惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

B
Blog RSS Feed
有赞技术团队
有赞技术团队
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
阮一峰的网络日志
阮一峰的网络日志
V
V2EX
Y
Y Combinator Blog
Jina AI
Jina AI
G
Google Developers Blog
Last Week in AI
Last Week in AI
博客园 - 叶小钗
H
Hackread – Cybersecurity News, Data Breaches, AI and More
L
LangChain Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
V
Visual Studio Blog
aimingoo的专栏
aimingoo的专栏
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
J
Java Code Geeks
Stack Overflow Blog
Stack Overflow Blog
IT之家
IT之家
The GitHub Blog
The GitHub Blog
D
Docker
量子位
罗磊的独立博客
腾讯CDC

Show HN

The Two Pillars: Mixer Mode and Meta-Software in the Reorganization of Software Work After AI GitHub - JaiCode08/teleport-env What 1,000+ Harness Experiments Taught Me About Self-Improving Agents Show HN: Liiists, a Markdown-first, iOS and CLI list app SwiperTab – Get this Extension for 🦊 Firefox (en-US) GitHub - kouhxp/fftext: Summarize, explain, fact-check, or translate any text, URL, or file. No GPU. No cloud. One command GitHub - sweetpad-dev/sweetpad: Develop Swift/iOS projects using VSCode GitHub - dogmaticdev/IRON: IRON a.k.a. Intermediate Representation Object Notation is a Interpreter/Database that is used to create Programming Languages. GitHub - sjhalani7/vaen: Package your AI coding harness into a portable .agent file, and share it across repos, teams, & the community without ever having to copy-paste instructions, skills, MCP config, or secrets. Show HN: Gandalf the Grader Show HN: Citadeld – replay any CI failure locally from a single file GitHub - tdortman/cuSBF: High-Performance GPU Super Bloom Filter coral-ai/claude-code-token-xray at main · Coral-Bricks-AI/coral-ai GitHub - ulyssestenn/funes: Funes is a Git-based framework for LLM-managed knowledge work: an AI Librarian ingests raw sources, builds an interlinked Markdown knowledge base, and uses it to produce cited reports, analyses, and other outputs. GitHub - ThatXliner/gah: Git Add Hunk, built for agents to use GitHub - harmont-dev/harmont-cli: Command-line client for the Harmont CI platform GitHub - brooksmcmillin/mcp-authflow: OAuth 2.0 Authorization Server framework for MCP servers GitHub - javaid-codes/audit-supply-chain-agents GitHub - amorey/gochan: A small library of common channel architectures for Go, inspired by Rust GitHub - arifozgun/OpenGem: Free, Open-Source AI API Gateway with Gemini, OpenAI & Anthropic Compatibility in 1 file GitHub - Pranesh950/BioPetals: 🌸 Run BIOxAI models at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading GitHub - cnguyen14/bounty-doctor: Diagnose a GitHub bounty issue before you waste hours: detects honeypot scam repos, AI-bot attempt swarms, and stale contests. Show HN: CoreMCP – MCP Server for On-Prem DBs Show HN: KittyHTML – Render HTML/CSS as an inline image in your terminal GitHub - bingud/filemat: Web-based file manager Show HN: TruthLens – Free multi-signal deepfake image detector GitHub - apexlocal-jz/claude-usage-tray: Windows system-tray app showing your Claude Code rate-limit usage at a glance. Zero deps, ~300 lines of PowerShell. Cross-IDE (works regardless of VS Code, Cursor, plain terminal). Release v0.1.2.1 · kouhxp/yapsnap GitHub - noopolis/moltnet: Self-hostable chat network for AI agents. Pre-built bridges for Claude Code, Codex, and the Claws. Rooms, DMs, history. No Slack bots, no Matrix, no glue code. GitHub - tamerh/enju: Coordinating Humans, AI Agents, and Compute as Peers on a Shared Workflow Graph
Releases · ue-patcher/Code_Stitcher
aihatterer · 2026-06-23 · via Show HN

AI Code Stitcher v2.30

Paste AI-generated code. It just works.

Copy output from Claude, Cursor, ChatGPT, Aider or any other LLM and AI Code Stitcher intelligently applies it to the correct locations in your existing codebase.

stitcher.example.done_2.mp4

What's new in v2.30?

