Yum and RPM with FC2

[Nifty Hat Mitch]

On Sun, Jul 18, 2004 at 01:33:18PM +0100, Timothy Murphy wrote:
> Tim and Alison wrote:
> 
...
> > It has downloaded all the new RPM files but I am getting the following
> > error when it tries to install the packages
> > 
> >  error: rpmdbNextIterator: skipping h#     286 Header V3 DSA signature:
> > BAD, key ID 4f2a6fd2
> 
> I've no idea how to solve your problem,
> but it reminds me how bad the error messages in yum are -
> if it works (as it does 90% of the time) it is a marvellous resource.
> But if it fails, you can be sure the error message will be unintelligible.
> 
> This is a general fault with Linux, I think
> (and to a lesser extent with Windows) -
> few developers seem to consider when penning their error messages
> what is likely to help someone who encounters the error.
> Frequently the error message seems to be a message to the developer
> rather than to anyone else.
> 
> I think all error messages should try to advise the reader
> what action he might take, if that is at all possible.

I almost agree.

IMO....
When an error message requires a magic decoder ring that only the
author has then the author is asking you to file a bug.

In one view of things there are two classes of errors.  User errors
and application internal errors.  The common user errors should
generate messages that a user can decode and fix his use of the
application.   Internal application errors are another pile of
fish.  

The common issue of user input errors commonly generates input
checkers and parsers.  Consider applications like named-checkconf,
named-checkzone, pwck, grpck, testprns, sgmlcheck, mbchk....

In this case "286 Header V3 DSA signature: BAD, key ID 4f2a6fd2" tells
me that the file is corrupt (bad download) or you do not have the key
for this package registered/ loaded.  See the rpm man page and
archives for stuff along this line:

 rpm {-K|--checksig} [--nosignature] [--nodigest]  PACKAGE_FILE ...
 rpm --import PUBKEY ...

For the installation of a major update expect some mismatch in the
keys.  Also if you include a variety of distribution packages/
packagers there is a need to import their digital keys.





-- 
	T o m  M i t c h e l l 
	/dev/dull where insight begins.