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

Re: qiime REMOVED from testing



On Wed, 22 Jan 2014 08:26:56 +0100 (CET)
"Steffen Möller" <steffen_moeller@gmx.de> wrote:

> The problem with the qiime removal imho is less the extra work that
> you ask for, but the message sent to our users of testing that they
> cannot rely on us - I know it says "testing", but this is exactly
> what they should be allowed to do reliably.

0: The bug has been open for some time - if there is that much demand,
then someone needs to fix the bug.

1: The package is only removed from the repository, not installed
systems. The removal only affects new installs.

Stop whining and fix the bug already!
 
> Researchers using qiime use the latest version since the scientific
> field (identify the relative abundance of microbiota) develops so
> quickly.

If the package was developing so quickly, why was the package allowed to
get sufficiently stale that the bug wasn't fixed?

All the bug needed was a new version of the package.

> The individuals who decided for qiime on Debian (not the
> upstream-provided binary distribution)  find testing a natural
> environment. They are on the latest scientifically and run on the
> latest technically. This goes together.

Until technical bugs in the software force someone else to do the work
instead, i.e. the auto-removal from testing process.
 
> Now, when you retract packages for no scientific reason or for a
> technical reason that would affect them, then they will look to
> Debian with big eyes asking "why did you do this to us?"

Because the people entrusted with looking after the package in Debian -
the maintainers - did not fix the bug.

The technical reason does affect them - it blocks updates of other
packages, some of which may contain much needed updates.

> I know, the problem is old. And you may not have any immediate answer
> that would make me happy. Just kindly think about it anyway. 

The general process has been thought about long and hard, many, many
times and this is the correct solution. If the maintainers cannot find
time to fix the bug properly, it is up to Debian to get this package
out of the way so that other packages which follow the rules can update
correctly.

Blocking supported packages with abandonware is a sufficiently
important problem that automated removal of packages is entirely
warranted.

Fix the bug and the entire problem goes away!

-- 


Neil Williams
=============
http://www.linux.codehelp.co.uk/

Attachment: signature.asc
Description: PGP signature


Reply to: