Skip to content

FUNCTIONAL_ISSUES_SUMMARY.md

Latest commit

 

History

History
130 lines (93 loc) · 4.06 KB

FUNCTIONAL_ISSUES_SUMMARY.md

File metadata and controls

130 lines (93 loc) · 4.06 KB
Raw

Functional Issues Investigation Summary

Overview

After fixing the two critical root causes (RETURNING clause and LEAST/GREATEST type mapping), 375 tests (40.5% of remaining 925 failures) have actual functional issues requiring code-level fixes, not just test baseline updates.

Issue Categories

1. "Sequence Contains No Elements" - 175 Tests (18.9%)

Nature: Query execution returns empty result sets where data is expected.

Possible Causes:

  • Query translation differences between MySQL and MariaDB
  • Missing data in test fixtures for MariaDB
  • WHERE clause conditions that filter out all rows on MariaDB
  • Aggregate operations (First, Single, etc.) failing on empty sets

Investigation Needed:

  • Run subset of failing tests to capture actual error context
  • Determine if issue is data-related or query translation
  • Check if MariaDB-specific SQL syntax differences affect results

2. ComplexJSON Operations - 95 Tests (10.3%)

Nature: JSON operations failing on MariaDB 11.6.2.

Known Issues:

  • JSON bulk updates
  • JSON projections
  • JSON set operations
  • Potential differences in JSON function support between MySQL and MariaDB

Investigation Needed:

  • Verify MariaDB 11.6.2 JSON function compatibility
  • Check if JSON_TABLE, JSON_EXTRACT, etc. behave differently
  • Determine if JSON type mapping needs MariaDB-specific handling

3. Composing Expression Errors - 47 Tests (5.1%)

Nature: Query translation failures when composing complex expressions.

Likely Causes:

  • Nested query operations not translating correctly
  • Complex LINQ expressions not supported by query translator
  • Expression visitor issues with MariaDB-specific SQL

Investigation Needed:

  • Identify specific expression patterns that fail
  • Check if EF Core 10 changes affect expression composition
  • Determine if MySQL-specific optimizations don't work on MariaDB

4. Other Functional Issues - 58 Tests (6.3%)

Nature: Miscellaneous failures not fitting other categories.

Potential Issues:

  • LEFT JOIN not supported in certain contexts
  • Type conversion differences
  • Collation/character set issues
  • Stored procedure compatibility

Investigation Approach

Phase 1: Data Analysis (2-3 hours)

  1. Run sample tests from each category
  2. Capture full error context and stack traces
  3. Identify common patterns within each category
  4. Determine root causes for each category

Phase 2: Targeted Fixes (Variable - depends on findings)

  1. Sequence Contains No Elements:

    • If data issue: Update test fixtures
    • If query issue: Fix query translation

  2. ComplexJSON:

    • Check MariaDB JSON compatibility
    • Update JSON function mappings if needed

  3. Composing Expressions:

    • Fix expression visitor logic
    • Add MariaDB-specific handling

  4. Other Issues:

    • Address individually based on root cause

Phase 3: Verification (1-2 hours)

  1. Re-run affected test subsets
  2. Verify fixes don't break other tests
  3. Update documentation

Time Estimates

  • Investigation: 2-3 hours
  • Fixes: 5-15 hours (depends on complexity)
  • Testing: 1-2 hours
  • Total: 8-20 hours

Compare to baseline updates: 18-45 hours for test assertions only.

Current Status

✅ COMPLETED

  • RETURNING clause compatibility fix
  • LEAST/GREATEST type mapping fix
  • 90.6% failure reduction achieved
  • Comprehensive analysis and categorization

🔄 IN PROGRESS

  • Functional issue investigation

⏳ PENDING

  • Functional issue fixes (this document provides roadmap)
  • SQL baseline updates (550 tests - can be deferred)

Recommendations

  1. Priority: Fix functional issues first (actual bugs)
  2. Defer: SQL baseline updates (test assertions only)
  3. Incremental: Address functional issues by category
  4. Verify: Each fix with targeted test subset

Next Steps

  1. Run sample tests from "Sequence contains no elements" category
  2. Capture and analyze error patterns
  3. Implement targeted fixes
  4. Repeat for other categories

This provides a more valuable improvement than updating 550 test baselines that don't affect production functionality.