Skip to content

Latest commit

 

History

History
58 lines (41 loc) · 2.41 KB

tasks.md

File metadata and controls

58 lines (41 loc) · 2.41 KB

Tasks

The Package Control website has a minimal python-based script runner that may be invoked by calling:

python task.py [task_name]

The [task_name] if the filename of one of the files in app/tasks/, without the .py extension. The following is the current list of valid tasks:

  • cleanup_renames - updates various tables to move stats from an old name to a new one.

  • compute_package_hashes - compute the SHA-256 hash for the various packages that are made available for download via non-secure HTTP

  • crawl - uses the Package Control channel info from config/crawler.yml to crawl GitHub, BitBucket and other repositories for package data. Looks for packages that have not been seen in the past hour. Only crawls 200 sources per run.

  • deploy - uses SSH to deploy the latest copy of code on GitHub to the server

  • gather_system_stats - pulls in updated info about the number of users, packages, labels and installs, bytes served, etc for the previous day.

  • generate_channel_json - builds the channel.json file that contains the 2.0 schema version channel info used by Package Control 2.x.

  • generate_channel_v3_json - builds the channel_v3.json file that contains the 3.0.0 schema version channel info used by Package Control 3.x.

  • generate_legacy_channel_json - build the repositories.json file that contains the 1.2 schema version channel info used by Package Control 1.x.

  • generate_signature - generate an ECDSA signature of the Package Control.sublime-package file. It is written using ASCII armor of BEGIN PACKAGE CONTROL SIGNATURE.

  • mark_removed - looks for packages that haven't been seen in a long time and marks them as removed, so they don't should in search results

  • parse_log_files - reads through the nginx log files for the previous day to extract statistics about the number of requests and bytes served.

  • rebuild_stats - completely reparses the (giant) usage table and recreated all install statistics about packages. In production, the usage table has 40M entries (as of August 2013), so this can take quite a while. This is not normally used unless an error is detected in calulcating stats.

  • refresh_package_stats - recaclulates the installs and trending ranks for the packages.

  • update_package_control_lib - grabs the latest copy of the package_control lib from Package Control