  • Better AIDER / Unified Diff feedback and processing for any hunks already applied,
  • Hint handling feedback and mechanism now more streamlined
  • AST processing on all patches not just full source replacement
  • "Global" and "Module level" now serve as hints for top level CONF
  • Too many things to mention

STITCH VIEWER VERSION 2.00 - File editing suite / Micro IDE

  • The Stitch Viewer operates as an event-driven comparative file/diff viewer that parses directory structures to rebuild and analyze revision histories.
  • The application initializes a low-latency rendering pipeline with hardware-batching to ensure responsive draw calls on modern displays.
  • It retrieves historical backup files alongside live project code and feeds them into an index-matching comparative engine. This comparative engine processes the files line-by-line, translating differences into logical blocks (equal, delete, insert, or replace) and aligning them side-by-side onto a unified virtual line index.
  • This alignment allows developers to inspect alterations across different historical runs and selectively merge or discard localized code segments through dynamic gutter controls.

example_stitch_viewer

The Anti-Agentic Approach

While the industry chases increasingly autonomous AI agents...
Code Stitcher is built on a simpler, more reliable philosophy:

You remain in control.

AI Code Stitcher excels at precisely integrating AI code output into real projects — with smart elastic matching, preview, and undo capabilities.

No mysterious black-box changes. No waiting hours to see what an agent did. Just fast, surgical updates under your direct supervision.

Key Features

  • Intelligent context matching (even when AI uses wrong filenames or incomplete context)
  • Safe preview before applying changes
  • Support for partial updates, function replacements, and larger refactors
  • Works with any editor (VS Code, Neovim, JetBrains, etc.)
  • Runs locally and privately

STITCH VIEWER TECHNICAL SPECIFICATION & USER MANUAL

Classification: Technical Documentation / System Integration & Operations Manual

SECTION 1: Architectural Philosophy & Core Processing Pipelines

The Stitch Viewer is a low-latency, highly specialized graphical comparative
utility designed to reconstruct, analyze, and modify file revision histories in
a non-destructive manner. Engineered as a lightweight comparative environment,
it operates independently of bloated IDE frameworks, providing instantaneous
load times and zero-allocation execution paths.

1.1 Rendering Pipeline Architecture

The application runs on a customized Pygame rendering pipeline optimized for
modern high-DPI displays. By setting the environment variable
SDL_HINT_RENDER_BATCHING to 1 and forcing DPI awareness through permonitorv2,
the rendering engine bypasses typical software-blitting bottlenecks.

+---------------------------------------------------------------------------------+
| Pygame Canvas |
+---------------------------------------------------------------------------------+
| |
| +--------------------+ +---------------------------------------------------+ |
| | Left Tree | | Comparative Panes | |
| | Explorer | | | |
| | | | +-------------------+ +-------------------+ | |
| | [File Header] | | | Historical Pane | | Current Pane | | |
| | └─ Run ID | | | (Read-only, dark) | | (Active/Edit) | | |
| | └─ Revision | | +-------------------+ +-------------------+ | |
| +--------------------+ +---------------------------------------------------+ |
| |
+---------------------------------------------------------------------------------+

Graphics rendering is strictly event-driven. A state-dirtying flag prevents
redundant recalculations, ensuring redrawing only occurs during:

  1. Active user mouse actions or dragging.
  2. Viewport and scroll-offset adjustments.
  3. Active real-time text input.
  4. Active transition animations (sliding layouts).
  5. Polled filesystem change detections.

1.2 The Opcode comparative Engine

The comparative algorithm relies on the sequence matching of lines. By executing
standard sequence matching, the system translates comparative differences into
distinct action blocks:

  • equal: Interleaved lines matching exactly, rendered with syntax-specific
    colors.
  • delete: Line ranges existing inside historical revisions but absent in
    current versions.
  • insert: Line ranges added to the current file but missing from the history.
  • replace: Multi-line blocks transformed or updated between versions.

These blocks map directly onto unified virtual line indices. This ensures that
comparative sections align side-by-side, regardless of offset variations or line
insertions.

