Re: absolute path errors in 2.1.16

[Mark Burgess]

Patch for this problem now uploaded to svn.

Mark

On Wed, 2005-10-05 at 09:53 -0500, Chip Seraphine wrote:
> Mark Burgess wrote:
> 
> >You both need to give decent error reports. I don't have any trouble
> >with this. Give me an example. Nothing much has changed in the code, so
> >it doesn't make a lot of sense to me.
> >
> >M
> >
> Sorry, I was working on getting that together, but I wanted to go ahead 
> and get the word out asap.
> 
> Basically, in places where I did this:
> 
>     control:
>        foo= ( /path/to/bin )
>        bar= ( /path/to/data )
>        result= ( ExecResult(${foo}/command ${bar}/file) )
> 
> ... resulted in things not expanding to "/path/to/bin/command 
> /path/to/file".  I have not actually tried the example above, although I 
> will once I get some fires put out. 
> 
> Anyway, here's some -I output showing the funky expansions:
> 
> monaco: Executing script /var/cfengine/bin/update-smf-manifest clearingd 
> cfengine/clearingd.smf.xml              
> cfengine/log/smf-update.last...(timeout=0,uid=-1,gid=-1)
> monaco:f-manifest clea: /var/cfengine/bin/update-smf-manifest: 
> cfengine/log/smf-update.last: cannot create
> monaco: Finished script /var/cfengine/bin/update-smf-manifest clearingd 
> cfengine/clearingd.smf.xml               cfengine/log/smf-update.last
> 
> 
> Reinstalling 2.1.15 with no changes whatsoever to the configs or 
> anything else, I got this:
> 
> monaco: Executing script /var/cfengine/bin/update-smf-manifest clearingd 
> /opt/tradelink/clearingd/etc/clearingd.smf.xml                  
> /var/opt/clearingd/log/smf-update.last...(timeout=0,uid=-1,gid=-1)
> monaco: Finished script /var/cfengine/bin/update-smf-manifest clearingd 
> /opt/tradelink/clearingd/etc/clearingd.smf.xml           
> /var/opt/clearingd/log/smf-update.last
> 
> 
> As you can see, the "/opt/tradelink/clearingd/etc/clearingd.smf.xml" 
> somehow expanded as "cfengine/clearingd.smf.xml", and 
> "/var/opt/clearingd/log/smf-update.last" became 
> "cfengine/log/smf-update.last".
> 
> I have not yet done proper debug runs; I'm tied up on damage control 
> right now (this broke a lot of my stuff).  I saw nothing like this on my 
> test boxes, although I can't say whether that was because I wasn't 
> looking carefully enough or because it didn't happen on them.
> 
>