[arch-dev-public] starting the repo cleanup
belanger at ASTRO.UMontreal.CA
Fri Oct 26 20:51:52 EDT 2007
On Fri, 26 Oct 2007, Dan McGee wrote:
> On 10/26/07, Andreas Radke <a.radke at arcor.de> wrote:
>> I will start removing packages now. It seems nobody else has the time
>> or interest to do it. I will use Eric's list in our dev wiki.
>> step 1: orphaned packages that are no (make)depends for any other pkg.
>> I wonder about Eric's comment
>> "Package | Reason to remove it. No reason implies we keep it"
>> I think it should be the opposite. If any developer wants to keep it he
>> should adopt it before it will be removed. If we won't do it that way
>> we will stay for months for comments what the reasons are why we want
>> to remove them. The reason is the poor (non existing) maintainership
>> and low usage according the public wiki list.
> Don't change the rules on your own just yet- for now, I'm sure there
> are plenty of removals that can proceed that have been commented on.
> However, I do agree that the list will probably just sit there if we
> can't touch a package that hasn't been commented on.
> **NOTE to devs- take a quick look at this list and comment on the
> packages you know can be purged from extra!**
As I said to Andy on jabber, I went through that list last night and will
continue tonight. I'll sent the current list of removal canditate with
their reason on this ML for discussion/approval. Also I have only checked
for make/dependencies for extra packages. They might be depends for
unstable, community or optional depends through post-install message.
This should be checked before doing the removal. If a package is a depends
for a community package, it should remain in extra until its added in
community by a dev or TU as we don't want to break depends.
>> Is it ok to rm the whole cvs files including history and the package
>> itself? I see no important reason why we should put the cvs history and
>> pkg somewhere else. If any user or TU wants to pick it up again they
>> would have to do it from scratch again. But we can also hold the cvs
>> entries somewhere for backup for a while. I would need permission on
>> that directory. Your choice.
> -1 here. Why on earth would we have to do that? Version control is
> used for a reason, and rewriting history is rarely a smart thing to
> do. If empty directories in your checkouts are an issue, try using a
> .cvsrc file:
> $ cat .cvsrc
> cvs -q
> diff -uN
> rdiff -u
> checkout -P
> update -dP
The removed packages should be put in unsupported; the only exception
being packages that are removed because the sources have disappeared. The
reason is that a lot of work has been put in these packages. We'll
probably remove some packages with non-trivial PKGBUILD with accompaning
files such as patches, .install file, scripts, etc. It only makes sense
to let the new maintainer benefit from this work.
I am ready to do all this and will work intensily on the cleanuip over
next week. It's only a question of going through the list to make sure we
don't remove a package that should remain in extra.
Another point that hasn't been mentionned yet, after the cleanup/adoption,
we'll need to do the same with the bug tracker. A lot of packages have
changed maintainers so bug reports will need to be reassigned to the new
maintainer. Bugs for removed packages will need to be closed with the
links posted on the packages page in AUR to let any potential maintainer
aware about them. We can worry about that later after the cleanup though.
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
More information about the arch-dev-public