Re: backporting Litecoin [not-in-testing]
On 06/04/2018 09:23 AM, Dmitry Smirnov wrote:
> On Monday, 4 June 2018 4:44:52 PM AEST Alexander Wirt wrote:
>> Then we agree to disagree. We don't want backports that are not in testing.
>
> I can't think of a better example of a package that worth exception for "not-
> in-testing" rule. Can you be convinced? Have you ever granted such exception
> for non-security related reason?
>
> Of course you know that sometimes package can not be in testing due to
> library transitions or removals. "not-in-testing" should be less strict and
> reasonable exceptions should be possible to get.
The reason for it is to make sure there's always an upgrade path. If a
package is retained because of a library transition, then just wait. If
you don't, then anyone upgrading from stable+backports to testing will
have a broken setup (with possibly packages linked against the wrong
version of the shared lib).
> Your objection to backporting Litecoin is not reasonable because it makes no
> sense.
It does make sense. Instead, why not addressing problems in litecoin so
it gets in Testing?
> Backports are not the same as "stable" - if they were the same then there
> would be no need for backports. If "not-in-testing" rule is for quality
> purposes
Probably it is for quality reason also, but as I understand, the main
problem is to provide an upgrade path. Pushing packages to backports
without having them in testing breaks that.
Cheers,
Thomas Goirand (zigo)
Reply to: