feat: IRIS pgwire compatibility fixes for DDL, parameters, and timestamps

- Fixed multi-statement DDL with comments by enhancing DdlSplitter and stripping comments
- Consolidated parameter translation ( -> ?) in SQLTranslator across all query paths
- Implemented DefaultValuesTranslator to rewrite INSERT statements with DEFAULT in VALUES
- Normalized ISO 8601 timestamps to IRIS-compatible format
- Translated PostgreSQL ALTER TABLE SET DATA TYPE/DROP NOT NULL to IRIS-compatible syntax
- Enhanced IRIS data type mapping for PostgreSQL OID compatibility
- Added comprehensive unit and E2E integration tests
- Bumped version to 1.0.6

Author: tom-dyar

AGENTS.md

@@ -5,6 +5,8 @@ Auto-generated from all feature plans. Last updated: 2026-01-02
 ## Active Technologies
 - Python 3.11 + python>=3.11, psycopg[binary], iris-devtester, intersystems-irispython (026-address-gaps-in)
 - PostgreSQL (via InterSystems IRIS) (026-address-gaps-in)
+- Python 3.11 + intersystems-irispython, psycopg[binary], iris-devtester (034-issues-that-likely)
+- InterSystems IRIS (via PostgreSQL wire protocol) (034-issues-that-likely)
 
 - Python 3.11+ + `iris-devtester`, `intersystems-irispython`, `psycopg[binary]` (033-devtester-skills)
 
@@ -24,6 +26,7 @@ cd src [ONLY COMMANDS FOR ACTIVE TECHNOLOGIES][ONLY COMMANDS FOR ACTIVE TECHNOLO
 Python 3.11+: Follow standard conventions
 
 ## Recent Changes
+- 034-issues-that-likely: Added Python 3.11 + intersystems-irispython, psycopg[binary], iris-devtester
 - 026-address-gaps-in: Added Python 3.11 + python>=3.11, psycopg[binary], iris-devtester, intersystems-irispython
 
 - 033-devtester-skills: Added Python 3.11+ + `iris-devtester`, `intersystems-irispython`, `psycopg[binary]`

CHANGELOG.md

@@ -5,42 +5,22 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
+## [1.0.6] - 2026-01-16
 
 ### Added
-- **Multi-Action ALTER TABLE Splitting**: Automatically decomposes PostgreSQL-style multi-column `ALTER TABLE` statements (e.g., multiple `ADD COLUMN` or `DROP COLUMN` actions) into individual IRIS-compatible statements.
-- **IRIS Bridge Gaps** (Feature 026): Comprehensive performance and functionality enhancements for IRIS integration
-  - **Fast Path Bulk Insert**: Protocol-level batching achieving 3,700+ rows/second (11× improvement)
-  - **HNSW Index Translation**: Support for PostgreSQL `USING hnsw` translated to IRIS `AS HNSW`
-  - **Recursive JSON Pathing**: Support for nested `->` and `->>` operators translated to `JSON_VALUE`
-  - **DDL Idempotency**: Full `IF NOT EXISTS` support for `CREATE TABLE` and `CREATE INDEX`
-  - **Simple Query Translation**: Enabled full SQL translation for the standard PostgreSQL query protocol
-- **P6 COPY Protocol** (Feature 023): PostgreSQL COPY FROM STDIN and COPY TO STDOUT for bulk data operations
-  - Bulk data import/export with CSV processing and streaming
-  - 1000-row batching for memory efficiency (<100MB for 1M rows)
-  - Transaction integration with automatic rollback on errors
-  - Query-based export support (`COPY (SELECT ...) TO STDOUT`)
-  - Performance: 600+ rows/second sustained throughput
+- **Multi-statement DDL with comments**: Updated `DdlSplitter` to be fully comment-aware and strip comments before execution to prevent IRIS parsing errors.
+- **Prepared statement translation ($n → ?)**: Consolidated parameter translation logic into `SQLTranslator` for consistency across all query paths.
+- **Default keyword in VALUES clause**: Implemented `DefaultValuesTranslator` to rewrite `INSERT` statements using `DEFAULT` within `VALUES` lists.
+- **Timestamp binding normalization**: Updated `DATETranslator` and `IRISExecutor` to normalize ISO 8601 timestamps (stripping `T`, `Z`, and offsets) into IRIS-accepted ODBC formats.
+- **ALTER TABLE translation**: Updated `DdlSplitter` to translate PostgreSQL `SET DATA TYPE` and `DROP NOT NULL` syntax to IRIS-compatible `ALTER COLUMN` commands.
 
 ### Fixed
-- Fixed critical bug in translation cache causing `TypeError` with mixed naive/aware datetimes
-- Standardized all internal timestamps to timezone-aware UTC
-- Fixed `CREATE INDEX` idempotency via specialized comment marker `/* IF_NOT_EXISTS */`
-- Dynamic versioning recognition in package metadata validation
-- Python bytecode cleanup (95+ artifacts removed from git)
-- Black code formatting (20 files reformatted to compliance)
-- asyncpg parameter type OID inference from CAST expressions
-- PostgreSQL compatibility documentation improvements
+- Fixed IRIS execution error "Input encountered after end of query" by improving semicolon and comment handling in the DDL splitter.
+- Resolved "LITERAL (1) found" errors during migrations by avoiding no-op SELECT injections for skipped DDL.
+- Enhanced IRIS data type mapping to better handle PostgreSQL OIDs (e.g., `BIGINT`, `TINYINT`).
 
-### Security
-- Upgraded authlib to 1.6.5 (fixes 3 HIGH severity CVEs)
-- Upgraded cryptography to 46.0.3 (fixes 1 HIGH severity CVE)
+## [1.0.5] - 2026-01-15
 
-### Performance
-- High-performance DML Fast Path: Buffering and collapsing `Sync` cycles into bulk IRIS calls
-- IRIS executemany() optimization for 4-10× performance improvement in bulk operations
-- COPY protocol optimized for 600+ rows/second sustained throughput
-- Memory-efficient streaming for large result sets
 
 ## [0.1.0] - 2025-01-05
 

specs/034-issues-that-likely/checklists/requirements.md

@@ -0,0 +1,35 @@
+# Specification Quality Checklist: IRIS pgwire compatibility fixes
+
+**Purpose**: Validate specification completeness and quality before proceeding to planning
+**Created**: 2026-01-16
+**Feature**: [Link to spec.md](../spec.md)
+
+## Content Quality
+
+- [x] No implementation details (languages, frameworks, APIs)
+- [x] Focused on user value and business needs
+- [x] Written for non-technical stakeholders
+- [x] All mandatory sections completed
+
+## Requirement Completeness
+
+- [x] No [NEEDS CLARIFICATION] markers remain
+- [x] Requirements are testable and unambiguous
+- [x] Success criteria are measurable
+- [x] Success criteria are technology-agnostic (no implementation details)
+- [x] All acceptance scenarios are defined
+- [x] Edge cases are identified
+- [x] Scope is clearly bounded
+- [x] Dependencies and assumptions identified
+
+## Feature Readiness
+
+- [x] All functional requirements have clear acceptance criteria
+- [x] User scenarios cover primary flows
+- [x] Feature meets measurable outcomes defined in Success Criteria
+- [x] No implementation details leak into specification
+
+## Notes
+
+- Items marked incomplete require spec updates before `/speckit.clarify` or `/speckit.plan`
+- Validation pass after 1 iteration; no clarifications required.

