[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Release management and testing problems

On Sat, Aug 03, 2002 at 01:08:59AM +0200, Raphael Hertzog wrote:
> Le Fri, Aug 02, 2002 at 04:27:24PM -0000, Moshe Zadka ?crivait:
> > >  56 of them have "unsatisfiable Depends".
> > So, if your idea worked *perfectly*, it would get these 56 in? 
> In aj's mode : 
> You don't know what you're talking about, go away.
> In my usual mode :
> Not at all, those "unsatisfiable Depends" are usually depencies on 
> non-existent packages.

Pretty much always these days, since I stopped outputing that line for
dependencies that could be satisfied by updating some other package.

> Most of the time the problem is specific to
> some architectures (because the dependency is not compiled on
> that architecture for example). Those problems happens frequently with
> meta-packages because they inconditionnaly (arch: all) depend
> on a large number of binary packages and some of them don't exist
> on some ports (look at meta-gnome for example).

Of course, testing ignores arch:all dependencies on !i386, so such issues
aren't relevant anyway.

> > > And finally 96 of them are "Valid candidate". The oldest "valid
> > > candidate" is 778 days old. The average "stuck time" for actual valid
> > > candidates is 84 days.
> Those are the packages that are stuck that shouldn't be stuck. The
> reason explaining why they are stuck is because one of their dependency
> is stuck. Why is the dependency stuck ? 

Often because two packages need to be installed at once, and because
fixing that's not tried all that frequently, since the testing scripts
aren't great at handling it.

Other problems are that some other package has to be updated and hasn't
been yet; this has often been the case when apache bumps it's version
number, and also happens when a new mozilla's been uploaded but a new
galeon hasn't, and at plenty of other times. There're whole bunches of
situations where packages are just fundamentally linked and can't be
updated separately.

> > I have long suspected the testing scripts can be improved algorithmically.
> > However, sadly, I have no time to do it, so I won't comment any more
> > until and unless someone who does have time asks me for input.
> Share the knowledge ! If someone on -devel likes your idea, he will
> implement it.

That's delusional, IME. If you like your idea, you get to implement it.

> > > This is an introductory statement yes, the problem was described in the
> > > paragraph below and it's a real problem. It's one of the worst source
> > > of "stuck packages".
> > Which constitute just 56/10k packages.
> We don't have 10k sources packages. And the right figure was 96.

We have 6535 source packages in unstable, making for a 1.47% hit rate.


Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.

 ``If you don't do it now, you'll be one year older when you do.''

Reply to: