Skip to content

TESTING_CHECKLIST.md

Latest commit

 

History

History
431 lines (324 loc) · 15.2 KB

TESTING_CHECKLIST.md

File metadata and controls

431 lines (324 loc) · 15.2 KB

guIDE v2.0.0 — Full Feature Testing Checklist

Generated: 2026-02-20 | Use this after installing guIDE-Setup-2.0.0.exe

Log Monitoring: After launching the app, tail the log file in a PowerShell window:

Get-Content "$env:APPDATA\guide-ide\logs\guide-main.log" -Wait -Tail 80

The log default is now DEBUG level — every internal call is captured.

Before You Start

  • Install dist-electron\guIDE-Setup-2.0.0.exe (388.9 MB)
  • Open PowerShell and start the log monitor: 
    Get-Content "$env:APPDATA\guide-ide\logs\guide-main.log" -Wait -Tail 80
  • Open a second PowerShell window for running test scripts (Python/JS etc.)
  • Have a test project folder ready (anything with .js, .py, .html files)

1. App Launch & UI

  • App launches without crash or splash error
  • Custom titlebar shows "guIDE" text in Audiowide brand font
  • Status bar visible at bottom (CPU %, RAM %)
  • Activity bar visible on left (icons for Explorer, Search, Git, Debug, Extensions, Browser)
  • Welcome screen appears in editor area (keyboard shortcuts listed)
  • F11 — toggles fullscreen
  • CPU & RAM gauges update in real-time in status bar

2. File Explorer (Ctrl+Shift+E)

  • Panel opens/closes with Ctrl+Shift+E
  • Open a folder — Ctrl+O → pick a project dir → file tree populates
  • Folder tree is collapsible/expandable
  • Click a file → opens in editor tab
  • Right-click file → context menu: Rename, Delete, New File, New Folder, Copy Path
  • Create a new file via right-click → appears in tree
  • Create a new folder via right-click → appears in tree
  • Rename a file → name changes in tree and tab
  • Delete a file → removed from tree
  • Drag a file to a new folder → moves successfully
  • Drop a file from Windows Explorer into the tree → file appears
  • Grid view toggle (if present) — switches to icon grid layout
  • Multi-select: Ctrl+Click multiple files
  • Ctrl+C / Ctrl+V to copy/paste a file

3. Core Editor

  • Open a file — correct syntax highlighting for language
  • Edit content — "●" dirty indicator on tab
  • Ctrl+S — saves file, dirty indicator disappears
  • Ctrl+Shift+S — saves all open files
  • Ctrl+N — creates a new untitled file
  • Ctrl+P — Quick Open: fuzzy-search by filename, press Enter to open
  • Ctrl+\ — Split Editor: two panes side-by-side
  • Middle-click a tab — closes it
  • Drag tab to reorder — order changes in tab bar
  • Minimap visible on the right side of editor
  • Code folding: click gutter arrow to collapse a function/block
  • Bracket matching: click { → matching } is highlighted
  • Ctrl+Shift+[ — selects all content between matching brackets
  • Ctrl+F — opens in-file Find bar
    • Search term highlights in editor
    • Regex mode toggle works
    • Case-sensitive toggle works
    • Whole-word toggle works
  • Ctrl+H — opens Find & Replace bar
    • Replace single match
    • Replace all matches
  • Ctrl+Shift+P — Command Palette opens; type a command name and execute it

4. Themes (Ctrl+Shift+P → "Switch Theme")

Test each of the 8 themes:

  • Dark Default
  • Monokai
  • Dracula
  • Nord
  • Solarized Dark
  • GitHub Dark
  • Light
  • Catppuccin Mocha

Verify: editor, sidebar, and status bar all change color.

5. AI Chat (Ctrl+J)

Basic Chat

  • Panel opens/closes with Ctrl+J
  • Type a message and press Enter → response streams in
  • Streaming renders token-by-token, no layout jumps
  • Context window indicator (SVG ring) shows token usage near input
  • Clear chat button works

Cloud Providers

Test at least 2 cloud models:

  • Gemini 2.5 Flash — send a message, get a response
  • Groq Llama — send a message, get a response
  • Cerebras — send a message, get a response
  • OpenRouter — send a message (verify model list loads from API)

Model Switching

  • Open model picker dropdown
  • Switch from one cloud model to another while idle
  • If a local GGUF model is installed: load it, send a message, wait for GPU/CPU inference

AI Capabilities

  • Mermaid diagrams — ask: "Draw me a mermaid flowchart of how a login system works" → collapsible diagram renders in chat
  • Code blocks — ask for a Python function → syntax-highlighted code block appears
  • Plan Mode toggle — enable Plan Mode, ask AI to do something, verify it produces a plan first
  • Thinking models — if a thinking model is selected (e.g. DeepSeek), verify <think> block is collapsible

