[pca] operands issue + little feature request proposal

Martin Paul martin at par.univie.ac.at
Wed Apr 6 14:51:55 CEST 2011


Hi Daniel,

Sorry for the delay - I took advantage of the nice wheather and spent a few days 
hiking in Austria's Waldviertel. Very nice, BTW :)

> i have found, that allrs list is bigger then default, which should be
> just *all (if i understand correctly to --man), so i was little bit
> shocked and suspecting if package selection algorithm is correct ..

The default operand when running pca without an argument is "missing". This 
group consists of all patches which apply to a system and which are not 
installed in their most current revision.

The "all" group on the other hand includes all patches which apply to a system, 
even if they are already installed. This is great for documentation: Keep a list 
of "pca -l all" from all of your systems and use grep to find out if a certain 
patch is installed on all of them (to which it applies). You wouldn't want to 
install "all" on a system, therefore.

The r/s/rs postfix is used to reduce any of the groups (missing, all, etc.) to 
those carrying the Recommended and/or Security flag. I think you just 
misunderstood the difference between the default (missing) and all. The sample 
outputs you included look fine.

> i'm also suggesting with this little proposal: please kick out patch
> unzipping phase while -I ..

It's there for a reason: PCA needs to read the patch README and the patchinfo 
file even when only pretending to install a patch, both to determine whether a 
reboot is required (this information is missing from the xref file, 
unfortunately) and to get the list of included files for the "--safe" option. 
Plus, it also tests whether the patch archive can be expanded successfully. Even 
though this could be done with the "-t" option to unzip, this isn't really much 
faster than extracting the archive to /tmp, and the is no "test" option which 
could be used for tar archives (which PCA still supports for old patches).

So theoretically I could work around the extraction when using "--pretend", but 
it needs changes in more than one place, and is probably not worth the effort.

Martin.

> regards and thanks martin for pca,
> ave daniel
> 
> 


-- 
                       Martin Paul \ System Administrator
               University of Vienna \ http://www.par.univie.ac.at/
         Faculty of Computer Science \ Nordbergstrasse 15/C/3, A-1090 Wien
  Research Group Scientific Computing \ Tel +43 1 427739403 Fax +43 1 42779394



More information about the pca mailing list