- Import shared Chrome test helpers
- Add test_singlefile_with_chrome_session() to verify CDP connection
- Add test_singlefile_disabled_skips() for config testing
- Update existing test to use get_test_env()
- Add get_machine_type() to chrome_test_helpers.py
- Update get_test_env() to include MACHINE_TYPE
- Refactor test_chrome.py to import from shared helpers
- Removes ~50 lines of duplicate code
- Add setup_test_env, launch_chromium_session, kill_chromium_session
to chrome_test_helpers.py for extension tests
- Add chromium_session context manager for cleaner test code
- Refactor ublock, istilldontcareaboutcookies, twocaptcha tests to use
shared helpers (~450 lines removed)
- Refactor screenshot, dom, pdf tests to use shared get_test_env
and get_lib_dir (~60 lines removed)
- Net reduction: 228 lines of duplicate code
This change consolidates duplicated logic between chrome_utils.js and
extension installer hooks, as well as between Python plugin tests:
JavaScript changes:
- Add getExtensionsDir() to centralize extension directory path calculation
- Add installExtensionWithCache() to handle extension install + cache workflow
- Add CLI commands for new utilities
- Refactor all 3 extension installers (ublock, istilldontcareaboutcookies,
twocaptcha) to use shared utilities, reducing each from ~115 lines to ~60
- Update chrome_launch hook to use getExtensionsDir()
Python test changes:
- Add chrome_test_helpers.py with shared Chrome session management utilities
- Refactor infiniscroll and modalcloser tests to use shared helpers
- setup_chrome_session(), cleanup_chrome(), get_test_env() now centralized
- Add chrome_session() context manager for automatic cleanup
Net result: ~208 lines of code removed while maintaining same functionality.
Simplifies the comma-separated parsing logic to:
- If value contains '[', parse as JSON array
- Otherwise, parse as comma-separated values
This prevents incorrect splitting of arguments containing internal commas
when there's only one argument. For arguments with commas, users should
use JSON format: CHROME_ARGS='["--arg1,val", "--arg2"]'
Also exports getEnvArray in module.exports for consistency.
Co-authored-by: Nick Sweeting <pirate@users.noreply.github.com>
Multiple hooks in the same plugin directory were overwriting each
other's stdout.log, stderr.log, hook.pid, and cmd.sh files. Now each
hook uses filenames prefixed with its hook name:
- on_Snapshot__20_chrome_tab.bg.stdout.log
- on_Snapshot__20_chrome_tab.bg.stderr.log
- on_Snapshot__20_chrome_tab.bg.pid
- on_Snapshot__20_chrome_tab.bg.sh
Updated:
- hooks.py run_hook() to use hook-specific names
- core/models.py cleanup and update_from_output methods
- Plugin scripts to no longer write redundant hook.pid files
<!-- IMPORTANT: Do not submit PRs with only formatting / PEP8 / line
length changes. -->
# Summary
<!--e.g. This PR fixes ABC or adds the ability to do XYZ...-->
# Related issues
<!-- e.g. #123 or Roadmap goal #
https://github.com/pirate/ArchiveBox/wiki/Roadmap -->
# Changes these areas
- [ ] Bugfixes
- [ ] Feature behavior
- [ ] Command line interface
- [ ] Configuration options
- [ ] Internal architecture
- [ ] Snapshot data layout on disk
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Prevented hook file collisions by giving each hook its own stdout,
stderr, pid, and cmd filenames. This fixes mixed logs and ensures
correct cleanup and status checks when multiple hooks run in the same
plugin directory.
- **Bug Fixes**
- hooks.py: write hook-specific stdout/stderr/pid/cmd files and exclude
them from new_files; derive cmd.sh from pid for safe kill.
- core/models.py: read hook-specific logs; exclude hook output files
when computing outputs; cleanup and background detection use *.pid.
- Plugins: stop writing redundant hook.pid files; minor chrome utils
cleanup.