SECTION 2: Environment Configuration & Host Integration

The Stitch Viewer integrates natively with host applications (such as the main
CodeStitcher parent program) using low-level operating system APIs,
single-instance mutex locks, and standard input/output structures.

2.1 Command-Line Interface (CLI) Specification

The execution binary accepts several parameters to initialize window positions,
load appropriate monitored directories, and establish secure process-to-process
communication links:

stitch_viewer.exe --from-stitcher --parent-pid 10424 --monitor "C:\Project" --pos "150,150" --height 800 --font "consolas" --font-size 13 --pro

CLI Argument Reference:

  • --from-stitcher: Mandatory. Failsafe validation flag. If absent, the viewer
    instantly terminates execution to prevent orphaned loop execution.
  • --parent-pid : Specifies the Process ID of the calling application.
  • --monitor : Specifies directories to scan for backups and unversioned
    assets. Multiple directories are scanned sequentially.
  • --pos <x,y>: Assigns the default spawning coordinates on the desktop
    workspace.
  • --height : Sets the physical pixel height of the program frame.
  • --font : Overrides the default system typeface for code rendering.
  • --font-size : Configures baseline font sizes.
  • --pro: Unlocks directory tree traversal paths.

2.2 Process Heartbeat & Failsafe Shutdown

To prevent orphaned processes in the memory space, the viewer maintains a
non-blocking parent-process heartbeat monitored on a 120-tick interval (~2
seconds). The background loop uses the Win32 API to check process states:

process = ctypes.windll.kernel32.OpenProcess(0x1000, False, parent_pid)

If the Win32 query return value flags that the parent process has terminated,
the comparative window closes immediately to free hardware assets.

2.3 Single-Instance Mutex Locking

To enforce single-instance comparative execution, the application attempts to
acquire an OS-level Mutex lock:

self.mutex = ctypes.windll.kernel32.CreateMutexW(None, False, "Global\CodeStitcherCodeStitcher_MUTEX")

If the system returns error code 183 (ERROR_ALREADY_EXISTS), a pre-existing
instance of the comparative environment is already running. The new instance
terminates immediately, preventing memory collisions.

SECTION 3: Comparative View Interface & Gutter Mechanics

The layout divides the screen into four functional zones: the directory explorer
tree, the historical revision pane, the interactive current version pane, and
the minimap navigation gutter.

+-----------------------------------------------------------------------------------------+
| Title Bar _ X |
+-----------------------------------------------------------------------------------------+
| Tree Explorer | Comparative View Header: Old File vs New File |
+---------------+-------------------------------------+-----------------------------------+
| | [Line numbers on right] | [Bookmarks] [Line numbers] |
| | | |
| File.py | (Def / Code Text) | (Def / Code Text) |
| └─ Run 01 | | |
| └─ 12:00 | | |
| | | |
| | | |
+---------------+-------------------------------------+-----------------------------------+
| | (Dynamic Status Bar / Linter Warnings) |
+-----------------------------------------------------------------------------------------+

3.1 Compact Directory Explorer Tree

Rendered on the far left, the file explorer uses a compact structure to display
version control nodes:

  1. File Nodes: Base targets sorted chronologically by their last modification
    timestamp.
  2. ...

Read more

AI Code Stitcher v2.20

Paste AI-generated code. It just works.

Copy output from Claude, Cursor, ChatGPT, Aider or any other LLM and AI Code Stitcher intelligently applies it to the correct locations in your existing codebase.

ai_code_stitcher_in_action.mp4

STITCH VIEWER VERSION 2.00 --- --- --- MICRO-IDE. --- --- --- TOTAL FILE EDITING CONVENIENCE

Just press [F11] to run!

  • The Stitch Viewer operates as an event-driven comparative file/diff viewer that parses directory structures to rebuild and analyze revision histories.

  • The application initializes a low-latency rendering pipeline with hardware-batching to ensure responsive draw calls on modern displays.

  • It retrieves historical backup files alongside live project code and feeds them into an index-matching comparative engine. This comparative engine processes the files line-by-line, translating differences into logical blocks (equal, delete, insert, or replace) and aligning them side-by-side onto a unified virtual line index.

