[Pkg-puppet-devel] Bug#493452: Downcasing of facts

[Micah Anderson]

tag 493452 + forwarded http://reductivelabs.com/redmine/issues/show/1515
thanks

The discussion related to this bug seems to have wandered off of the
BTS itself, resuming on the pkg-puppet-devel list, I've summarized the
discussion thus far below to return it. 

I've also filed this bug upstream[0].

micah

0. http://reductivelabs.com/redmine/issues/show/1515

micah writes:
> > What if you endorse capitalism and change your selector to:
> > Package {
> > > >      provider => $operatingsystem ? {
> > > >        Debian => aptitude,
> > >> >      },                                           
>> }
> > Does it find it then? Since facter is reporting:
> > > lapse:~|master|% facter | grep operat                                                                             
> > > operatingsystem => Debian                                                                          

madduck writes:
> It seems so, yes. As Russ says though, this is a regression...

mpalmer writes:
> The contents of facts stopped getting downcased with 0.22.0, and was made
> configurable in 0.22.1.  This option, 'downcasefacts', was defaulted to
> false in 0.22.1, and is defaulted similarly in 0.24.4 and 0.24.5.  I would
> look to local config files for changes.

rra writes:
> > The contents of facts stopped getting downcased with 0.22.0, and was made
> > configurable in 0.22.1.  This option, 'downcasefacts', was defaulted to
> > false in 0.22.1, and is defaulted similarly in 0.24.4 and 0.24.5.  I would
> > look to local config files for changes.

> The change wouldn't be downcasing of facts, but rather case-sensitive
> matching of fact values in case and selector statements.  We see Debian
> when we interpolate the $operatingsystem fact, but a case statement
> matching on debian still works in 0.24.4.


mpalmer writes:
On Tue, Aug 05, 2008 at 12:03:28AM -0700, Russ Allbery wrote:
> > Matthew Palmer <mpalmer at debian.org> writes:
> > 
> > > The contents of facts stopped getting downcased with 0.22.0, and was made
> > > configurable in 0.22.1.  This option, 'downcasefacts', was defaulted to
> > > false in 0.22.1, and is defaulted similarly in 0.24.4 and 0.24.5.  I would
> > > look to local config files for changes.
> > 
> > The change wouldn't be downcasing of facts, but rather case-sensitive
> > matching of fact values in case and selector statements.  We see Debian
> > when we interpolate the $operatingsystem fact, but a case statement
> > matching on debian still works in 0.24.4.

> > Hmmm, well then, that would be handled by the 'casesensitive' config
> > parameter, but again, that config parameter default hasn't changed upstream
> > between 0.24.4 and 0.24.5, and there's no changes in the code that uses that
> > config parameter.

> > Martin, what does puppetmasterd --genconfig say about the value of
> > casesensitive on your 0.24.5 system?

> I assume you meant puppetd, since puppetmasterd is still running on
> a 0.24.4 system.

> puppetd on 0.24.5 says, with the comment characters:

> # casesensitive = false
> # downcasefacts = false
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-puppet-devel/attachments/20080816/b2c159d5/attachment.pgp