about defining repository snapshots

[jeffss]

it seems some number of problems on upgrading the system is related to the time it takes between the upgrades, which is also dependent on the time it takes to resolve the dependencies, possible conflicts and then emerge time. As it is currently the options range from, to just regularly update the repository, independent of the status of the upgrade but each time it is done new problems are included before the upgrade is complete, while on the other side it involves waiting for the upgrade to be complete but as time increases so does the complexity of the upgrade. A possible management solution could be allowing the user to include some number of checkpoints, which means versioning the repositories folder and then progress through those versions from upgrade to upgrade; and some policy like removing the previous snapshots when emerging anything utilizing a newer

[Hu]

Gentoo makes the Portage tree available as a git repository, so you can theoretically take any particular developer's commit as a checkpoint. If you want the experience of having done weekly upgrades, despite now being more than a week behind, you can use git to visit commits a week apart and update to current as of the visited commit, then repeat. This is not exactly accurate since upstream distfiles may have been removed, but it should work well enough.

[NeddySeagoon]

jeffss,

There are also monthly squashfs snapshots.

The further back you need to go, the more difficult it gets to find distfiles.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.