[arch-haskell] State of Affairs: Summarizing 83 days worth of experience

[Tasha Buckley]

On Sat, Jan 8, 2011 at 4:03 AM, Peter Simons <simons at cryp.to> wrote:

> Anyway, this issue is beside the point I was trying to make. I am concerned
> about the lack of communication regarding the maintenance of [extra] and
> [community]. Whatever is going on in these repositories, the members of this
> list are not involved in the process even though we are directly affected by
> changes that occur in these places. This is not an ideal situation, and it
> already *did* cause a significant amount of trouble in the past.

Well, who on this list is even an Arch dev with access to [extra]?
That would be Vesa and Rémy, who is now the primary maintainer I
believe. Rémy is very active on the list, so I don't think it's quite
accurate to characterize the problem as one of disconnectedness.
Rather, one and a half people are responsible for maintaining a set of
packages with sometimes cascading rebuilds required without, according
to you, a more complete automatic way to check for consistency with
the rest of the packages that we want to build against Haskell
[extra]. As an aside: in the past few months, I have not really seen
any issue take more than a couple of weeks to resolve with the average
response time seeming more like a few days. So I am one person who
thinks this aspect has noticeably improved (though obviously still not
where you would like it to be).

To summarize my main point, I would actually characterize this problem
as a lack of people with access to [extra] who care about the Haskell
packages or have the free time/resources to get a turnaround on issues
consistently to a few days at most. With better tools, having only two
devs would clearly be less of a blocker, but that's another part of
your post.

As far as the TUs who have been helping with [community], Xyne is
always helpful and communicative on this list and the forums. While
I'm not sure if Sergej has posted to the list, his ~1400 packages are
very close to up-to-date and he's very responsive about updating when
flagged out of date or making sure his haskell packages track HP
(alex) or are updated.

Please don't take this as me trying to dismiss your concerns. I am
just trying to offer a counterpoint on this particular aspect as it
seems an end user's perspective could be helpful too.