specs/034-issues-that-likely/contracts/alter-table-compat.md

@@ -0,0 +1,22 @@
+# Contract: ALTER TABLE Compatibility
+
+## Purpose
+Define how ALTER TABLE SET DATA TYPE and DROP NOT NULL behave for IRIS.
+
+## Inputs
+- ALTER TABLE statements that change column type or nullability.
+
+## Outputs
+- IRIS-compatible ALTER TABLE statements or clear errors when unsupported.
+
+## Rules
+1. **SET DATA TYPE translation**: `ALTER COLUMN x SET DATA TYPE T` MUST be translated to `ALTER COLUMN x T` when supported.
+2. **DROP NOT NULL translation**: `ALTER COLUMN x DROP NOT NULL` MUST be translated to `ALTER COLUMN x NULL` when supported.
+3. **Best-effort behavior**: When IRIS rejects the operation due to constraints, return a clear, actionable error.
+
+## Error Handling
+- Errors MUST indicate the column and operation that failed and suggest IRIS constraints as the cause.
+
+## Acceptance Criteria
+- Supported ALTER TABLE changes execute successfully in IRIS.
+- Unsupported changes return clear, actionable errors.

specs/034-issues-that-likely/contracts/ddl-splitting.md

@@ -0,0 +1,23 @@
+# Contract: DDL Splitting and Comment Handling
+
+## Purpose
+Ensure multi-statement SQL and DDL with comments are split safely without corrupting statement boundaries.
+
+## Inputs
+- SQL text potentially containing multiple statements, comments, and string literals.
+
+## Outputs
+- Ordered list of individual SQL statements, preserving semantics.
+
+## Rules
+1. **Comment-aware splitting**: Semicolons inside `--` or `/* */` comments MUST NOT split statements.
+2. **String-aware splitting**: Semicolons inside single or double-quoted strings MUST NOT split statements.
+3. **Leading comments**: Leading comment blocks MUST be preserved with their associated statement.
+4. **ALTER TABLE multi-action**: Multi-action `ALTER TABLE` statements MUST be decomposed into single-action statements for IRIS execution.
+
+## Error Handling
+- If DDL decomposition fails, return a clear error that includes the original statement context.
+
+## Acceptance Criteria
+- DDL scripts with comments and multiple statements execute in order without corruption.
+- Multi-action ALTER TABLE statements are split into valid IRIS statements.

specs/034-issues-that-likely/contracts/default-values.md

@@ -0,0 +1,22 @@
+# Contract: DEFAULT in VALUES Handling
+
+## Purpose
+Provide IRIS-compatible behavior for INSERT statements that use `DEFAULT` within a VALUES list.
+
+## Inputs
+- INSERT statement with column list and VALUES list that may contain `DEFAULT`.
+
+## Outputs
+- INSERT statement rewritten to omit DEFAULT column/value pairs.
+
+## Rules
+1. **DEFAULT removal**: Columns paired with `DEFAULT` in VALUES MUST be removed from both column list and values list.
+2. **Row integrity**: Remaining columns and values MUST preserve ordering and alignment.
+3. **Row-level DEFAULT VALUES**: `INSERT INTO table DEFAULT VALUES` passes through unchanged.
+
+## Error Handling
+- If DEFAULT removal would empty the column list while values remain, return a clear error.
+
+## Acceptance Criteria
+- INSERT statements with per-column DEFAULT values execute successfully in IRIS.
+- Defaults are applied as defined by table schema.

specs/034-issues-that-likely/contracts/parameter-translation.md

@@ -0,0 +1,22 @@
+# Contract: Prepared Statement Parameter Translation
+
+## Purpose
+Guarantee that prepared statements use IRIS-compatible parameter placeholders across all protocol paths.
+
+## Inputs
+- SQL statement containing PostgreSQL positional parameters (`$1`, `$2`, ...).
+
+## Outputs
+- SQL statement with IRIS-compatible `?` placeholders and translated casts.
+
+## Rules
+1. **Placeholder translation**: All `$n` placeholders MUST be translated to `?`.
+2. **Type cast translation**: PostgreSQL `::type` casts MUST be rewritten as IRIS `CAST(... AS type)`.
+3. **Apply to all query paths**: Translation MUST run for simple and extended protocol flows.
+
+## Error Handling
+- If translation encounters unsupported casts, return a clear error specifying the cast.
+
+## Acceptance Criteria
+- Prepared statements using `$n` placeholders execute successfully on IRIS.
+- Parameter description inference aligns with translated casts.

