[Pkg-puppet-devel] Are we ok with the state of 0.25.0 ?

[Nigel Kersten]

On Wed, Oct 14, 2009 at 6:15 AM, Micah Anderson <micah at riseup.net> wrote:

> * Nigel Kersten <nigel at explanatorygap.net> [2009-10-14 08:39-0400]:
> > On Wed, Oct 14, 2009 at 5:30 AM, Micah Anderson <micah at riseup.net>
> wrote:
> >
> > > * Andrew Pollock <apollock at debian.org> [2009-09-26 17:55-0400]:
> > > > On Sat, Sep 26, 2009 at 11:48:21AM -0400, Micah Anderson wrote:
> > > > >
> > > > > I'm really glad this got sorted out, I haven't tried to go through
> the
> > > > > same steps you have yet, so my brain isn't fully wrapped around how
> > > this
> > > > > works, or why. I'm afraid that when I do need to do this, I will
> have
> > > to
> > > > > dig through the mailing list archives and try to reconstruct the
> steps
> > > > > based on what folks reported here... it might be better if we
> > > maintained
> > > > > a page to preserve these details, maybe on wiki.debian.org?
> > > >
> > > > Yes, we should definitely maintain some documentation, either on the
> Wiki
> > > or
> > > > on the Alioth project page.
> > >
> > > This was a minor digression into the need for documentation of the git
> > > procedures of the package, I'm sorry if it distracted from the main
> > > point, which was getting 0.25 uploaded.
> > >
> > > What, if anything, are we waiting on at this point?
> > >
> >
> > 0.25.1, which will be very very soon, within days.
>
> Sounds reasonable.
>
> > There was a potential data corruption bug that I was very concerned about
> in
> > 0.25.0, but it's looking like that's actually doing more correct erb
> > template evaluation than 0.24.x was.
>
> Do you have a reference to this? if its not an actual bug, I'm going to
> want to check my templates before I corrupt things :) In fact, it might
> be worth adding to NEWS?
>

The bug history is a little confusing, but it gets clearer towards the end.

http://projects.reductivelabs.com/issues/2615

There are a bunch of other associated bugs, but it turns out that the
problem has to do with bugs in serialization libraries or differences
between versions.