Advanced concepts

• Customizing the output of Mercurial
  • Using precanned output styles
  • Commands that support styles and templates
  • The basics of templating
  • Common template keywords
  • Escape sequences
  • Filtering keywords to change their results
    • Combining filters
    • Adding logic
    • Setting a default template
  • From templates to commands
    • The simplest of template aliases
  • Templates by example
    • Identifying mistakes in templates
    • Uniquely identifying a repository
    • Listing files on multiple lines
    • Mimicking Subversion's output
  • Parsing Mercurial output
  • Overview of the template system
• Handling repository events with hooks
  • An overview of hooks in Mercurial
  • Hooks and security
    • Hooks are run with your privileges
    • Hooks do not propagate
    • Hooks can be overridden
    • Ensuring that critical hooks are run
  • A short tutorial on using hooks
    • Performing multiple actions per event
    • Controlling whether an activity can proceed
  • Writing your own hooks
    • Choosing how your hook should run
    • Hook parameters
    • Hook return values and activity control
    • Writing an external hook
    • Telling Mercurial to use an in-process hook
    • Writing an in-process hook
  • Some hook examples
    • Writing meaningful commit messages
    • Checking for trailing whitespace
  • Bundled hooks
    • acl—access control for parts of a repository
  • Configuring the acl hook
  • Testing and troubleshooting
    • bugzilla—integration with Bugzilla
  • Configuring the bugzilla hook
  • Mapping committer names to Bugzilla user names
  • Configuring the text that gets added to a bug
  • Testing and troubleshooting
    • notify—send email notifications
  • Configuring the notify hook
  • Testing and troubleshooting
  • Information for writers of hooks
    • In-process hook execution
    • External hook execution
    • Finding out where changesets come from
  • Sources of changesets
  • Where changes are going—remote repository URLs
  • Hook reference
    • changegroup—after remote changesets added
    • commit—after a new changeset is created
    • incoming—after one remote changeset is added
    • outgoing—after changesets are propagated
    • prechangegroup—before starting to add remote changesets
    • precommit—before starting to commit a changeset
    • preoutgoing—before starting to propagate changesets
    • pretag—before tagging a changeset
    • pretxnchangegroup—before completing addition of remote changesets
    • pretxncommit—before completing commit of new changeset
    • preupdate—before updating or merging working directory
    • tag—after tagging a changeset
    • update—after updating or merging working directory
• Changing history
  • The good and bad of history rewriting
    • Why change history?
    • Why avoid changing history?
  • Adapting the last changeset
  • Move changesets around in history using rebase
  • Edit a group of changesets using histedit
  • Safely changing history
    • Overriding default behaviour
  • Conclusion
• Scaling Mercurial
  • Scalability issues
  • Scaling up to many changesets with remotefilelog
  • Handle large binaries with the largefiles extension
  • Scaling repositories with many files
  • Improving server scalability and cloning speed
    • Generating bundle files
    • Specifying correct manifest lines
  • Scaling repositories with many branches
• Behind the scenes
  • Mercurial's historical record
    • Tracking the history of a single file
    • Managing tracked files
    • Recording changeset information
    • Relationships between revisions
  • Safe, efficient storage
    • Efficient storage
    • Safe operation
    • Fast retrieval
  • Aside: the influence of video compression
    • Identification and strong integrity
  • Revision history, branching, and merging
  • The working directory
    • What happens when you commit
    • Creating a new head
    • Merging changes
    • Merging and renames
  • Other interesting design features
    • Clever compression
  • Network recompression
    • Read/write ordering and atomicity
    • Concurrent access
  • Safe dirstate access
    • Avoiding seeks
    • Other contents of the dirstate