Branko Vukelic schrieb:
> So, the question is... When do you think it is okay to go over the
> 'edge'? How often does it happen that a dev decides to use stable
> development snapshots, and why? Should there be a more strict rule
> like allowing absolutely NO non-stable versions in extra, and just put
> everything non-stable in unstable?

Just an example here: When gaim (now pidgin) was near the 2.0 stage, the
1.X version was unmaintained. According to Jan, it needed like 20
patches to even compile and fix all the critical bugs. He switched to
gaim 2.0 beta packages, that were much more stable, but without any
patches. In such cases, I would say it is okay to move to development
versions.

If there is no stable version for a package yet, but it is widely used,
then using a development version is okay as well (Example: I added the
iwl4965 driver when it was in a very early stage, because people needed it).

In some rare cases, we need features of a development version as a
dependency. Such cases are decided by the responsible developer,
depending on the stability of the development version (this happened
sometimes with wireless_tools).

In case of gcc and glibc, we sometimes use snapshots that are not yet
released, but Jan should answer the why and when questions here.


Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
arch mailing list
[email protected]
http://archlinux.org/mailman/listinfo/arch

Reply via email to