Review: Syncing from testing a success?

Matthias Andree matthias.andree at
Thu Apr 8 23:23:58 BST 2010

Am 08.04.2010, 21:41 Uhr, schrieb Martin Pitt:

> So, in your opinion, did syncing from testing
>  (1) help to avoid introducing larger breakage into Ubuntu (for the
>      domain you are usually watching)
>  (2) meant a smaller or larger amount of review and sync requests
>  (3) made it easier or harder to merge with Debian and get changes
>      integrated back upstream
>  (4) made library transitions easier or harder
>  (5) anything else that caused or eased problems that you can think
>      of
> In order to not bias the discussion, I'll send my own experience as a
> followup later on.
> Thank you in advance for your answers, and let's have an outstanding
> Lucid release!

As upstream maintainer of a couple of packages, I must say I'm rather  
unhappy that I see outdated versions currently in pipe for Lucid, namely

- bogofilter. For Lucid, it is currently 1.2.0, but should be 1.2.1, which  
addresses critical bugs that massively impair parsing particular messages.  
The newer version, albeit available for more than half a year, hasn't made  
it into testing because bogofilter-tokyocabinet is waiting for  
tokyocabinet builds on arcane architectures unsupported by Ubuntu, and  
Ubuntu - AFAIR - wouldn't use the tokyocabinet version anyways, only -bdb  
and -sqlite.
Here, this means missed upstream fixes and also an update request  

Note that either bogofilter or SpamAssassin is a requisite for Evolution  
spam filtering, and hence used by less-skilled users. Note also that there  
is a pertinent bogofilter bug that wasn't forwarded upstream  
<> - and  
there is no way that syncing alone could fix that. It's pure coincidence  
this was discovered last year by me so I could fix it upstream.

- fetchmail is at 6.3.9-rc2, but should be 6.3.16 - I'm not sure 6.3.16  
will make the Ubuntu freeze in time, because that was uploaded only today.  
I'm wondering why 6.3.13 wasn't picked up, it was available in testing  
since 2010-02-10. Again, lots of missed upstream fixes, among them users'  
pet peeves.  

(1) using testing may occasionally prevent breakage, but at the same time  
it may also withhold bug fixes from users.

(2) not my call,

(3) harder because you're not using the latest version, however "get  
changes integrated back upstream" just doesn't happen. I have NEVER seen  
Ubuntu bug reports or patches forwarded upstream, or even to Debian.  
Maintainer contacts with Debian are tight for the packages mentioned, but  
Ubuntu - for upstream - is just a blank box of people weasling away  
whenever you need one.

(4) no experience

(5) caused problems as described above.

I know things will be the other way around for other packages. That  
depends on the package, and on the point in time that you look at a  
particular package. It will change over time, even for the same package.

I think what's more important than to discuss merging modes is getting  
bugs triaged and dispatched (upstream, request for more info, ...) much  
quicker (at most 90 days after report), and COMMUNICATE with Debian and  
with upstream maintainers. Without that you're not going to know if  
testing or unstable is "better" or "easier" or whatever trait you're  
looking at.

Matthias Andree

More information about the ubuntu-devel mailing list