NVIDIA / nodewright / 32314880358
83%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 61
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 80.329% (+1.1%) from 79.182%
32314880358

push

github

web-flow 
test(operator): cover pure-logic gaps in the nodewright API, wrapper, version and CLI utils (#498)

* test(operator): cover pure-logic gaps in the nodewright API, wrapper, version and CLI utils

The nodewright API group had no types tests at all, while the legacy skyhook
group has two. Its pure logic was only reached sideways, through envtest webhook
specs and e2e, so branches that no rollout happens to take were unexercised:
StateToStatus and WasUpdated at 0%, getSafetyLimit at 0%, getFailureThreshold at
43%, Compartment.Validate at 50%.

Adds the two missing types test files alongside the existing webhook suite
(matching how api/v1alpha1 is laid out), a first test file for internal/version,
and specs for the wrapper taint helpers, Compartment.GetNode and the CLI table
and YAML output helpers, none of which had direct tests.

Coverage of the touched files:

  api/nodewright/v1alpha1/deployment_policy_types.go  81.5% -> 100.0%
  api/nodewright/v1alpha1/nodewright_types.go         86.0% ->  97.4%
  internal/version/version.go                         68.4% -> 100.0%
  internal/wrapper/node.go                            86.2% ->  92.3%
  internal/wrapper/compartment.go                     90.5% ->  94.0%
  internal/cli/utils/utils.go                         73.7% ->  80.3%

Merged across all CI suites this moves the total from 78.7% to 79.8%.

Tests only, no production code touched, so there is no behaviour change to
document and no RELEASE_NOTES.md entry.

Signed-off-by: Alex Yuskauskas <ayuskauskas@nvidia.com>

* test(operator): strengthen assertions from review feedback

BuildGraph only checked that the graph was non-nil, which passes even if the
dependsOn edge is dropped. Walk it instead: base is ready first, tuning unblocks
once base is done.

The two narrow-table cases asserted too little to catch a regression. The no-rows
case only counted lines, so two wrong lines would pass; the missing-WideExtract
case only asserted node-1 was absent, so empty ... (continued)

11275 of 14036 relevant lines covered (80.33%)

8.09 hits per line

Coverage Regressions

Lines Coverage ∆ File
4
80.35
 -0.09% operator/internal/controller/skyhook_controller.go
Jobs
ID Job ID Ran Files Coverage
1 32314880358.1 61
80.33
 GitHub Action Run
Source Files on build 32314880358