[Rpm-metadata] metadata list and discussion over tapas tonight

Adrian Likins alikins at redhat.com
Tue Aug 5 21:56:01 UTC 2003


On Tue, Aug 05, 2003 at 08:11:48AM -0400, Peter Bowen wrote:
> > distribution information: 
> >    which repositories/channels exist where and when they were last
> > updated, etc - useful for giving a place to look for the other files and
> > seeing what's changed, quickly.
> 
> Good, this matches somewhat with Red Carpet's channels.xml file.
>
	It's based on apts "Release", rc's "channels.xml"
(or was it distributions.xml) and up2dates "login"

	It would include basic info (name, description, etc)
plus some sort of id for the rest of the metadata. Maybe
a version number of each file (for cache validation). Checksums
of the other files would be useful as well. Especially if
its decided to gpg sign this file (presumably with a detached
sig paranoid clients can verify). 

	This file should be kept fairly svelte though, since
it's probabaly going to be downloaded everything (ignoring
If-Last-Modified/ETags style "caching").
	 
> > package information not including file lists: 
> >    all the misc information considered important per package (to be
> > decided in terms of entries but it must include byte-ranges for the
> > headers per package)
> 
> Again, seems to jive with what Red Carpet has now.
> 
> > Should include almost complete file lists for
> > certain packages that can be 'whitelisted' in b/c they end up being used
> > in a file dependency in a large percentage of the time.
> 
> So common file deps, or those that the server knows the client will
> need, are included here?
> 
	Being lazy, I'd say just included everything. With
some client side smarts, you rarely need to update this file.
And for "update" channels, it probabaly doesnt change that
often. 

> > package filelists: complete file listings per package
> 
> One file per package on the server?
> 
	one file for the channel. Along the lines of
deb apt-get's "Contents.gz" file.

> > obsoletes: listing of obsoletes tags - this could be alternatively
> > included in package information not including file lists, depends on
> > some more discussion - nothing too crazy.
> 
> I would like to see this in the package information file.
> 
	That could be doable, but obsoletes info changes slowly,
while package info changes alot. So if they were combined, you
repeatedly download the same obsoletes info. 

> > - send in your list of things that should appear in the non-file
> > packages listing. What data from the header is critical/useful/desired.
> 
> I think that the absolute minimum list of things that need to be in the
> file are Name, Epoch, Version, Release, Arch, package file size, and
> absolute/relative path to the package file.
> 
> Looking at the packageinfo.xml that Red Carpet uses now, the main things
> that I see that we would want to are possibly section, pretty name,
> install only, importance, license, plus some internal information.
> 
> Section and pretty name are simply for display purposes, section roughly
> mapping to group, and pretty name allowing the client to display a more
> friendly name for a package.
> 
> Importance is an indication of how important the update/errata is to the
> end user.  We use minor, feature, suggested, urgent, and necessary.  The
> middle three basically map to Enhancement, Bug Fix, and Security
> erratas.
> 
> Install only should be a boolean and indicate that the package should be
> installed with the INSTALL_INSTALL flag, rather than the INSTALL_UPGRADE
> flag.
> 
> Finally, Red Carpet has the ability to require a user to accept a
> license prior to installing certain software.  The license field should
> have the name of the license they must agree to in the license field.
> 
	I tend to think these should live in a seperate file. You
can join it with the packagelist file to get the complete dataset.
This allows a few things. The first being that clients that dont
support this can simply ignore it, and not have to download it. 
The other is the same local cache invalidation problem already
mentioned. This info might not change on every repo update, so
could save some bandwidth having it seperate. 


> These are the main things that I see that Red Carpet uses that are not
> directly found in RPM headers.  We also have a couple of internal use
> fields that are essentially opaque data to the client, plus some fields
> that are duplicated from the rpm headers.  As long as we either allow
> for extra fields that some clients may not know about, or allow for XML
> Namespaces, extra data in the XML should not be problematic.
> 
	Namespaces etc would work for not breaking clients, but
it still means downloading data that a given client doesnt care
about. 

	But as I already mentioned, saving bandwidth is the
primary goal of my proposal, and if thats not actually 
important, other ideas will work. 


Adrian


More information about the Rpm-metadata mailing list