← Back to team overview

graphite-dev team mailing list archive

Re: branch merges, docs, collaboration

 

Doh... I keep thinking I've figured out bzr, but it keeps surprising me.
Apparently all these merges got committed in my local checkout but not
pushed to trunk somehow (I thought automatic pushing was the whole point of
a checkout but what do I know). I won't be able to fix this until later
tonight. Sorry for the inconvenience, I'll send another note when everything
is actually in trunk and testable.

-Chris

On Sun, Jul 10, 2011 at 10:17 PM, Chris Davis <chrismd@xxxxxxxxx> wrote:

> Hey everyone, I've finally gotten around to reviewing all the branches that
> got proposed for merging into trunk. My apologies for everyone who waited so
> long for feedback. So all the sudden there are some really cool things in
> trunk: Lucio's awesome client-side rendered flot UI, some much needed
> refactoring courtesy of Sidnei, also a new lineMode=connected option from
> katzj. After a few minor issues are worked out there will be several other
> branches merging in soon including lapsu's holtwinters branch, ziggy's
> whitelist/blacklist feature, and Sidnei's migration to twistd. Also don't
> forget Nick's new second Y-axis feature was just committed to trunk last
> week. Also the Dashboard UI in trunk is 10x better than what was put out in
> 0.9.8. Plenty of awesome stuff to try out, testing & feedback would be much
> appreciated.
>
> Despite my best efforts, the vast majority of my work on Graphite lately
> has been related to new features rather than docs or bug fixing, which is
> not what I intended but thats just how things have played out. This means
> the next release will not be a 1.0, too much new stuff coming in that needs
> time to bake first. That said, 0.9.9 will probably be ready much sooner than
> 1.0 would've anyways. I am still trying my best to focus on the
> documentation but for various reasons that hasn't been easy lately.
>
> Also I wanted to say that I think that my lack of time to spend on Graphite
> in recent weeks has created a bit of a bottleneck for development progress.
> I have a proposal for how we can address this, feedback welcomed. If anyone
> has a change (bug, branch, crazy idea, whatever) that is awaiting feedback
> and I'm not responsive enough to it, feel free to engage the rest of
> graphite-dev by emailing this list. Many eyes make all bugs shallow, and
> there are some great coders/eyes on this list that can review your changes.
> Also remember that every member of graphite-dev has commit access to trunk,
> and by virtue of you receiving this message that means you :).
>
> Code submissions generally fall into one of these two categories:
>
> 1. A small, targeted change that addresses a specific issue. The change has
> a well defined and narrow scope, creates a minimal diff, and has a clear
> purpose.
> 2. An exploratory accretion of ideas, involving design changes or
> significant new functionality. Usually these are many changes rolled into
> one branch.
>
> What I'd like to see happen, is for more changes of type #1 to get
> committed to trunk directly by the author. If you aren't fully confident in
> your change, ask someone on graphite-dev or #graphite to review it, simple
> as that. But you all have commit access and I'm not afraid of someone
> committing something broken once and a while, just go for it. But please try
> to put exactly 1 logical change in each commit rather than a bunch of things
> in one big commit/merge. This will help speed up development because many
> small changes are easier to review & merge than a few mega-changes, which
> are very costly to review.
>
> For changes of type #2, I think we should do more upfront discussion to
> avoid wasting effort as much as possible. If you've got an idea for Graphite
> that you think is cool, but maybe too substantial to just sneak into trunk
> under cover of darkness, please send a note to graphite-dev or #graphite so
> we can all give feedback, preferably before pouring in lots of effort
> writing the code. Sometimes people are already working down parallel paths
> but don't even know it because of lack of communication.
>
> tl;dr
>
> Anyways, that's enough of a rant for now. As always, feedback
> welcomed. Thanks,
>
> -Chris
>

Follow ups

References