[Nagiosplug-devel] Plugin guideline suggestion

[Jason Martin]

On Wed, Jan 19, 2005 at 08:28:11AM -0000, Voon, Ton wrote:
> Maybe I haven't asked the right question, so let me try again: what are you
> trying to solve with these version identifiers?

My particular need is that I have >500 servers with plugins on
them and am looking to design an elegant way of surveying  /
tracking their versions. Of course I'll have to upgrade if this
change goes in to use the new functionality but I can handle
that.

Another thought I had about managing this would be to have each
plugin return its version in the performance data, but I realize
that doing so is a much larger change to the standard
non-standard and would interfere with existing perfdata tools.
Hence the less impactful store-it-in-the-binary method.

Additional reasons / benefits:
* CVS files should all (IMHO) already have a $Header$ or $Id$
string in their text in at least a comment. It can also be used
in the --version data.  This mechanism exposes that header in
the final product.
* The tools to get this data are standard across platforms
* It is a pretty standard method of file identifications which
integrates easily with 'software inventory' tools.
* It does not depend on the plugin developer implementing robust
argument processing. 
* The plugin does not have to be functional to get the data (ie
on a central depot that stores plugins for all platforms)
* Implementing it does not require linking to any external
libraries (perl's GetOpt::Long and such). This is important
since adding them to existing code requires more testing than
defining a variable.

> by all compilers" which immediately makes me reticent.
The #ident C pragma was just an idea, there are very easy
compiler-generic ways of embedding that data. I have a document
somewhere that describes how to do it in many common languages
that I can provide if we go this route.

> But I don't want to introduce "another way" for the sake of it. Unless you
> have a very good reason.
I understand; if the above isn't convincing then I'll work out
some other way to manage the data and perhaps start submitting
some --version patches.

Thanks,
-Jason Martin
-- 
Smoking is a leading cause of statistics.
This message is PGP/MIME signed.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 211 bytes
Desc: not available
URL: <https://www.monitoring-plugins.org/archive/devel/attachments/20050119/4829c462/attachment.sig>