Re: New team member
On Thu, Jul 24, 2014 at 12:56 PM, Cédric Boutillier <boutil@debian.org> wrote:
> Hi Caitlin,
>
> On Thu, Jul 24, 2014 at 01:17:35AM -0400, Caitlin Matos wrote:
>> Hi Cedric,
>
>> On 23/07/14 05:26 AM, Cédric Boutillier wrote:
>> >
>> >My point of view is the following: if you just prepare a one-time upload
>> >to fix a bug or to upgrade to a newer version without much change, then
>> >you can make a "Team upload". If you add these words to the changelog
>> >entry, then there is no need to add yourself to the Uploaders field in
>> >debian/control.
>> >
>> >If instead, you found a package on which you spent some time, and you
>> >particularly like to maintain, then add your name to the Uploaders
>> >field.
>
>> Okay, great to know.
>
>> The main problem I foresee is if I start working on a package locally, and
>> it ends up being quite a bit of work, and after some time, I discover
>> someone else has now uploaded it. Or, potentially worse, I work on and
>> upload something that someone's already been working at for several
>> days/weeks. I'm trying to figure out what kind of co-ordination there is in
>> that respect. I know PET will tell you about repos with a new version ready
>> or where there appears to be some work in progress, but if someone keeps it
>> all local until the end, you can't tell!
A good rule of thumb is to not keep changes local for long. Push as soon
as possible.
I have had one, or maybe two, race conditions like this. (Mostly because I have
left the work almost done and then not finished.) I didn't mind that someone
else finished the work though, it was quite welcome.
Usually I chat on IRC if I am doing work that spans longer time. If there are
more complex situations I bring it up on the list.
>> As a general rule, if there's an available update or fairly new bug for a
>> package with the uploader listed as someone who seems to be quite active on
>> these boards (e.g., Jonas Gennant, Hleb Valoshka), I'm guessing I should
>> steer clear of those. And for ones where there's not been a new Debian
>> version for years, those sound safe. But, of course, the only way to find
>> those is to browse through them all manually (even after filtering for just
>> those with new versions upstream, there's a lot of them). And then there's
>> some more complicated situations....
I don't think you should stay clear of anything if you would like to work on it.
If in doubt, ask the active people to join the effort!
>> Here's a real example I just ran into. Sticking with those rules, I found a
>> package (ruby-http) that met all of the above, AND has an RC bug, AND whose
>> uploader seems to have disappeared from Debian. I've already done quite a
>> bit of work now on this and it looks ready to upload. I then noticed that
>> Antonio Terceiro and Christian Hofstaedtler have done some updates on the
>> package on several occasions in the past couple of years. Should I therefore
>> have left it up to them? Is someone going to feel insulted if I upload it?
>> Or relieved? (I know that depends on the person, I'm just musing aloud a bit
>> here).
>
>> I know my questions are probably a bit annoying and long-winded and whiny.
>> Believe me, this is actually the shortened and revised version of my e-mail.
>> I'm sorry! I'm just new to this, and don't want to get someone upset with
>> me!
>
> I understand your concerns, as I had the same a few years ago.
It is sad that it is like this but I think most new contributors (and even
experienced ones) feel this. (Asheesh Laroia have had several talks
and workshops on DebConf about this subject.)
Just dig in, and, again, if you want certainty, ask around. It is also good to
socialize and get to learn the team in general, it reduces doubts like this
and makes the work more fun IMHO.
In Debian in general, Non-maintainer uploads (NMU for short) is an
example when we help eachother out. This is a good thing! We are in
this together and cooperate to build Debian. Help should be welcomed!
--
Per
Reply to: