Page 1 of 1

Re: How long the gap between testing to testing transition?

Posted: 2017-04-16 20:02
by pylkko
Job wrote:This is silly I don't have a handle on the inner working of Debian circle but a testing made stable a day ago can't be much different from a testing established a day ago.
I see what you are saying and in theory this is true. That is, when testing is made stable, what happens is that a "snapshot" is taken from a "rolling" package collection and that stays as it is while from where it is taken from continues on rolling. In this sense one would expect testing and stable to be more or less equal in some short period after the transition. However, in practice, during the freeze very little if any new packages from upstream sources are put into the Debian package "universe" (in other words into unstable...some are passed around in experimental but others not at all) and package managers are focusing on other things. Therefore, right after stable is finally stable, most package managers have a lot to catch up to and there is a more-rapid-than-normal influx of new stuff after the transition.

Re: How long the gap between testing to testing transition?

Posted: 2017-04-16 21:25
by phenest
bester69 wrote:i dont understand how this transiction is done, if it happens in one day, one week, or a month
Once Stretch is Stable, a new "testing" is created immediately. However, it can take a few weeks before any changes start to happen.

Re: How long the gap between testing to testing transition?

Posted: 2017-04-16 21:38
by No_windows
phenest wrote:
bester69 wrote:i dont understand how this transiction is done, if it happens in one day, one week, or a month
Once Stretch is Stable, a new "testing" is created immediately. However, it can take a few weeks before any changes start to happen.
I believe the phrase that I've read in the wiki, or Deb pages someplace is that the "new" Testing is simply a copy of the new Stable...... in other words for a moment in time they are equivalent. I believe Bester's question is how long, and I would theorize the answer is that "it depends", and "it doesn't matter".

The simple solution here is to change to the code name ahead of time.......I'll have to look, but I think I'm already using "jessie" not "stable", so when Jessie passes to oldstable nothing will change for me.

There doesn't need to be any "what if" here, change it now and be done with it.

Re: How long the gap between testing to testing transition?

Posted: 2017-04-17 03:58
by pendrachken
Job wrote:This is silly I don't have a handle on the inner working of Debian circle but a testing made stable a day ago can't be much different from a testing established a day ago.


Not "silly" at all.

Stable is released > testing gets opened up to all the juicy new packages that sat in unstable during the freeze prior to release, and packages flood in. Anyone who wanted to migrate to stable after bug hunting in the frozen testing updates after even the first package comes into testing now has a not purely stable system.

This could be within minutes or hours of the stable release, and I would wager almost a certainty within 24-36 hours of release. It's not worth the risk if you wanted to run stable. That is the whole reason for codenames - to avoid situations like this. You ONLY have "testing" in your sources.list if you ONLY want to run testing, otherwise you use the current testing branch code name so as to make a smooth transition to a sane and stable system upon the next stable release.