  • This alignment allows developers to inspect alterations across different historical runs and selectively merge or discard localized code segments through dynamic gutter controls.

example_stitch_viewer


AI Code Stitcher + Diff Viewer/File Editor

What's new in v2.12?

  • AIDER patch manual rescue interface over-haul. Interactive match for the AIDER search block with intuitive overlay

  • Decorators in Python (@) now handled intelligently.

  • Fuzzy Match Engine Early-Exits: Optimized best match by index searching by introducing an early-exit math heuristic. If the remaining possible lines cannot mathematically achieve the confidence threshold, the search loop aborts immediately. Also added a boundary breakout if the first 15 lines of a search block fail to find any anchor.

  • Heuristic String Pre-Filtering: Integrated fast length-ratio checks along with quick ratio filters to skip expensive SequenceMatcher evaluations on dissimilar lines.

  • Match Window Short-Circuiting: Added an early breakout for unique candidate windows. Once a high-confidence match (≥0.98) is resolved in the highest-density window, evaluations on remaining lower-scoring windows are skipped.

  • Step-Size Fallback Tuning: Adjusted the fallback exact sequence matcher to step through search windows at a wider interval, dropping overall execution times on large files.

  • Indentation & Alignment Safety

  • Copy-Paste Indentation Repairer: Refactored dedenting mechanism to automatically detect when a block header (ending in :) has nested body statements indented disproportionately deeper (>8 spaces). The engine now automatically dedents the body block to a standard 4-space alignment relative to the header.

  • Visual Preview Synchronization: Integrated the indentation repair algorithm directly into headless snippet alignment so that terminal previews match the final on-disk output.

  • Structural Gap/Span Penalty: Introduced a file-span density penalty to filter out false-positives that span across huge sections of unrelated code. The penalty is automatically bypassed if explicit lazy comments (...) are present in the patch.

The Anti-Agentic Approach

While the industry chases increasingly autonomous AI agents...
Code Stitcher is built on a simpler, more reliable philosophy:

You remain in control.

AI Code Stitcher excels at precisely integrating AI code output into real projects — with smart elastic matching, preview, and undo capabilities.

No mysterious black-box changes. No waiting hours to see what an agent did. Just fast, surgical updates under your direct supervision.

Key Features

  • Intelligent context matching (even when AI uses wrong filenames or incomplete context)
  • Safe preview before applying changes
  • Support for partial updates, function replacements, and larger refactors
  • Works with any editor (VS Code, Neovim, JetBrains, etc.)
  • Runs locally and privately

+---------------------------------------------------------------------------------------------------------------------------------------------------+

STITCH VIEWER TECHNICAL SPECIFICATION & USER MANUAL

Classification: Technical Documentation / System Integration & Operations Manual

SECTION 1: Architectural Philosophy & Core Processing Pipelines

The Stitch Viewer is a low-latency, highly specialized graphical comparative
utility designed to reconstruct, analyze, and modify file revision histories in
a non-destructive manner. Engineered as a lightweight comparative environment,
it operates independently of bloated IDE frameworks, providing instantaneous
load times and zero-allocation execution paths.

1.1 Rendering Pipeline Architecture

The application runs on a customized Pygame rendering pipeline optimized for
modern high-DPI displays. By setting the environment variable
SDL_HINT_RENDER_BATCHING to 1 and forcing DPI awareness through permonitorv2,
the rendering engine bypasses typical software-blitting bottlenecks.

+---------------------------------------------------------------------------------+
| Pygame Canvas |
+---------------------------------------------------------------------------------+
| |
| +--------------------+ +---------------------------------------------------+ |
| | Left Tree | | Comparative Panes | |
| | Explorer | | | |
| | | | +-------------------+ +-------------------+ | |
| | [File Header] | | | Historical Pane | | Current Pane | | |
| | └─ Run ID | | | (Read-only, dark) | | (Active/Edit) | | |
| | └─ Revision | | +-------------------+ +-------------------+ | |
| +--------------------+ +---------------------------------------------------+ |
| |
+---------------------------------------------------------------------------------+

Graphics rendering is strictly event-driven. A state-dirtying flag prevents
redundant recalculations, ensuring redrawing only occurs during:

  1. Active user mouse actions or dragging.
  2. Viewport and scroll-offset adjustments.
  3. Active real-time text input.
  4. Active transition animations (sliding layouts).
  5. Polled filesystem change detections.

