Metakit 2.4.9.6

[jcw]

A new release of Metakit has been checked into CVS.
Change history since 2.4.9.5 attached.

Enjoy,
-jcw


2006-11-15    ############################################    MK 2.4.9.6

     Please upgrade, this consolidates a year with various fixes.

     Added some notes to the README about building the Python and
     Tcl extensions using various approaches.

     Bumped hardwired Python version number to 2.5 in unix/configure.

2006-11-11    More precise file head/tail marker checks

     The tests for valid head / tail file markers have been relaxed a
     tiny bit, to allow for future file format changes.  This has no
     effect on current uses of MK, the file format remains the same.

2006-11-11    Reformatted all C++ source files

     To get rid of some tab/indent problems accumulated over the last
     few years, I've run all C++ sources through a pretty-print app
     (SourceFormatX, a shareware Windows app).  These sources are all
     being checked-in, so the list of affected files is quite large.

     No semantic changes whatsoever to MK.  All tests pass, as before.

2006-10-31    Fixes for the Python binding

     Fix 64-bit compile with Python 2.5 by Ralf Schmitt, see
     http://www.equi4.com/pipermail/metakit/2006-October/002236.html

     Memory leak fix in PWOMapping.h, by Brian Warner, see
     http://www.equi4.com/pipermail/metakit/2006-October/002229.html

     Fix 64-bit ptr -> int conversions in {PyRowRef,PyView}.cpp, see
     http://systemexit.de/repo/metakit?cs=f8f1ef2f6099

2006-10-31    Fix int overflow in mapped file check

     As reported by Max Kaehn, there is a test in column.cpp which is
     sensitive to integer overflow when two pointers are > 2 Gb apart.
     This may have been an old problem for anyone who uses the default
     memory-mapping files enabled.  Embarassing that this was not caught
     or resolved before, but I was never able to reliably reproduce it.

     This is a potentially serious problem, upgrading is recommended.

2006-10-31    Bring this change log up to date from CVStrac timeline

     I've decided to not adopt the CVStrac system for this change log,
     because having this log inside the source tree is more convenient.

     (08-01) persist.cpp: maintain stable-storage space usage on re-open
     (05-08) view.cpp: fix resource leak in view.cpp
     (01-26) PyStorage.cpp: fix metakit.py pathnames are required to
             be of type str & support unicode filenames in storage()
     (01-25) mk4tcl.cpp: fix crash in mk::select with -keyword search
     (01-24) fileio.cpp: fix compile error in fileio.cpp with mingw
     (01-13) configure.in: tweaks to configure.in to use Tcl stubs when
             compiling a dynamic Mk4tcl on macosx/darwin

2006-01-12    Don't call _open on WinNT, make commit fail on > 2 Gb

     Fixes for the Dec 5 patch to still always support Win98, thx Matt.

     Fix ticket #2, return false on commit if files would exceed 2 Gb.

2006-01-12    Started using CVStrac for this project, makefile tweak

     CVStrac is a CVS browser, bug tracking system, and more.  All files
     have gone through a dummy check-in to complete the transition.  The
     CVS repository is not affected by this change, but this does mean
     that CVStrac only really started tracking changes from here on.

     If it works out well, this CHANGES file will probably be abandoned
     in favor of the "timeline" feature in CVStrac which offers the same
     functionality and *much* more (lots of cross-/hyper-links for one).

     Tweaked Makefile so "make test" does not complain about CVS dir.

2005-12-05    Fix the broken hpux change of 2005-09-06

     It turns out that the 2005-09-06 change, which was rolled back on
     2005-10-01, was simply incomplete.  Now that it works, it has been
     adopted on all platforms because it is more portable and  
equivalent.
     Thx again Matt.

2005-12-05    Better support for UTF-8 file & property names

     Patch submitted by Matt Newman.  Fixes handling of property names
     in a description string to support UTF-8 and %'s in names.  Also
     fix file name handling (convert UTF-8 to native before open).

2005-11-22    ############################################    MK 2.4.9.5

_____________________________________________
Metakit mailing list  -  Metakit@...
http://www.equi4.com/mailman/listinfo/metakit

[John Fletcher-2]

Where can I get it?  The web site

http://www.equi4.com/metakit.html

has 2.4.9.5


John

To:             Metakit list <metakit@...>
From:           Jean-Claude Wippler <jcw@...>
Date sent:       Wed, 15 Nov 2006 14:22:26 +0100
Subject:         [Metakit] Metakit 2.4.9.6

