Pkgcore

----------------------------

- new portage configuration feature- 'ignore-paludis-versioning'. This
directs pkgcore to ignore nonstandard -scm ebuilds instead of complaining
about them.
Note this does *not* affect the installed pkgs database- if there is a
-scm ebuild in the vdb, pkgcore *must* deal with that ebuild, else if it
silently ignores vdb -scm pkgs it can result in overwriting parts of the
-scm pkg, and other weirdness. If you've got a -scm version pkg installed,
it's strongly suggested you uninstall it unless you wish to be bound to that
nonstandard behaviour of paludis.

Finally, it's not yet covering *all* paludis version extensions- that will
be expanded in coming versions.

- pkgcore is now aware of installed -scm pkgs, and gives a cleaner error
message.

- a few versions of portage-2.2 automatically added PKGSET items to the
world file; due to how portage has implemented their sets, this would
effectively convert the data to portage only. As such, that feature was
reversed (thank you genone); that said, a few world files have pkgset
entries from these versions. Pkgcore now ignores it for worldfiles, and
levels a warning that it will clear the pkgset entry.

- ticket 174; ignore bash style comments (leading ) in pkgsets, although
they're wiped on update. If folks want them preserved, come up with a way
that preserves the location in relation to what the comment is about- else
wiping seems the best approach.

- ticket 14; tweak PORT_LOGDIR support a bit, so that build, install,
and uninstall are seperated into different logs.

- added '' operator to pmerge as an alias for --set; for example,
'pmerge system' is the same as 'pmerge --set system'.

- fallback method of using the file binary instead of libmagic module is
fixed; ticket 183.


----------------------------

----------------------------

- correct a flaw in repository searching that slipped past the test harness.
effectively breaks via inverting the negate logic for any complex search.


--------------------------

--------------------------

- prepstrip was updated to match current portage semantics, minus stripping
and splitdebug functionality (we handle that via a trigger). Via this,
FEATURES=installsources and basic bincheck (pre-stripped binaries) is now
supported.

- FEATURES='strip nostrip splitdebug' are now supported in portage
configuration (trigger is pkgcore.merge.triggers.BinaryDebug).

- added cygwin ostype target for development purposes. In no shape or form
is this currently considered supported, although anyone interested in
developing support for that platform, feel free to contact us.

- in candidate identification in repository restriction matching, it was
possible for a PackageRestriction that was negated to be ignored, thus
resulting in no matches. This has been corrected, although due to
collect_package_restrictions, it's possible to lose the negation state
leading to a similar scenario (no known cases of it currently). This
codepath will need reworking to eliminate these scenarios.

- mercurial+ sync prefix is now supported for hg.

- triggers _priority class var is now priority; overload with a property if
custom functionality is needed.


--------------------------

--------------------------

- filelist sets (world file for example) are now sorted by atom comparison
rules. ticket 178.

- pquery --restrict-revdep-pkgs and --revdep-pkgs were added: they're
used to first match against possible pkgs, then do the revdep looking for
pkgs that revdep upon those specific versions. Functionality may change,
as may the outputting of it. ticket 179.

- pebuild breakage introduced in 11/07 is corrected; back to working.

- 'info' messages during merging are now displayed by default- new debug
message type was added that isn't displayed by default.

- ebuild domain now accepts triggers configuration directive.

- FEATURES=unmerge-buildpkg was added; this effectively quickpkgs a pkg
before it's unmerged so you have a snapshot of its last state before
it is replaced.

- FEATURES=pristine-buildpkg was added; this is like FEATURES=buildpkg,
but tbzs the pkg prior to any modification by triggers. Upshot of this,
you basically have an unmodified binpkg that can be localized to the merging
host rather then to the builder. Simple example, with this if your main
system is FEATURES=strip, it tucks away a nonstripped binpkg- so that
consumers of the binary repo are able to have debug symbols if they want
them.

- FEATURES=buildsyspkg is now supported.

- FEATURES=buildpkg is now supported.

- the engine used for install/uninstall/replace is now configurable via
engine_kls attribute on the op class.

- dropped exporting of USER='portage' if id is portage. Ancient var setting,
can't find anything reliant on it thus punting it.

- add SunOS to known OS's since its lchown suffices for our needs.

- added eapi awareness to atoms, so that an eapi1 atom only allows the
slot extension for example.

- remove a stray printf from cpy atom; visible only when repository atoms
are in use.


--------------------------

--------------------------

- fix collision unprotect trigger exceptions (typically KeyError).
ticket 165

- correct invalid passing of force keyword down when the repository isn't
frozen. Occasionally triggered user visible tracebacks in pmaint copy.

- portage broke compatibility with pkgcore a while back for our binpkgs-
for some inane reason, portage requires CATEGORY and PF in the xpak
segment. This is being removed from portage in 2.2, but in the interim
pkgcore now forces those keys into the binpkgs xpak for compatibility
with portage.

Shorter version: pmaint copy generated binpkgs work with portage again.

- cbuild/chost/ctarget are available via pquery --attr, and are written to
binpkg/vdb now.

- stat removal work: FEATURES=-metadata-cache reuses existing eclass cache
object, thus one (and only one) scan of ${PORTDIR}/eclass

- metadata, flat_hash, and paludis_flat_list cache formats configuration
arg 'label' is no longer required, and will be removed in 0.5. If they're
unspecified, pkgcore will use location as the place to write the cache at,
else it'll combine location and label.

- cdb, anydbm, sqlite, and sql_template cache backends have been removed
pending updating the code for cache backend cleanups. If interested in
these backends, contact ferringb at irc://freenode.net/pkgcore .


--------------------------

--------------------------

- merging/replacing performance may be a bit slower in this release- the level
of stats calls went up in comparison to previous releases, with several
duplicates. This will be corrected in the next release- releasing in the
interim for bugfixes this version contains.

- add CBUILD=${CBUILD:-${CHOST}}; couple of odd ebuilds rely on it despite
being outside of PMS.

- protective trigger was added blocking unmerging of a basic set of
directories/syms; mainly /*, and /usr/*.

- when a merge passes through a symlink for path resolution, that sym is
no longer pulled in as an entry of that pkg. Originally this was done for
protective reasons, but it serves long term as a way to inadvertantly hold
onto undesired junk from the users fs, and opens the potential to unmerge
system/global symlinks when that pkg/slot's refcount hits zero.

- detection, and predicting merge locations for syms was doing an unecessary
level of stat calls; this has been reduced to bare minimum.

- ticket 159; force an realpath of CONTENTS coming from the vdb due to other
managers not always writing realpath'd entries, thus resulting in occasional
misidentification of what to remove.

- pkgcore.util.parserestrict no longer throws MalformedAtom, always
ParseError. Removes ugly commandline tracebacks for bad atoms supplied
to pmerge.

- ticket 158; honor RSYNC_PROXY for rsync syncer.
Thanks to user Ford_Prefect.

- pmerge -N now implies --oneshot.

- correct a flaw in tbz2 merging where it repeatedly try to seek in the bz2
stream to generate chksums, instead of using the on disk files for
chksumming.

- pmaint regen w/ > 1 thread no longer throws an ugly set of tracebacks upon
completion.

- binpkg repositories now tell you the offending mode, and what is needed
to correct it. No longer cares if the specified binpkg base location is
a symlink also.

- pmaint --help usage descriptions are far more useful now.


--------------------------