<sup>Written for commit 754b096193.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
Multiple hooks in the same plugin directory were overwriting each
other's stdout.log, stderr.log, hook.pid, and cmd.sh files. Now
each hook uses filenames prefixed with its hook name:
- on_Snapshot__20_chrome_tab.bg.stdout.log
- on_Snapshot__20_chrome_tab.bg.stderr.log
- on_Snapshot__20_chrome_tab.bg.pid
- on_Snapshot__20_chrome_tab.bg.sh
Updated:
- hooks.py run_hook() to use hook-specific names
- core/models.py cleanup and update_from_output methods
- Plugin scripts to no longer write redundant hook.pid files
- Add comprehensive default CHROME_ARGS in config.json with 55+ flags
for deterministic rendering, security, performance, and UI suppression
- Update chrome_utils.js launchChromium() to read CHROME_ARGS and
CHROME_ARGS_EXTRA from environment variables (set by get_config())
- Add getEnvArray() helper to parse JSON arrays or comma-separated
strings from environment variables
- Separate args into three categories:
1. baseArgs: Static flags from CHROME_ARGS config (configurable)
2. dynamicArgs: Runtime-computed flags (port, sandbox, headless, etc.)
3. extraArgs: User overrides from CHROME_ARGS_EXTRA
- Add CHROME_SANDBOX config option to control --no-sandbox flag
Args are now configurable via:
- config.json defaults
- ArchiveBox.conf file
- Environment variables
- Per-crawl/snapshot config overrides
- Add _derive_persona_paths() in configset.py to automatically derive
CHROME_USER_DATA_DIR and CHROME_EXTENSIONS_DIR from ACTIVE_PERSONA
when not explicitly set. This allows plugins to use these paths
without knowing about the persona system.
- Update chrome_utils.js launchChromium() to accept userDataDir option
and pass --user-data-dir to Chrome. Also cleans up SingletonLock
before launch.
- Update killZombieChrome() to clean up SingletonLock files from all
persona chrome_user_data directories after killing zombies.
- Update chrome_cleanup() in misc/util.py to handle persona-based
user data directories when cleaning up stale Chrome state.
- Simplify on_Crawl__20_chrome_launch.bg.js to use CHROME_USER_DATA_DIR
and CHROME_EXTENSIONS_DIR from env (derived by get_config()).
Config priority flow:
ACTIVE_PERSONA=WorkAccount (set on crawl/snapshot)
-> get_config() derives:
CHROME_USER_DATA_DIR = PERSONAS_DIR/WorkAccount/chrome_user_data
CHROME_EXTENSIONS_DIR = PERSONAS_DIR/WorkAccount/chrome_extensions
-> hooks receive these as env vars without needing persona logic
- Move hardcoded default args from Python to config.json YTDLP_ARGS
- Add get_ytdlp_args() function to read from YTDLP_ARGS env var
- Keep format arg with max_size in code (depends on YTDLP_MAX_SIZE)
- YTDLP_ARGS can be overridden as JSON array in environment
Replace old `output` field with new fields across the codebase:
- output_str: Human-readable output summary
- output_json: Structured metadata (optional)
- output_files: Dict of output files with metadata
- output_size: Total size in bytes
- output_mimetypes: CSV of file mimetypes
Files updated:
- api/v1_core.py: Update MinimalArchiveResultSchema to expose new fields
- api/v1_core.py: Update ArchiveResultFilterSchema to search output_str
- cli/archivebox_extract.py: Use output_str in CLI output
- core/admin_archiveresults.py: Update admin fields, search, and fieldsets
- core/admin_archiveresults.py: Fix output_html variable name bug in output_summary
- misc/jsonl.py: Update archiveresult_to_jsonl() to include new fields
- plugins/extractor_utils.py: Update ExtractorResult helper class
The embed_path() method already uses output_files and output_str,
so snapshot detail page and template tags work correctly.
- All install hooks now respect their respective XYZ_BINARY env vars
(e.g., WGET_BINARY, CHROME_BINARY, YTDLP_BINARY, etc.)
- Support both absolute paths (/usr/bin/wget2) and binary names (wget2)
- Dynamic bin_name used in Dependency JSONL output
- Updated 11 install hooks to follow the new pattern
- Mark checklist items as complete in TODO_hook_architecture.md
- Rename 13 on_Crawl__00_validate_* hooks to on_Crawl__00_install_*
- This better reflects what these hooks actually do (check/install binaries)
- Update TODO_hook_architecture.md to reflect renamed hooks