1.2 The Opcode comparative Engine

The comparative algorithm relies on the sequence matching of lines. By executing
standard sequence matching, the system translates comparative differences into
distinct action blocks:

  • equal: Interleaved lines matching exactly, rendered with syntax-specific
    colors.
  • delete: Line ranges existing inside historical revisions but absent in
    current versions.
  • insert: Line ranges added to the current file but missing from the history.
  • replace: Multi-line blocks transformed or updated between versions.

These blocks map directly onto unified virtual line indices. This ensures that
comparative sections align side-by-side, regardless of offset variations or line
insertions.

SECTION 2: Environment Configuration & Host Integration

The Stitch Viewer integrates natively with host applications (such as the main
CodeStitcher parent program) using low-level operating system APIs,
single-instance mutex locks, and standard input/output structures.

2.1 Command-Line Interface (CLI) Specification

The execution binary accepts several parameters to initialize window positions,
load appropriate monitored directories, and establish secure process-to-process
communication links:

stitch_viewer.exe --from-stitcher --parent-pid 10424 --monitor "C:\Project" --pos "150,150" --height 800 --font "consolas" --font-size 13 --pro

CLI Argument Reference:

  • --from-stitcher: Mandatory. Failsafe validation flag. If absent, the viewer
    instantly terminates execution to prevent orphaned loop execution.
  • --parent-pid : Specifies the Process ID of the calling application.
  • --monitor : Specifies directories to scan for backups and unversioned
    assets. Multiple directories are scanned sequentially.
  • --pos <x,y>: Assigns the default spawning coordinates on the desktop
    workspace.
  • --height : Sets the physical pixel height of the program frame.
  • --font : Overrides the default system typeface for code rendering.
  • --font-size : Configures baseline font sizes.
  • --pro: Unlocks directory tree traversal paths.

2.2 Process Heartbeat & Failsafe Shutdown

To prevent orphaned processes in the memory space, the viewer maintains a
non-blocking parent-process heartbeat monitored on a 120-tick interval (~2
seconds). The background loop uses the Win32 API to check process states:

process = ctypes.windll.kernel32.OpenProcess(0x1000, False, parent_pid)

If the Win32 query return value flags that the parent process has terminated,
the comparative window closes immediately to free hardware assets.

2.3 Single-Instance Mutex Locking

To enforce single-instance comparative execution, the application attempts to
acquire an OS-level Mutex lock:

self.mutex = ctypes.windll.kernel32.CreateMutexW(None, False, "Global\CodeStitcherCodeStitcher_MUTEX")

If the system returns error code 183 (ERROR_ALREADY_EXISTS), a pre-existing
instance of the comparat...

Read more

AI Code Stitcher v2.02

VERSION 2.0 NOW AVAILABLE!!

Paste AI-generated code. It just works.

Copy output from Claude, Cursor, ChatGPT, Aider or any other LLM and AI Code Stitcher intelligently applies it to the correct locations in your existing codebase.

ai_code_stitcher_in_action.mp4

What's new in v2.0?

  • Fuzzy Match Engine Early-Exits: Optimized best match by index searching by introducing an early-exit math heuristic. If the remaining possible lines cannot mathematically achieve the confidence threshold, the search loop aborts immediately. Also added a boundary breakout if the first 15 lines of a search block fail to find any anchor.

  • Heuristic String Pre-Filtering: Integrated fast length-ratio checks along with quick ratio filters to skip expensive SequenceMatcher evaluations on dissimilar lines.

  • Match Window Short-Circuiting: Added an early breakout for unique candidate windows. Once a high-confidence match (≥0.98) is resolved in the highest-density window, evaluations on remaining lower-scoring windows are skipped.

  • Step-Size Fallback Tuning: Adjusted the fallback exact sequence matcher to step through search windows at a wider interval, dropping overall execution times on large files.

  • Indentation & Alignment Safety

  • Copy-Paste Indentation Repairer: Refactored dedenting mechanism to automatically detect when a block header (ending in :) has nested body statements indented disproportionately deeper (>8 spaces). The engine now automatically dedents the body block to a standard 4-space alignment relative to the header.

