nastiest of men, for
the nastiest of reasons, will somehow work for the benefit of us all.
    -- John Maynard Keynes

From john at stumbles.org.uk  Tue Oct 30 12:15:36 2007
From: john at stumbles.org.uk (John Stumbles)
Date: Tue, 30 Oct 2007 19:15:36 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
In-Reply-To: <47276DA6.3030403@nerc.ac.uk>
References: <47276DA6.3030403@nerc.ac.uk>
Message-ID: <472782D8.7080106@stumbles.org.uk>

Greg Matthews wrote:

> hope this makes sense and someone can point me the right direction

Expect.pm?

-- 
John Stumbles

From stephenca at ls26.net  Wed Oct 31 05:07:57 2007
From: stephenca at ls26.net (Stephen Cardie)
Date: Wed, 31 Oct 2007 12:07:57 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
In-Reply-To: <47276DA6.3030403@nerc.ac.uk>
References: <47276DA6.3030403@nerc.ac.uk>
Message-ID: <4728701D.5080103@ls26.net>

Greg Matthews wrote:
> [cut] 
> hope this makes sense and someone can point me the right direction
> 

Greg,

You can do this with IPC::Run.  A trivial example is shown below.  The 
@hosts data structure contains a list of hosts to interrogate; each 
element of this is a reference to a hash, the keys of which are 
hostname, username and cmds.  The last is an array of filename and 
command pairs, the command will be executed on the remote host and the 
output recorded in the file. Obviously, this could be the same for every 
host, in which case you'd remove if from the hash ref and just re-use 
the array for each host (i.e. call it @cmds and remove the line 
'my(@cmds) = @{$host->{cmds}};' below.

This is quite Expect-like, and, as has been suggested, you could achieve 
the same thing with Expect.pm, if you are more comfortable with that 
environment.

### BEGINS ####
#!/usr/bin/perl
use strict;
use warnings;

use IO::File;
use IPC::Run qw( start pump finish timeout );

my(@hosts) = (
     { hostname => 'myhost.mydomain.com',
       username => 'myusername',
       cmds     => [
           # Filename      Command
         [ 'uname_a.txt', "uname -a\n"],
         [ 'uname_s.txt', "uname -s\n"],
       ],
     },
);

my($ssh_bin) = '/usr/bin/ssh';
my($ssh_opts) = '-2';

for my $host (@hosts) {
     my(@login) =
         ($ssh_bin, $ssh_opts, join '@' => @{$host}{qw(username hostname)});
     my(@cmds) = @{$host->{cmds}};

     my($h,$in,$out,$err);
     $h = start( \@login, \$in, \$out, \$err, timeout( 3 ) );

     $in .= "\n";

     while (@cmds) {
         my($file,$cmd) = @{(shift @cmds)};
         $in .= $cmd;

         pump $h until ( $out=~/\n/ );
         my($fh) = IO::File->new($file, 'w');
         defined( $fh ) or die "Can\'t write to $file:$!";
         $fh->print( $out );
         $fh->close;
         $out = '';
     }

     $h->finish;
}

exit;
#### ENDS #####





From gmatt at nerc.ac.uk  Wed Oct 31 05:11:16 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Wed, 31 Oct 2007 12:11:16 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
In-Reply-To: <200710301836.18804.adam.trickett@iredale.net>
References: <47276DA6.3030403@nerc.ac.uk>
	<200710301836.18804.adam.trickett@iredale.net>
Message-ID: <472870E4.1000104@nerc.ac.uk>

Adam Trickett wrote:
> On Tuesday 30 Oct 2007, Greg Matthews wrote:
>> Seems like I'm the only one to ask for advice here...
>>
>> I'm using ssh to connect to a series of machines like so:
>>
>> open( SSH, "$ssh_bin $ssh_opts $host $command |") or die...
>> while (<SSH>) {
>> 	....
>> }
>> close SSH;
>>
> 
> Did you consider the Net::SSH::Perl or Net::SSH module, or do you really need 
> to do it this way?

yes, I looked at these initially, my main concern is that I can't 
guarantee that these modules will be available. IIRC at least one of 
these will do what I want.

