Draft: sylva-units: linearize "all-unit-dependencies" (!7459) · Merge requests · Sylva-projects / sylva-core

When working on !7425 (merged) I realized that we might possibly improve sylva-units rendering performance by linearizing the "all-unit-dependencies" recursive named template, flattening it to not have it call itself recursively.

I wasn't sure what gains this would give, and if this was worth the effort, but I thought it was worth throwing Claude Code at it.

The prompt was simple:

can you linearize the following recursive function

(please also indent the code properly)

The first attempt was produced wrong results :-/

I tried.

please thoroughly review the code to make sure that:

it is equivalent

it does not miss dependencies which are there

it does not add dependencies which are not there

The second attempt was still wrong :-/

I tried:

I still get wrong results some dependencies are missing

And it catched remaining issues that lead to an implementation that produces the same result as the current code does. :-)

The result: 🤩

runtime reduced by 2 (17s -> 8.8s)
memory consumption reduced by 16 (1.2Gi -> 81 Mi)

Now, the resulting code is quite heavy and complex, and would be problematic to merge as such IMHO.

🤔

After taking a step back, I think that what will really be beneficial is to simplify the way we compute dependencies:

today we have something complex, because of the way we inject the cluster-machines-ready dependency dynamically in all units that aren't dependencies of cluster-machines-ready (yes, this sentence is a bit self-referential and this relates strongly with the fact that the corresponding code is complexe)
we can simplify this
if we do, we should be able to have a single pass global dependency loop detection that we should be able to implement as a non-recursive single pass loop, without needed fancy features like caching or the ability to ignore units along the way

CI configuration

Below you can choose test deployment variants to run in this MR's CI.

Click to open to CI configuration

Legend:

Icon	Meaning	Available values
☁️	Infra Provider	`capd`, `capo`, `capm3`
🚀	Bootstrap Provider	`kubeadm` (alias `kadm`), `rke2`, `okd`, `ck8s`
🐧	Node OS	`ubuntu`, `suse`, `na`, `leapmicro`
🛠️	Deployment Options	Deployment option list and description
🎬	Pipeline Scenarios	Available scenario list and description
🟢	Enabled units	Any available units name, by default apply to management and workload cluster. Can be prefixed by `mgmt:` or `wkld:` to be applied only to a specific cluster type
🔴	Disabled units	Any available units name, by default apply to management and workload cluster. Can be prefixed by `mgmt:` or `wkld:` to be applied only to a specific cluster type
🏗️	Target platform	Can be used to select specific deployment environment Available platform list and description

Global config for deployment pipelines

autorun pipelines
allow failure on pipelines
record sylvactl events

Notes:

Enabling autorun will make deployment pipelines to be run automatically without human interaction
Disabling allow failure will make deployment pipelines mandatory for pipeline success.
if both autorun and allow failure are disabled, deployment pipelines will need manual triggering but will be blocking the pipeline

Be aware: after configuration change, pipeline is not triggered automatically. Please run it manually (by clicking the run pipeline button in Pipelines tab) or push new code.

Edited Apr 03, 2026 by Thomas Morin

Draft: sylva-units: linearize "all-unit-dependencies"

CI configuration

Global config for deployment pipelines

Merge request reports