> A new release of Metakit has been checked into CVS.
> Change history since 2.4.9.5 attached.
>
> Enjoy,
> -jcw
>
>
> 2006-11-15    ############################################    MK
> 2.4.9.6
>
>      Please upgrade, this consolidates a year with various fixes.
>
>      Added some notes to the README about building the Python and Tcl
>      extensions using various approaches.
>
>      Bumped hardwired Python version number to 2.5 in unix/configure.
>
> 2006-11-11    More precise file head/tail marker checks
>
>      The tests for valid head / tail file markers have been relaxed a
>      tiny bit, to allow for future file format changes.  This has no
>      effect on current uses of MK, the file format remains the same.
>
> 2006-11-11    Reformatted all C++ source files
>
>      To get rid of some tab/indent problems accumulated over the last
>      few years, I've run all C++ sources through a pretty-print app
>      (SourceFormatX, a shareware Windows app).  These sources are all
>      being checked-in, so the list of affected files is quite large.
>
>      No semantic changes whatsoever to MK.  All tests pass, as before.
>
> 2006-10-31    Fixes for the Python binding
>
>      Fix 64-bit compile with Python 2.5 by Ralf Schmitt, see
>      http://www.equi4.com/pipermail/metakit/2006-October/002236.html
>
>      Memory leak fix in PWOMapping.h, by Brian Warner, see
>      http://www.equi4.com/pipermail/metakit/2006-October/002229.html
>
>      Fix 64-bit ptr -> int conversions in {PyRowRef,PyView}.cpp, see
>      http://systemexit.de/repo/metakit?cs=f8f1ef2f6099
>
> 2006-10-31    Fix int overflow in mapped file check
>
>      As reported by Max Kaehn, there is a test in column.cpp which is
>      sensitive to integer overflow when two pointers are > 2 Gb apart.
>      This may have been an old problem for anyone who uses the default
>      memory-mapping files enabled.  Embarassing that this was not
>      caught or resolved before, but I was never able to reliably
>      reproduce it.
>
>      This is a potentially serious problem, upgrading is recommended.
>
> 2006-10-31    Bring this change log up to date from CVStrac timeline
>
>      I've decided to not adopt the CVStrac system for this change log,
>      because having this log inside the source tree is more
>      convenient.
>
>      (08-01) persist.cpp: maintain stable-storage space usage on
>      re-open (05-08) view.cpp: fix resource leak in view.cpp (01-26)
>      PyStorage.cpp: fix metakit.py pathnames are required to
>              be of type str & support unicode filenames in storage()
>      (01-25) mk4tcl.cpp: fix crash in mk::select with -keyword search
>      (01-24) fileio.cpp: fix compile error in fileio.cpp with mingw
>      (01-13) configure.in: tweaks to configure.in to use Tcl stubs
>      when
>              compiling a dynamic Mk4tcl on macosx/darwin
>
> 2006-01-12    Don't call _open on WinNT, make commit fail on > 2 Gb
>
>      Fixes for the Dec 5 patch to still always support Win98, thx
>      Matt.
>
>      Fix ticket #2, return false on commit if files would exceed 2 Gb.
>
> 2006-01-12    Started using CVStrac for this project, makefile tweak
>
>      CVStrac is a CVS browser, bug tracking system, and more.  All
>      files have gone through a dummy check-in to complete the
>      transition.  The CVS repository is not affected by this change,
>      but this does mean that CVStrac only really started tracking
>      changes from here on.
>
>      If it works out well, this CHANGES file will probably be
>      abandoned in favor of the "timeline" feature in CVStrac which
>      offers the same functionality and *much* more (lots of
>      cross-/hyper-links for one).
>
>      Tweaked Makefile so "make test" does not complain about CVS dir.
>
> 2005-12-05    Fix the broken hpux change of 2005-09-06
>
>      It turns out that the 2005-09-06 change, which was rolled back on
>      2005-10-01, was simply incomplete.  Now that it works, it has
>      been adopted on all platforms because it is more portable and  
> equivalent.
>      Thx again Matt.
>
> 2005-12-05    Better support for UTF-8 file & property names
>
>      Patch submitted by Matt Newman.  Fixes handling of property names
>      in a description string to support UTF-8 and %'s in names.  Also
>      fix file name handling (convert UTF-8 to native before open).
>
> 2005-11-22    ############################################    MK
> 2.4.9.5
>
> _____________________________________________
> Metakit mailing list  -  Metakit@...
> http://www.equi4.com/mailman/listinfo/metakit
>

--
Dr John P. Fletcher   Tel: [phone] (direct line)
Chemical Engineering and Applied Chemistry (CEAC),
School of Engineering and Applied Science (SEAS),
Aston University, Aston Triangle, BIRMINGHAM B4 7ET  U.K.              
CEAC Web site http://www.ceac.aston.ac.uk/
FAX: [phone]


_____________________________________________
Metakit mailing list  -  Metakit@...
http://www.equi4.com/mailman/listinfo/metakit

[jcw]

John Fletcher wrote:

> Where can I get it?  The web site
>
> http://www.equi4.com/metakit.html
>
> has 2.4.9.5

Updated now.  The sources I announced were in CVS, for info see:
        http://www.equi4.com/mkdownload.html

The source tarfile has been updated, no new binaries at this stage.

(for Python it'd have to be for 2.4 or 2.5, but I don't know which)

Thanks,
-jcw


_____________________________________________
Metakit mailing list  -  Metakit@...
http://www.equi4.com/mailman/listinfo/metakit