You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
i've noticed an incompleteness of output using test=True, when migrating from a distribution package repo to a backport-repo providing newer versions.
let's assume htop is installed from ubuntu-precise.
running a highstate using test=True will just show that pkgrepo.managed is executed. it won't detect that the package needs to be upgraded, as applying highstate normally does:
state.highstate:
----------
State: - pkgrepo
Name: precise-bleed
Function: managed
Result: True
Comment: Configured package repo precise-bleed
Changes: repo: ppa:precisebleed/ppa
----------
State: - pkg
Name: htop
Function: latest
Result: True
Comment: The following packages were successfully installed/upgraded: htop.
Changes: htop: { new : 1.0.2-2~ppa1~precise1
old : 1.0.1-1
}
----------
i can understand why this probably happens.
since the pkgrepo is not added for read during simulation, there's no package-information-change available and salt cannot see the available upgrade.
since the simulation diverges from reality, please at least include a hint in such situations, that additional packages may be changed.
i guess that's all you can do without running a real simulation.
The text was updated successfully, but these errors were encountered:
Yes, you hit the nail on the head. Since we aren't actually making the pkgrepo change or running an update, we have no idea whether new packages will be available.
We should add a note if test=True to say that some packages might require updates even if the states say they don't. Is that acceptable?
i've noticed an incompleteness of output using test=True, when migrating from a distribution package repo to a backport-repo providing newer versions.
let's assume htop is installed from ubuntu-precise.
running a highstate using test=True will just show that pkgrepo.managed is executed. it won't detect that the package needs to be upgraded, as applying highstate normally does:
state.highstate:
i can understand why this probably happens.
since the pkgrepo is not added for read during simulation, there's no package-information-change available and salt cannot see the available upgrade.
since the simulation diverges from reality, please at least include a hint in such situations, that additional packages may be changed.
i guess that's all you can do without running a real simulation.
The text was updated successfully, but these errors were encountered: