IXIN 1.7 available

[Thien-Thi Nguyen]

release notes:

  OK, w/ this release, we've fully explored the "incremental access"
  side of IXIN purpose; now to turn to the "runtime-tweakable" side.
  It looks more and more like TWEAKS in each NODE-INDEX entry is not
  going to be sufficient.  For example, what if, within a node, some
  setting is changed many times?  The current thinking (hinted to in
  README) is to move settings out of NODE-DATA completely, such that
  rendering programs need not support two forms of settings, the one
  used in META, and the one expressed (balkily) in SXML.

  This release is the last to support Texinfo 4.13.  If you haven't
  already, please try using 4.13.93 (or later) for playing w/ IXIN.

  thi

README excerpt:

  IXIN is an EXPERIMENTAL distribution of EXPERIMENTAL code and
  some pre-built IXIN files.  IXIN stands for "indexed texinfo".

  All code is under GPLv3+.  All (.xml) docs are under GFDL.

  ‘spec/ixin.texi’ is The IXIN Chronicles, a document that
  defines the file format and provides related info.  There
  are various output formats, including .info, .pdf and .ixin.

  ‘c/’ contains some simple command-line utilities to write and
  read the file format, and a proof-of-concept rendering program.

  ‘d/’ contains the some pre-built .sxml and .ixin files.
  ‘d/GNUmakefile’ supports "make demo" and "make demo-zow",
  among other commands.

NEWS excerpt:

  - 1.7 | 2013-01-11

    - bugfix: a1-nf3-mixp now respects attribute ‘xml:space’

        That is, when this attribute has value "preserve", a1-nf3-mixp
        no longer touches the whitespace in an element's immediate
        children.

    - bugfix: NODE-INDEX entries' LENGTH includes trailing newline

        This brings the utils in line w/ the specification, which DOES
        count the trailing newline.  Previously, mkixin ignored it and
        retrieve added it back in.

    - bugfix: retrieve repl takes EOF as ‘(quit)’

        Regression introduced w/ IXIN 1.6 (2013-01-03); used to work
        fine w/ IXIN 1.5 (2012-12-30).  Blech, progress...

    - META ... DESCRIPTION can be a renderable sequence

        This used to be specified as a single-line string, which is too
        restrictive.  See d/spec.ixin, for example.

    - images no longer inline; new blocks BLOBS-INDEX, BLOBS

        Placing the images inline is counter to the goal of supporting
        incremental access.  Oops, sorry about that.

        Another advantage (arguably) w/ blobs is that it is easier to
        expand the set of allowed (MIME) types in the future.  At the
        moment, only image/* is specified, but it's not difficult to
        imagine audio/* also.  On the other hand, probably video/* is
        out of the question.  Feedback in this area most welcome.

    - new "retrieve repl" commands: backward, forward, blob

        The first two change the "current node" numerically, ignoring
        the ‘next’, and ‘prev’ pointers.

        The latter displays BLOB-INDEX entry and BLOB-DATA for the
        specified ID.

    - "retrieve dump" displays blobs, too

        The ‘dump’ and ‘dump/abbrev’ commands now also display blob
        index and blob data, analogously to the node data.

    - changes to spit.el

      - new commands: ‘[’, ‘]’

        These are ‘spit-%backward’ and ‘spit-%forward’, respectively,
        which in turn run the ‘backward’ and ‘forward’ repl commands.

      - prettier node names after ‘M-s’

        Node names are cached by both ‘M-i’ and ‘M-s’, the latter more
        fully and w/ precedence.  To see this in action, do:

         M-x spit RET d/coreutils.ixin RET

        then:

         M-d    ; node identifiers (not pretty)
         M-i    ; names w/ ‘variable-pitch’ face cached
         M-d    ; pretty
         M-s    ; names w/ fully-propertized face cached
         M-d    ; even more pretty
         M-i    ; no cache update
         M-d    ; still OK

        On a slow computer (like the author's), try d/rcs.ixin instead.

        Note that this caching conflates node names with sectioning tree
        names, which loses if these do not correspond "well enough", or
        if there are multiple sectioning elements "in" a node.  The
        latter causes Texinfo 4.13.93 to issue a warning; a rigorous
        renderer should do likewise, or compensate accordingly.

      - ‘M-m’ output prettier

        The INVITATIONS are now displayed w/ pretty face, etc.  This is
        quite dramatic for d/coreutils.ixin, for example.

      - ‘s’ decorates ‘(ixcc ...)’

        The ‘(ixcc ...)’ forms are displayed as buttons (that have no
        action) with the coordinate replaced with a cached rendering
        (from ‘M-d’, ‘M-i’ or ‘M-s’), if available.  For example:

         (ixcc "(node 0)")

        will be replaced with

         (ixcc "Top")

        where the "ixcc" part is a button.  There is at this time
        support for families ‘node’, ‘blob’ (see below) and ‘(dts NAME)’
        only.  When there is no cached rendering available, i.e., for
        family ‘(floatset TYPE)’, the coordinate is anyway decorated
        with the family in red, the position yellow.

        Relatedly, typing ‘tab’ or ‘backtab’ now runs command
        ‘forward-button’ or ‘backward-button’, respectively.

      - ‘C-M-d’ (aka ‘spit-%list-dts’) deleted

        The output for this command is largely redundant w/ that of
        ‘M-d’ (aka ‘spit-%dump-dts’).

      - image blobs displayed inline

        E.g., spec/ixin.ixin node 1 (Introduction) shows the logo.

    - spec document renamed: The IXIN Chronicles

        Its scope is now expanded beyond the spec proper to include
        other stuff (i.e., opportunities to exercise the utils), such as
        invocation (usage) and implementation notes for the utilities,
        interesting character sequences from d/prob.xml, how to compute
        the node base and blob base, etc.

        Relatedly, there are now many more terms in the indices.

    - mkixin uses temporary dir for temporary files

        This means you can use "make -j" in d/, now.  Previously, the
        temporary filenames were hardcoded, confusing peer processes,
        and resulting in corrupted output.

tarball, etc, in dir:

  http://www.gnuvola.org/software/ixin/

atom feed:

  http://www.gnuvola.org/NEWS.atom

-- 
Thien-Thi Nguyen ..................................... GPG key: 4C807502
.                  NB: ttn at glug dot org is not me                   .
.                 (and has not been since 2007 or so)                  .
.                        ACCEPT NO SUBSTITUTES                         .
........... please send technical questions to mailing lists ...........

pgpQHnezK0hv4.pgp
Description: PGP signature