Hello!


This time my status update will take you closer to the land of details.
 A few things are not new, but many are; read on please.

-   Decisions that rbu and I have agreed on (in doubt mistakes are
    mine):

    -   USE_EXPAND use flags (e.g. ALSA_CARDS) are now handled like
        normal use flags (e.g. ssl).  They were stripped away until
        recently.

    -   Overlay names from layman-global.txt are assumed to be globally
        unique and used for overlay identification.  One of the
        consequences of that is that you we can collect data from global
        overlays (e.g. sunrise) that you do not manage with layman.

    -   rsync excludes and filters from PORTAGE_RSYNC_EXTRA_OPTS will
        not be handled.  It would be an inadequately complex extra task
        while not really gaining much at the same time.  It would also
        require to write code at a level where we can currently use
        unmodified portage API.

    -   package.use data will not be collected:  We collect the set of
        actually used flags for each package installed and the set of
        global use flags already.  The source of a use flag beyond that
        (i.e. distinct USE= in env versus and entry package.use) does
        not seem useful to us.

    -   package.{keywords,unmask} and world file entries will not be
        collected separately but grouped with the list of installed
        packages as this data makes most sense together.  For each
        package we are now collecting:

        -   Category and package name
        -   Version and revision
        -   Slot
        -   Keyword status
                i.e. what kind of package.keywords entry would be needed
                to re-install the package now.  Can be one of:
                -   '' means the package is no less stable than the
                    system itself
                -   '~arch' means an unstable package on a stable system
                -   '**' catches case KEYWORDS='' mainly
        -   Masked flag
        -   Unmasked flag
        -   In-world-file flag
        -   Source overlay/tree name
                (e.g. 'gentoo' or 'sunrise')
        -   Relevant enabled flags
                (subset of ISUE from the related ebuild)

        package.mask cannot be combined with installed package stats in
        all cases and will be reported separately as planned before.

-   Collection and auto-filtering of Gentoo-specific machine data is now
    complete:

    -   Compile flags
        -   CFLAGS, CXXFLAGS, LDFLAGS (right, that's link flags ...)
        -   Basic GCC-oriented parsing

    -   Global use flags
        -   Parsing of profiles/...
            -   arch.list
            -   use.desc
            -   use.local.desc
            -   desc/*.desc
        -   Filtering away other use flags

    -   Installed packages
        -   Filtering away packages that are not from the main tree or
            a layman-global overlay
        -   More details given above

    -   Make options
        -   Basic GNU make-oriented parsing

    -   Mirrors
        -   Distfiles mirrors (filtered against synced copy of
            mirrors3.xml)
        -   rsync mirror (filtered against gentoo.org host)

    -   Overlays
        -   Filtered against synced copy of layman-global.txt
        -   Works without Layman installed

    -   User package.mask
        -   Filtering away packages that are not from the main tree or
            a layman-global overlay

    -   System profile

    -   Trivial portage scalars:
        -   ARCH
        -   CHOST
        -   USE_ORDER

    -   Trivial portage vectors:
        -   ACCEPT_KEYWORDS
        -   FEATURES

-   On my list of further data to collect at the moment:
    -   Alpha CPU details
        (see mail "Project summaries - Alpha Arch Team")

-   All code over here [1], review welcome:


Sebastian


[1]
http://git.goodpoint.de/?p=smolt-gentoo.git;a=tree;f=client/distros/gentoo;hb=refs/heads/gentoo