[simon-mundy]

Types of changes

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• QA improvement (unit/integration tests, CS fixes, etc.) that does not change code behavior

Checklist:

• An issue has been created for this pull request.
• I have added tests for my changes.
• All new and existing tests passed.
• Documentation has been updated accordingly.
• This pull request targets the correct branch.

Description

PR 5 of 6 from #189: all 26 throw sites in src/Adapter, and the removal of VunerablePlatformQuoteException.

Stacked on #190. GitHub cannot base a cross-fork PR on a fork branch, so this targets 0.6.x and shows both commits. Only the second, 9870e985, belongs to this PR. Merge #190 first; I will rebase after. It does not depend on #191, #192 or #193.

Message text is byte-identical, verified template by template against the original literals.

• 13 templates and factories on Adapter\Exception\RuntimeException, 6 on its InvalidArgumentException, 2 on InvalidQueryException.
• The drifted PDO param message converges on the quoted form, per decision 3 on [RFC]: Named exception constructors and complete per-component exception namespaces #189. The two variants differed only in the quotes around %s.
• The four 'Data does not exist for this name/position' copies in ParameterContainer collapse onto one template.

VunerablePlatformQuoteException is removed

It held no state, implemented no external contract, and reported a fail-fast setup error a caller cannot recover from, so it met none of the three tests for a condition earning its own class. Its two throw sites now call RuntimeException::forVulnerablePlatformQuote(), which renders the identical message, and the misspelling goes with the class.

That also retires the one exemption in ExceptionHierarchyTest, which existed solely for this class. No exception class is exempt from the constant naming rule any more.

There is no BC shim, deliberately. The class was final, so nobody can have subclassed it, but a class_alias() to Adapter\Exception\RuntimeException would be worse than the break: catch (VunerablePlatformQuoteException) would silently start catching every adapter runtime failure. It already extended Adapter\Exception\RuntimeException, so anything catching that is unaffected.

The driver-error passthroughs

Pdo\Statement and Pdo\AbstractPdoConnection throw with PDO's own error text as the whole message. Both now use a DRIVER_ERROR template of '%s', documented on the constant as carrying the driver's wording verbatim. It is a thin template, but it keeps the throw site free of message assembly and keeps every factory under the naming guard.

One deviation from the plan on #189

InvalidConnectionParametersException keeps its constructor rather than gaining a named constructor. Its message is supplied by the adapter that throws it, so there is no template for core to own — an adapter-specific message does not belong in a core constant. It earns its own class on the extra state it carries.

Motivation and Context

Refs #189. Depends on #190.

Downstream work this creates. phpdb-pgsql and phpdb-sqlite each have one test file naming the removed class (3 and 2 references). Those need a one-line import swap and expectException(RuntimeException::class) in their own repositories. Both already require php-db/phpdb: ^0.6.0, which is unreleased, so those changes cannot go green until 0.6.0 ships — they should follow this merge, not precede it. phpdb-mysql has no references.

How has this been tested?

mago format --check, mago lint and mago analyze all exit 0. Full unit suite passes: 1681 tests, 3581 assertions.

Every factory is covered: 13/13, 6/6 and 2/2. All 21 templates verified byte-identical against the original literals by a throwaway script before the call sites were touched, including a check that forVulnerablePlatformQuote() reproduces the removed class's message exactly.

Baseline and configuration changes

Three baselined entries are re-attributed, not added to, because the same findings moved rather than appeared:

Was	Now
mixed-argument on Exception::__construct in AbstractPdoConnection	on InvalidQueryException::forDriverError
unhandled-thrown-type VunerablePlatformQuoteException in Sql92::quoteValue	RuntimeException
unhandled-thrown-type VunerablePlatformQuoteException in AbstractPlatform::quoteValue	RuntimeException

• src/Adapter/Exception/ is added to the too-many-methods exclude in mago.toml, as Route the Sql throw sites through named constructors #192 and Route the TableGateway and RowGateway throw sites through named constructors #193 do for their components. The threshold is 10 and the catalogue holds 13. This will conflict with Route the Sql throw sites through named constructors #192 and Route the TableGateway and RowGateway throw sites through named constructors #193 on the same line — the paths need merging into one array as they land.
• One halstead reading on Pdo\Statement is baselined. Removing long string literals reduces distinct operands, which raises halstead difficulty.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!