Re: [rdiff-backup-users] Plans?

[Kevin Fenzi]

On Fri, 24 May 2013 12:45:26 +0000
"Edward Ned Harvey (rdiff-backup)" <address@hidden> wrote:

> I'm technically a maintainer, but I don't want to set your hopes too
> high.  My big first motiviation was just the fact that the webpage
> was covered with broken links, and it was difficult to join this
> mailing list (because the link to the mailing list was broken.)  

We all start scratching some itch. ;) 
> 
> Even if there isn't a lot of development effort going on, things like
> that make it all look like run down garbage that nobody cares
> about ... and discourages anyone from thinking about contributing.
> 
> In my discussions with former developers / maintainers here so far,
> there has been a consistent theme:  They've consistently told me what
> happened to them was there was too much to do, they pushed too hard,
> they burned out.  And the one regret they have is putting their
> effort into development rather than finding more developers.

Understandable. 

> So here are my plans:  I'm going slow.  I plan to reach out to
> colleges and see if they will offer co-op student contribution for
> course credit.  I personally am not interested in rdiff-backup on
> windows or mac - but I recognize the importance of supporting it - 
> 
> Before ANYthing can move forward, we need the ability to perform
> regression testing.  The old tests exist in the repo, but I haven't
> looked yet at how to set it up.  We need test machines to run these
> tests on.  I can probably provide some of this - I have a
> virtualization infrastructure in my basement (who doesn't?)

Sure. If we can document how to setup and test we could also get others
doing them. Also, might look into something like buildbot or jenkins. 

> Anyway, I anticipate the need for some more overhead tasks before any
> development can begin.  And once that happens, I anticipate the first
> thing to do is look over the bug reports, to see what types and what
> severities there are.  See if any of them can be reproduced or
> included into regressions.  Some of them might get attention, some
> might not.  

Yep. 
 
> I'm not sure yet, what kind of data validation rdiff-backup does, but
> I have a feeling it should be improved.  And there are some frontend
> features I'd like to see added, such as ability to reference backups
> by rev number instead of date, ability to record log message at the
> time of backup, obviously ability to review log, and ability to
> compare rev A against rev B (without requiring a complete restore of
> both).  And hopefully, make validation a strong feature, and put it
> front and center.  (Like, mention it on the webpage, for example.)
> Right now, --verify *might* be a great feature.  But it's not on the
> examples page, (which is basically the only thing most users will
> read) and since "v" is late in the alphabet, they have to read the
> whole man page before they even find it.  So ...  It would be cool if
> literally the only change is to give that feature a more prominent
> position on the Examples page.

Sure. I think it's important to get maint back on track before really
digging into any development. Just fixing easy bugs/docs and getting a
new maint release out there will also cause people to see that things
are flowing again and potentially add a bunch of developers to help
with new features too. 

Thanks for the thoughts. Do post to the list when/if you need help on
anything. 

kevin

signature.asc
Description: PGP signature