6. Inline Chat (Ctrl+I)

  • Open a code file and position cursor inside a function
  • Press Ctrl+I → inline chat overlay appears at cursor
  • Type an edit request (e.g. "add error handling to this function")
  • AI response appears as a diff in the editor
  • Accept button applies the change
  • Reject button reverts to original
  • Escape closes the inline chat

7. Next Edit Suggestions (Ghost Text)

  • Open a TypeScript or JavaScript file
  • Type something or pause after an edit
  • Ghost text appears in dim/colored text suggesting the next edit
  • Press Tab to accept the suggestion
  • Ghost text disappears and suggestion is applied

8. MCP Tools (AI Tool Use)

Ask the AI each of the following and verify the tool executes correctly (watch the log for tool call entries):

File Tools

  • "List all files in the current project" → calls list_directory, returns file list
  • "Read the contents of [filename]" → calls read_file, shows content
  • "Create a new file called test-mcp.txt with the text 'hello world'" → calls write_file, file appears
  • "Edit test-mcp.txt and add another line" → calls edit_file
  • "Delete test-mcp.txt" → calls delete_file, file gone
  • "Search for all TODO comments in my project" → calls grep_search
  • "Show me the project structure" → calls get_project_structure

Terminal Tool

  • "Run the command: echo hello from mcp" → calls run_command, output shown

Git Tools

  • "What are my uncommitted changes?" → calls git_status
  • "Show me the git log" → calls git_log
  • "What's changed in file X?" → calls git_diff

Web & Memory Tools

  • "Search the web for: latest JavaScript frameworks 2026" → calls web_search
  • "Remember that my project name is guIDE test" → calls save_memory
  • "What do you remember about my project?" → calls get_memory

Browser Tools

  • "Navigate to https://example.com" → calls browser_navigate, browser tab opens
  • "Click the More Information link on this page" → calls browser_click
  • "Get the content of this page" → calls browser_get_content

9. Agentic Mode

  • Ask a multi-step coding task: "Create a new Python file, write a Flask hello-world app in it, then show me how to run it" — verify AI chains multiple tool calls (write_file, list_directory, etc.) autonomously
  • Verify iteration counter doesn't exceed 100 (watch status in chat)
  • Ask to abort mid-run → generation stops
  • T2 guardrail test: Say "DO NOT create any files. Just show me the code for a React counter component" → verify NO write_file tool call is made (AI shows code only)

10. Background Agents

  • Type /agent in the chat input
  • Provide a task for the background agent
  • Verify agent status indicator appears in chat
  • Agent completes the task and reports back

11. Code Runner (F5)

Test at least 3 different file types:

  • Python (.py) — create a script with print("Hello Python"), press F5 → terminal opens, output shown
  • JavaScript (.js) — create console.log("Hello JS"), press F5 → terminal opens, output shown
  • TypeScript (.ts) — create console.log("Hello TS"), press F5 → runs via npx tsx
  • Bash (.sh) — create echo "Hello Shell", press F5 → runs via bash
  • HTML (.html) — press F5 → HTML preview opens (not code runner)

12. File Preview System

Toggle preview with the Eye icon on the tab or the Play button:

  • .html → HTML preview renders in iframe, links work
  • .md → Markdown preview with styled headers, code blocks, tables
  • .svg → SVG preview with zoom controls (25%-400%), background toggle
  • .json → Collapsible tree view, keys/values colored
  • .csv → Sortable table with row/column count
  • .yaml → Syntax-highlighted YAML view
  • .png / .jpg → Image displays inline

13. Terminal (Ctrl+`)

  • Terminal opens with Ctrl+`
  • Shell starts in the current project folder
  • Can run commands: pwd, ls, python --version
  • Multiple tabs — click + to open a second terminal tab
  • Rename a terminal tab — double-click the tab name
  • Close a terminal tab — click the ×
  • Terminal IntelliSense — start typing a command and verify AI suggestions appear

14. Search (Ctrl+Shift+F)

  • Panel opens with Ctrl+Shift+F
  • Type a search term → results appear across all project files
  • Click a result → jumps to that line in the editor
  • Regex mode toggle works (e.g. search \btodo\b)
  • Case-sensitive toggle works
  • Whole-word toggle works
  • Semantic Search toggle — switch to semantic mode, run a query

15. Git / Source Control (Ctrl+Shift+G)

  • In a Git repo, open Source Control panel
  • Changed files list is accurate
  • Click a changed file → diff view opens with +/- highlights
  • Stage a file (click + icon)
  • Write a commit message and commit
  • Branch indicator in status bar shows current branch
  • Create a new branch from the branch dropdown
  • Checkout an existing branch
  • Git Blame — open a file, verify inline blame decorations (author, date) appear per line

