Update to 1.1 final

Changes since 1.0.X:
- new simpler and more reliable solver
- shared libraries are now always tracked
- ssh:// is supported as a protocol to distribute packages (needs pkg 1.1+ on
  the server hosting the packages)
- multirepository is no longer considered experimental and works by default.
- incremental update of the catalog (only if the repository was created by pkg
  1.1+)
- simplification of the public API
- stabilisation of the public API (we will now try to keep it stable and if
  change are needed there will be deprecation time before removal of some old
  functions)
- new experimental pkg convert (can convert from and to legacy pkg database)
  pkg2ng now uses pkg convert (still recommanded to use pkg2ng)
- new pkg lock/unlock to prevent any manipulation of a given package (no
  upgrade,delete,etc)
- improved UI (now you can see the progress of an upgrade what is left to be
  done)
- new pkg annotation to allow one to add annotations (free form key/value) to a
  package)
- pkg audit is now able to directly parse the vuxml native format and not only
  the compact version
- pkg -vv now shows all available options and their current settings
- pkg -vvv now shows a description of all the available options
- pkg info now automatically considers the query as globbing if * is in the
  requested pattern
- new hook plugin interface (allows users to create hooks that get called at
  anytime during and upgrade/installation/deletion of a package)
- new cmd plugin interface (allows users to create new sub command available for
  pkg)
- pkg register can now register a port installation in the legacy database
  format
- repository can be defined in simple yaml files
- Explain why a package is being reinstalled
- A package can now be marked to only be upgraded from a given repository via
  annotations
- install and upgrade will show from which repository packages are taken from
- old/unused 'informations' field has been removed in favor of using
annotations
- pkg repo produce a repo.txz in pkg 1.0 (legacy) fromat

Internal:
- massive usage of hash tables (uthash), which simplifies a lot of the code,
  and improves performances
- lots of optimisation in plist and manifest parsing
- lots of optimisation in loading packages (mmap used when possible)
- lots of cleanup in memory usage
- regression test framework is now ready (using atf) regression test are slowly
  being added and populated.

0 files changed, 0 insertions, 0 deletions