-
Notifications
You must be signed in to change notification settings - Fork 461
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Detect pending upgrades using dist-upgrade
#634
Conversation
Rebasing on top of master now that #633 is merged made TravisCI happy 😄 |
I'm not sure this is a great idea... It's normal for
That said, I think seeing these packages could be useful but they should probably be a separate fact. Especially since the current fact might be used in places to automate updates which up until now has been safe to do based on this fact's output. Your suggested change risks breaking a system. |
Prior to puppetlabs-apt 2.3 (which switched away from the Example: I have a 16.04 VM which has just been booted from an old cloud image. Running
The
In my case, the results of running
I would prefer this puppet module go back to the previous behaviour of listing all packages that can be upgraded, either by using |
@daenney, since we are only listing outdated packages into a fact (without taking any action), I would expect the Relying on Do you have in mind some use-cases where having only a part of the available updates (those that will be installed with |
What I want to guard against is that by changing this fact you can cause problems for others. If this was in any way relied upon by people to, for example, automate the upgrades of their hosts this could cause issue. Since all these facts are queriable from PuppetDB for example there's no knowing what other tools have been built on top of it. However since it seems that this used to be the behaviour we had before I don't think it's a problem to change it again. But, it should be an explicit decision to do so and to ensure that in the future that is kept. We can't keep doing this where we change an implementation of a fact in point releases which changes the packages it will consider as out of date versus not. |
OK, I understand your point: I considered this to be a regression when updating from 2.2.2 to 2.3.0, that only deserved a minor bump to 2.4.0 according to this semver faq item, and as I use Pinging @robinelfrink, the author of #581 where this behavior has changed, in case he has anything to say in regard to this. If everyone agrees, maybe this PR can be labelled e.g. "3.0.0"? |
The `upgrade` action only reports updated packages that have no changed dependencies. You must use `dist-upgrade` to see such upgrades.
bc966be
to
ce7fe36
Compare
This PR becomes oldish… I just rebased this on top of master. I had no news since I proposed to add this as part of a future major release, and 2 major releases have shipped since then 😞 |
Ping @hunner |
i am late to the show here, does it make sense that this is a different fact ? so people can pick and choose ? |
@smortex Thank you for your contribution. Based on your suggestion, dist-upgrade has been added as a separate fact as per PR #719 , and should be released during the next release. A separate fact will allow end users to pick and choose between either. I'm therefore closing off this PR. Thank you |
Excellent! Thanks for the heads-up @willmeek! |
After changing some
apt::source
andapt::pin
'ing some packages, I was surprised to not see any upgrades in the$::apt_package_updates
fact.Relying on
dist-upgrade
instead of a bareupgrade
made them visible.