16. Debug Panel (Ctrl+Shift+D)

Node.js

  • Open a JavaScript file with a simple loop
  • Set a breakpoint (click the gutter)
  • Press F5 to start debug → execution pauses at breakpoint
  • Variables panel shows local variables with correct values
  • Call stack panel shows the current call chain
  • Step over (F10), step into (F11), step out (Shift+F11)
  • Continue (F5) resumes execution
  • Remove breakpoint

Python

  • Open a Python script
  • Set a breakpoint
  • Start debugging → pauses at breakpoint (uses debugpy)
  • Variables panel shows Python values

17. Browser Integration

  • Open Browser panel via activity bar (or via AI using browser_navigate)
  • URL bar visible — type https://google.com and press Enter
  • Page loads inside the panel
  • Back/Forward/Reload buttons work
  • Ask AI: "Go to google.com and tell me what's on the page" → AI navigates and reads content
  • Ask AI: "Search 'electron js' on the current page" → AI takes browser actions

18. RAG System

  • Open a project with several code files
  • RAG auto-indexing message appears in log (RAG, Indexed)
  • Ask AI a question about the codebase: "Where is the main configuration object defined in this project?" — AI cites the correct file without being told

19. Speech (if microphone available)

  • Click the microphone button in the chat input
  • Speak a prompt → Groq Whisper transcribes it → text appears in input
  • TTS: right-click an AI response and select "Read aloud" (or equivalent button)

20. Image / Vision Input

  • Drag an image (.png or .jpg) into the chat input → thumbnail appears
  • Paste a screenshot with Ctrl+V → thumbnail appears
  • Click the paperclip icon → file picker → select an image → thumbnail appears
  • Send the message with a vision-capable model (Gemini, GPT-4o) → AI describes the image

21. MCP Server Management (Ctrl+Shift+X)

  • Extensions / MCP Servers panel opens
  • Existing MCP servers listed
  • Can toggle a server on/off
  • Add a new external MCP server (fill in name + command)
  • Verify new server appears in AI tool calls

22. Paste with Auto-Imports

  • Open a React .tsx file
  • Paste code that uses useState without importing it
  • Verify import { useState } from 'react' is automatically inserted at the top

23. Developer Console

  • Open the Developer Console panel (button in chat area or Settings menu)
  • Verify it shows real-time LLM/GPU/AI backend logs
  • Color-coded entries (info/warn/error)
  • Entry count capped at ~500, old entries drop off
  • Close the console panel

24. System & Performance

  • GPU detection — status bar shows GPU name + VRAM (if NVIDIA/AMD present)
  • GPU/CPU indicator — when local model is active, icon shows green monitor (GPU) or red CPU
  • Tokens/sec — visible in status bar during local inference
  • Context ring — fills proportionally as conversation grows

25. Licensing

  • License key screen appears on first launch (if not activated)
  • Enter a valid license key → activates
  • App stores machine fingerprint
  • Help → About shows version 2.0.0

26. Custom Instructions (.prompt.md)

  • Create a .prompt.md file in the project root with text: Always respond like a pirate.
  • Send any message to AI → response should be pirate-themed
  • Delete .prompt.md → normal AI responses return

27. Keyboard Shortcuts — Full Sweep

Shortcut Expected Action ✓/✗
Ctrl+Shift+P Command Palette opens
Ctrl+B Sidebar toggles
Ctrl+` Terminal toggles
Ctrl+J AI Chat toggles
Ctrl+I Inline Chat at cursor
Ctrl+Shift+E File Explorer
Ctrl+Shift+F Search in files
Ctrl+Shift+G Git / Source Control
Ctrl+Shift+D Debug Panel
Ctrl+Shift+X Extensions / MCP
Ctrl+S Save current file
Ctrl+Shift+S Save all files
Ctrl+N New file
Ctrl+O Open file/folder
Ctrl+P Quick Open
Ctrl+F Find in file
Ctrl+H Find & Replace
Ctrl+\ Split Editor
Ctrl+Shift+[ Select bracket contents
F5 Run current file
Tab Accept ghost text suggestion
Escape Close inline chat/search
F11 Toggle fullscreen

Log Monitoring Reference

# Live tail — run BEFORE launching app:
Get-Content "$env:APPDATA\guide-ide\logs\guide-main.log" -Wait -Tail 80

# Filter errors only:
Get-Content "$env:APPDATA\guide-ide\logs\guide-main.log" -Wait -Tail 30 | Where-Object { $_ -match 'ERROR|FATAL|WARN' }

# Filter AI/LLM entries:
Get-Content "$env:APPDATA\guide-ide\logs\guide-main.log" -Wait -Tail 30 | Where-Object { $_ -match 'LLM|Chat|Tool|Agent|RAG' }

# View most recent session only:
Get-Content "$env:APPDATA\guide-ide\logs\guide-main.log" | Select-Object -Last 200

Issue Log

Use this section to record anything that fails during testing:

# Feature Observed Expected Severity

Total test items: ~120 checks across 27 categories