For now I think I'll live with the ssh timeouts - this is run from cron 
so its not /my/ time!

thanks all

GREG

-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From gmatt at nerc.ac.uk  Wed Oct 31 05:19:38 2007
From: gmatt at nerc.ac.uk (Greg Matthews)
Date: Wed, 31 Oct 2007 12:19:38 +0000
Subject: [Thamesvalley-pm] using ssh within perl scripts
In-Reply-To: <4728701D.5080103@ls26.net>
References: <47276DA6.3030403@nerc.ac.uk> <4728701D.5080103@ls26.net>
Message-ID: <472872DA.10704@nerc.ac.uk>

Stephen Cardie wrote:
> You can do this with IPC::Run.  A trivial example is shown below.  The 
> @hosts data structure contains a list of hosts to interrogate; each 
> element of this is a reference to a hash, the keys of which are 
> hostname, username and cmds.  The last is an array of filename and 
> command pairs, the command will be executed on the remote host and the 
> output recorded in the file. Obviously, this could be the same for every 
> host, in which case you'd remove if from the hash ref and just re-use 
> the array for each host (i.e. call it @cmds and remove the line 
> 'my(@cmds) = @{$host->{cmds}};' below.

this looks quite useful, I shall squirrel it away for future reference. 
Unfortunately, I can't guarantee the availability of IPC::Run. I shall 
have a close look at this module tho as it seems quite powerful.

thanks Stephen

GREG

> 
> This is quite Expect-like, and, as has been suggested, you could achieve 
> the same thing with Expect.pm, if you are more comfortable with that 
> environment.
> 
> ### BEGINS ####
> #!/usr/bin/perl
> use strict;
> use warnings;
> 
> use IO::File;
> use IPC::Run qw( start pump finish timeout );
> 
> my(@hosts) = (
>      { hostname => 'myhost.mydomain.com',
>        username => 'myusername',
>        cmds     => [
>            # Filename      Command
>          [ 'uname_a.txt', "uname -a\n"],
>          [ 'uname_s.txt', "uname -s\n"],
>        ],
>      },
> );
> 
> my($ssh_bin) = '/usr/bin/ssh';
> my($ssh_opts) = '-2';
> 
> for my $host (@hosts) {
>      my(@login) =
>          ($ssh_bin, $ssh_opts, join '@' => @{$host}{qw(username hostname)});
>      my(@cmds) = @{$host->{cmds}};
> 
>      my($h,$in,$out,$err);
>      $h = start( \@login, \$in, \$out, \$err, timeout( 3 ) );
> 
>      $in .= "\n";
> 
>      while (@cmds) {
>          my($file,$cmd) = @{(shift @cmds)};
>          $in .= $cmd;
> 
>          pump $h until ( $out=~/\n/ );
>          my($fh) = IO::File->new($file, 'w');
>          defined( $fh ) or die "Can\'t write to $file:$!";
>          $fh->print( $out );
>          $fh->close;
>          $out = '';
>      }
> 
>      $h->finish;
> }
> 
> exit;
> #### ENDS #####
> 
> 
> 
> 
> _______________________________________________
> Thamesvalley-pm mailing list
> Thamesvalley-pm at pm.org
> http://mail.pm.org/mailman/listinfo/thamesvalley-pm
> 


-- 
Greg Matthews           01491 692445
Head of UNIX/Linux, iTSS Wallingford

-- 
This message (and any attachments) is for the recipient only. NERC
is subject to the Freedom of Information Act 2000 and the contents
of this email and any reply you make may be disclosed by NERC unless
it is exempt from release under the Act. Any material supplied to
NERC may be stored in an electronic records management system.


From user_groups at pm.org  Wed Oct 24 10:13:11 2007
From: user_groups at pm.org (Dave Cross)
Date: Wed, 24 Oct 2007 17:13:11 -0000
Subject: [Thamesvalley-pm] generating HTML
In-Reply-To: <0F354FDFECA12647B75F7B154EC3828508798891@LNWEXMB58.msad.ms.com>
References: <47178044.7090102@nerc.ac.uk>
	<4717A24C.8010604@wardley.org>	<47186FB0.9020108@nerc.ac.uk>
	<0F354FDFECA12647B75F7B154EC3828508798891@LNWEXMB58.msad.ms.com>
Message-ID: <471F7D08.5000601@pm.org>

Ismail, Rafiq (IT) wrote:

> For another perspective.  I'm rather keen on HTML::Template these days,
> which Andy also mentioned.  The templating language itself is very
> restricted and simple.

That's all very well. But does it have a book with a badger on the cover?

Dave....