User Details
- User Since
- Apr 12 2019, 4:27 PM (347 w, 13 h)
Yesterday
Thu, Dec 4
Wed, Dec 3
Tue, Dec 2
Thu, Nov 27
Wed, Nov 26
Mon, Nov 24
Fri, Nov 21
Thu, Nov 20
A related issue is here: https://vyos.dev/T8000.
Wed, Nov 19
Tue, Nov 18
Mon, Nov 17
Fri, Nov 14
Thu, Nov 13
Wed, Nov 12
Tue, Nov 11
Based on the comments above, I am moving this to 'wishlist' for consideration after the legacy backend is retired. My expectation is that we will not want to drop the invariant as such, however, we will be in a position to consider refinements of the model, radical or conservative, once freed from the legacy constraints.
Thu, Nov 6
Nov 5 2025
Nov 4 2025
Nov 3 2025
Nov 2 2025
Nov 1 2025
Oct 29 2025
Oct 28 2025
Oct 26 2025
The assumption stated in the description is true, as discussed with @c-po at the time. The following invariant is fundamental to the operation of the commit algorithm (legacy and modern) and the config daemons: during a commit, the active and proposed configs are fixed --- no updates during the commit are reflected in any script's config dict, including those called by config dependencies. Evolution of this model will require a refinement of config dependencies, which is not possible until the legacy backend is retired.
Oct 23 2025
Oct 21 2025
Oct 20 2025
Oct 8 2025
Some background: the scripts in question relate to the vyatta-image-tools.pl (not the image installer) and related scripts, cf. https://vyos.dev/T4038, for changes present in current but not 1.4.
On the other hand, the package vyatta-op was deprecated for versions > 1.4 in https://vyos.dev/T6527, and the file vyatta-image-complete was not included in vyos-1x: it would appear that the work above was to have provided a replacement, but that needs to be checked.
This will need a bit of analysis before a PR, but either we include the legacy script, or correct the errors in the possible replacement.
Oct 7 2025
Oct 6 2025
This was automatically closed; re-open until confirmed by testing.