  • Visual Preview Synchronization: Integrated the indentation repair algorithm directly into headless snippet alignment so that terminal previews match the final on-disk output.

  • Structural Gap/Span Penalty: Introduced a file-span density penalty to filter out false-positives that span across huge sections of unrelated code. The penalty is automatically bypassed if explicit lazy comments (...) are present in the patch.

v2.02 Notes:

-Fixed file hint extraction from overly verbose LLM comments. Now will bypass file prompt automating all patch types.
-Cancel now reliably cancels out of entire patch application process.

The Anti-Agentic Approach

While the industry chases increasingly autonomous AI agents...
Code Stitcher is built on a simpler, more reliable philosophy:

You remain in control.

AI Code Stitcher excels at precisely integrating AI code output into real projects — with smart elastic matching, preview, and undo capabilities.

No mysterious black-box changes. No waiting hours to see what an agent did. Just fast, surgical updates under your direct supervision.

Key Features

  • Intelligent context matching (even when AI uses wrong filenames or incomplete context)
  • Safe preview before applying changes
  • Support for partial updates, function replacements, and larger refactors
  • Works with any editor (VS Code, Neovim, JetBrains, etc.)
  • Runs locally and privately

Head to https://sharples3.gumroad.com/l/codestitcher for Pro Edition enquires.

Please note that Windows Defender SmartScreen may flag this program/install as 'potentially dangerous'. This is due to the tool being new and limited install base. We promise there is no spy/malware in this program.

AI Code Stitcher v2.01

VERSION 2.0 NOW AVAILABLE!!

Paste AI-generated code. It just works.

Copy output from Claude, Cursor, ChatGPT, Aider or any other LLM and AI Code Stitcher intelligently applies it to the correct locations in your existing codebase.

ai_code_stitcher_in_action.mp4

What's new in v2.0?

  • Fuzzy Match Engine Early-Exits: Optimized best match by index searching by introducing an early-exit math heuristic. If the remaining possible lines cannot mathematically achieve the confidence threshold, the search loop aborts immediately. Also added a boundary breakout if the first 15 lines of a search block fail to find any anchor.

  • Heuristic String Pre-Filtering: Integrated fast length-ratio checks along with quick ratio filters to skip expensive SequenceMatcher evaluations on dissimilar lines.

  • Match Window Short-Circuiting: Added an early breakout for unique candidate windows. Once a high-confidence match (≥0.98) is resolved in the highest-density window, evaluations on remaining lower-scoring windows are skipped.

  • Step-Size Fallback Tuning: Adjusted the fallback exact sequence matcher to step through search windows at a wider interval, dropping overall execution times on large files.

  • Indentation & Alignment Safety

  • Copy-Paste Indentation Repairer: Refactored dedenting mechanism to automatically detect when a block header (ending in :) has nested body statements indented disproportionately deeper (>8 spaces). The engine now automatically dedents the body block to a standard 4-space alignment relative to the header.

  • Visual Preview Synchronization: Integrated the indentation repair algorithm directly into headless snippet alignment so that terminal previews match the final on-disk output.

  • Structural Gap/Span Penalty: Introduced a file-span density penalty to filter out false-positives that span across huge sections of unrelated code. The penalty is automatically bypassed if explicit lazy comments (...) are present in the patch.

The Anti-Agentic Approach

While the industry chases increasingly autonomous AI agents...
Code Stitcher is built on a simpler, more reliable philosophy:

You remain in control.

AI Code Stitcher excels at precisely integrating AI code output into real projects — with smart elastic matching, preview, and undo capabilities.

No mysterious black-box changes. No waiting hours to see what an agent did. Just fast, surgical updates under your direct supervision.

Key Features

  • Intelligent context matching (even when AI uses wrong filenames or incomplete context)
  • Safe preview before applying changes
  • Support for partial updates, function replacements, and larger refactors
  • Works with any editor (VS Code, Neovim, JetBrains, etc.)
  • Runs locally and privately

Head to https://sharples3.gumroad.com/l/codestitcher for Pro Edition enquires.