specs/034-issues-that-likely/contracts/sql-normalization.md

@@ -0,0 +1,25 @@
+# Contract: SQL Normalization Pipeline
+
+## Purpose
+Define the required ordering and behaviors for SQL normalization so that IRIS receives compatible statements without client-side rewrites.
+
+## Inputs
+- Raw SQL statement from PostgreSQL client.
+
+## Outputs
+- Normalized SQL statement suitable for IRIS execution.
+
+## Rules
+1. **Parameter translation precedes normalization**: `$n` placeholders MUST be translated to `?` before any normalization steps run.
+2. **Schema mapping**: Unqualified `public` schema references MUST map to the configured IRIS schema.
+3. **Identifier normalization**: Unquoted identifiers MUST be normalized for IRIS compatibility.
+4. **Date literal translation**: ISO date literals MUST be translated to IRIS-accepted forms.
+5. **JSON operator translation**: `->` and `->>` MUST be mapped to IRIS JSON equivalents.
+6. **Vector type normalization**: Vector type declarations MUST be normalized to IRIS vector syntax.
+
+## Error Handling
+- If translation fails, return a clear error indicating the unsupported construct.
+
+## Acceptance Criteria
+- Given a SQL statement with parameters and supported constructs, the output is valid IRIS SQL.
+- Given a statement with unsupported constructs, a clear, actionable error is returned.

specs/034-issues-that-likely/contracts/timestamp-normalization.md

@@ -0,0 +1,22 @@
+# Contract: Timestamp Normalization
+
+## Purpose
+Ensure timestamp values are accepted by IRIS when clients provide ISO 8601 strings or binary-encoded timestamps.
+
+## Inputs
+- Timestamp literals in SQL or bound parameter values.
+
+## Outputs
+- IRIS-compatible timestamp strings (`YYYY-MM-DD HH:MM:SS[.fff]`).
+
+## Rules
+1. **Timezone suffix stripping**: `Z` or offset suffixes MUST be removed for standard timestamp handling.
+2. **Binary decoding**: Binary timestamp parameters MUST be converted to IRIS-compatible strings.
+3. **Microseconds support**: Preserve fractional seconds when present.
+
+## Error Handling
+- If a timestamp cannot be normalized, return a clear error with the offending value.
+
+## Acceptance Criteria
+- ISO 8601 timestamps with `T`/`Z` or offsets are accepted after normalization.
+- Binary timestamp parameters execute successfully on IRIS.

specs/034-issues-that-likely/data-model.md

@@ -0,0 +1,33 @@
+# Data Model: IRIS pgwire compatibility fixes
+
+## Entities
+
+### SQL Statement
+- **Description**: Client-submitted SQL text that may contain comments, parameters, and multiple statements.
+- **Key Fields**:
+  - `raw_text`: Original SQL text from client.
+  - `normalized_text`: SQL after normalization/translation steps.
+  - `statement_list`: Parsed list of individual statements (order-preserving).
+- **Relationships**:
+  - Contains zero or more `Parameter Binding` entries.
+
+### Parameter Binding
+- **Description**: Positional parameter placeholders and runtime values supplied by the client.
+- **Key Fields**:
+  - `placeholder_style`: `$n` (PostgreSQL) or `?` (IRIS).
+  - `values`: Ordered list of bound parameter values.
+  - `types`: Optional inferred types for parameter descriptions.
+- **Relationships**:
+  - Belongs to a `SQL Statement`.
+
+### Timestamp Value
+- **Description**: Temporal values passed as literals or bound parameters that must be compatible with IRIS.
+- **Key Fields**:
+  - `input_format`: ISO 8601 variants (with/without timezone suffix).
+  - `normalized_format`: IRIS-accepted ODBC-like timestamp string.
+- **Validation Rules**:
+  - Timezone suffixes are stripped for standard timestamp columns.
+
+## State Transitions
+- **Statement Lifecycle**: `raw_text` → `normalized_text` → `statement_list` → executed.
+- **Parameter Lifecycle**: `$n` placeholders + values → translated placeholders + normalized values.