Re: three bugs and a wishes

[Mark . Burgess]

Sorry for the delay.

On 16 May, Uwe Zeisberger wrote:
> hello,
> 
> I have 2 bugs and 2 wishes concerning cfengine. I use version 2.0.5 and
> found these things not corrected reading the changelog. So I hope my
> report is usefull for you...
> 
> first the bugs, as they are more important:
> 
> 1)
> if I want to check /var (which is a locally mounted partition) for free
> space, ie:
> 
> required:
>     /var
>         freespace=200mb
>         inform=true


This is documented and requires a force option for crossing filesystems.


> I never get informed (if not force=true), because /var and /var/.. are
> on different file systems (see mount.c +144). If that's a feature - but
> I don't think so - it's at least not documented. There is only a hint
> concerning NFS mounted file systems.
> 
> 2)
> Searching for the reason of bug 1, I saw a unsensible check at 
> mount.c +112:
> 
> strncmp(host,VFQNAME,strlen(host)) == 0
> 
> Since strncmp("host", "hosts", strlen("host")) == 0, using
> 
> strncmp(host,VFQNAME,strlen(host) + 1) == 0
> 
> is better. Well, comparing two fqdn, it's quite unlikely that it make
> any difference, but maybe this construct is used anywhere else...
> 
> 3)
> I'm not able to negate an evaluated class. for example:
> 
> groups:
>     missing_var = ( !IsDir(/var) )
> 
> this results in: 'Wildcards cannot be used in this context'.


Fixed in 2.1.1

 
> Now to my wish:
> It is imho cumbersomely to check on which hosts cfengine did something.
> By now I use Inform = ( on ). Changing the config results in approx
> 40 mails. When (or if) I realize that there is a mail less than there
> have to be, I spend much time comparing these mails with a hostlist.
> Sadly I have no concrete proposition, but maybe you have.

Will think about this.
Thanks for your message
Mark


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Work: +47 22453272            Email:  address@hidden
Fax : +47 22453205            WWW  :  http://www.iu.hio.no/~mark
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~