On 6/12/06, a thing <[EMAIL PROTECTED]> wrote:
Declan Naughton wrote
> 1) In apt, if you update a package you must download a whole new
> package. In conary, you just download the changes. (I guess it could
> be said that apt operates with packages (RPM, DEB), while conary
> operates with "changesets". A changeset can be a whole new package, an
> update for a package, or they can instruct to remove a package.)
That sounds like it would be annoying to keep all those changesets in a repo,
like from 1.0 to 1.1, from 1.1 to 1.2, from 1.0 to 1.2, and on. Or are you
designing so if 1.0 is upgraded and 1.2 is the latest version, it downloads the
changeset to get to 1.1, then the changeset to get from 1.1 to 1.2 is
downloaded? Either way, it ends up cluttering the repository with tons of
changesets for ancient versions.
I've only briefly looked at how to produce packages for conary. I know
that you don't build a DEB or an RPM - you don't get a single file
that you can redistribute. You check the package in with the
repository immediatly.
The repository software, which is usually run through apache with
mod_python, will sort out changesets - generating + caching them etc.
--
Declan Naughton
_______________________________________________
gnu-system-discuss mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/gnu-system-discuss
