extension of lsb packages

Erik Troan ewt at redhat.com
Wed Mar 6 06:46:23 PST 2002


On Tue, 05 Mar 2002 12:54:08 -0600
"George Kraft IV" <gk4 at austin.ibm.com> wrote:

> 2) changing the rpm file format to determine if it is an LSB package is a
> design issue.  I *restarted* the lsb packaging taskforce  a year ago.  That

I'm not sure what this has to do with adding a byte to an RPM to distinguish LSB packages from non-LSB ones in a way that old versions of RPM and alien would completely ignore.

Afaict, the issue with this is folks want an easy way for scripts and such to determine if a package is intended for LSB systems. Right now that information will be encoded in dependencies, which is fine but require something like RPM to be on the system in order to parse them. Adding a byte to a reserved area that all old versions of RPM/alien would ignore hardly seems like a major design issue, esp when one of the people who needs convincing (me) doesn't mind at all. That byte would be used by file(1); RPM/alien will get absolutely no semantic information from it.

It doesn't resolve *naming* issues at all. *Suggesting* LSB packages be named .lsb is certainly reasonable, though I think the real world implications are almost nil. Having packages that are part of the lsb begin with lsb- is a good idea. As for having all packages that are lsb compliant being in lsb-, I really don't see the point. If this effort is successful to any reasonable degree, all 3rd party packages should end up as LSB packages, so the lsb- is a bit redundant. If that doesn't happen, the LSB hasn't been successful, so who cares?

Erik

-----------------------------------------------------------------------------
|               "Distracting? I'm inspiring"                                |
|                              - Sally Bowles in Cabaret                    |




More information about the lsb-discuss mailing list