Skip to content

doc(coverage): two of the three "untested" functions cannot be covered there - #61

Merged
kzangeli merged 1 commit into
mainfrom
doc/coverage-uncoverable-not-untested
Sep 4, 2026
Merged

doc(coverage): two of the three "untested" functions cannot be covered there#61
kzangeli merged 1 commit into
mainfrom
doc/coverage-uncoverable-not-untested

Conversation

@kzangeli

@kzangeli kzangeli commented Sep 4, 2026

Copy link
Copy Markdown
Collaborator

doc/coverage.md has said since yesterday that the never-entered bucket's untested-behaviour group is down to three functions. Only one of them is a gap in the suite.

geoEntityValidate lives in plugins/shared/, which no run excludes, and is called only from plugins/currentState/corDB/, which a mongoc run does exclude. It can never be non-zero in the mongoc column, whatever anyone writes — and all four of its diagnostics are already asserted, by geoproperty_degenerate_polygon and three others. attrInstanceOf is the same shape one level up: it is reached only when the broker evaluates q in-process, which a mongoc query does not do.

The mongoc run's never-entered list contains functions only the corDB path can reach, and vice versa. The file already says this about the corDB column — "a property of the run, not of the code" — and the mirror case is easy to miss, because a mongoc-only zero looks exactly like an untested function until you follow the caller. Both of these were chased as gaps before anyone did, which is how the correction came to be needed.

corRestClientResponseHeader remains, and is real: it needs a Context Source answering 3xx with a Location header, which ftClient's /mock/reply cannot express.

Documentation only — no code, no test impact.

…d there

This file has said since yesterday that the never-entered bucket's
untested-behaviour group is down to three functions. Only ONE of them is a
gap in the suite.

geoEntityValidate lives in plugins/shared/, which no run excludes, and is
called only from plugins/currentState/corDB/, which a mongoc run DOES
exclude. It can never be non-zero in the mongoc column, whatever anyone
writes - and all four of its diagnostics are already asserted, by
geoproperty_degenerate_polygon and three others. attrInstanceOf is the
same shape one level up: reached only when the broker evaluates `q`
in-process, which a mongoc query does not do.

⭐ The mongoc run's never-entered list contains functions only the corDB
path can reach, and vice versa. The file already says this about the corDB
column - "a property of the run, not of the code" - and the mirror case is
easy to miss, because a mongoc-only zero looks exactly like an untested
function until you follow the caller. Both of these were chased as gaps
before anyone did, which is how the correction came to be needed.

corRestClientResponseHeader remains, and is real: it needs a Context Source
answering 3xx WITH a Location header, which ftClient's /mock/reply cannot
express.
@read-the-docs-community

Copy link
Copy Markdown

Documentation build overview

📚 coraine | 🛠️ Build #34391437 | 📁 Comparing 459ff45 against latest (5bb5c5d)

  🔍 Preview build  

1 file changed
± coverage.html

@kzangeli
kzangeli merged commit 9c0707f into main Sep 4, 2026
3 checks passed
@kzangeli
kzangeli deleted the doc/coverage-uncoverable-not-untested branch September 4, 2026 11:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant