Commit graph

21 commits

Author SHA1 Message Date
recht
e914ab3288 hange extension.mk to also install optimized files for distutils packages.
Inspired by FreeBSD "ports".

Fix the PLISTs accordingly.
Also, while at it, remove now obsolete compileall.py calls in post-install
targets and insure that extension.mk is in included before builinlinks of
other Python modules.

Discussed with/ok'ed by drochner@.
2003-09-14 16:43:45 +00:00
grant
0155927c43 s/netbsd.org/NetBSD.org/ 2003-07-17 21:41:05 +00:00
jschauma
e366d0c694 Use tech-pkg@ in favor of packages@ as MAINTAINER for orphaned packages.
Should anybody feel like they could be the maintainer for any of thewe packages,
please adjust.
2003-06-02 01:15:31 +00:00
wiz
f0dfffd966 buildlink1 -> buildlink2. 2002-10-09 17:38:02 +00:00
jlam
e44bf515dc Strip the ".buildlink" from the names of the python application and
extension Makefile fragments, because they really don't have anything to
do with the buildlink[12] frameworks.  Change all the Makefiles that use
application.buildlink.mk and extension.buildlink.mk to use application.mk
and extension.mk instead.
2002-09-21 23:46:45 +00:00
drochner
9871f96c24 move to canonical pkg name (pyX.Y-pkg-vers),
append version suffix to extended python interpreter,
don't install docs (separate pkg now)
2002-01-17 12:02:25 +00:00
drochner
5ea947fc92 we need at least Python 2.0.1nb1 or 2.1.1nb1, 2.2 doesn't work at all 2002-01-16 20:32:42 +00:00
drochner
c511169aa8 use buildlink files for Python and py-Numeric 2002-01-16 19:45:26 +00:00
zuntum
98cf373d36 Move pkg/ files into package's toplevel directory 2001-11-01 00:32:23 +00:00
drochner
c8124ca977 use USE_PYTHON 2001-05-14 16:22:35 +00:00
drochner
b9ee1933ce set runtime library search path so that libnetcdf is found 2001-05-10 17:03:28 +00:00
jtb
0628e0b0ff Only install the include files once. 2001-04-22 20:15:46 +00:00
agc
fa5c767d84 Move to sha1 digests, and add distfile sizes. 2001-04-19 08:45:37 +00:00
agc
9e8d6c8b8d + move the distfile digest/checksum value from files/md5 to distinfo
+ move the patch digest/checksum values from files/patch-sum to distinfo
2001-04-17 11:33:31 +00:00
jtb
eb0cef2016 Change MAINTAINER section to packages@netbsd.org 2001-04-14 21:43:38 +00:00
jtb
8fce78ee75 Added defaults for EVAL_PREFIX. 2001-04-08 22:53:14 +00:00
jtb
a9668a7762 Updated to py-Scientific-2.2.
2.1.5 --> 2.2
-------------

Bug fixes:

- The sign convention for rotation angles was not correctly applied
  in Scientific.Geometry.Transformation and Scientific.Geometry.Quaternions.
  Everything was consistent, but with the wrong sign.

Improvements:

- Scientific.IO.PDB can handle a larger range of variant formats for DNA

New features:

- New module Scientific.Visualization.VPython.

- Scientific.IO.TextFile also handles bzip2 compressed files
  (extension .bz2). This only works if bzip2 is installed.


2.1.4 --> 2.1.5
---------------

New features:

- Scientific.TkWidgets.TkPlotCanvas:
  - Popup menu on right button
  - Value display on middle button
  - HorizontalLine, VerticalLine objects

Modifications:

- MPI interface no longer assumes that MPI_Op and MPI_Datatype are
  integers (they aren't in LAM, for example).


2.1.3 --> 2.1.4
---------------

New features:

- New MPI functions implemented by Jakob Schiotz:
  - nonblocking send and receive
  - abort
  - reduce and allreduce

Modifications:

- The Python interpreter lock is released during calls to the netCDF
  library, permitting other threads to continue during I/O operations.
  An internal lock has been added to prevent reentrant calls to the
  netCDF library, which is not thread safe.
  (If you don't understand any of this, you shouldn't care, it has
  no consequences unless you use threads.)
- Methods receive and receiveString on MPI communicator objects
  can be called without source and tag arguments (defaults to None).


Bug fixes:

- The compilation script did not work correctly for MPI support


2.1.2 --> 2.1.3
---------------

New features:

- Some small utility widgets have been added to Scientific.TkWidgets.
  They are not particularly "scientific", but not worth the effort
  of a separate distribution.

Bug fixes:

- The abs() function didn't work on DerivVar objects (modules
  Scientific.Functions.Derivatives and Scientific.Functions.FirstDerivatives)
  with negative values.

- The method projectionOf in Scientific.Geometry.Objects3D.Plane was
  wrong.


2.1.1 --> 2.1.2
---------------

Installation:

- The mpipython executable is compiled using 'mpicc', which should simplify
  installation significantly because all library specifications are
  automatically taken into account. In other words, MPI support should
  work on all platforms without any modifications to the Setup file.
  (Thanks to Jakob Schiotz for this suggestion!)

- Some 'extern' declarations were changed to 'staticforward' to please
  some compilers.

Documentation:

- the documentation for the methods receive and receiveString were
  wrong.


2.1.0 --> 2.1.1
---------------

Bug fixes:

- The upper bound for MPI message tags was not treated correctly,
  leading to a bound of zero on some platforms.

Modifications:

- The methods sendArray and sendString were combined into a single
  method send.

- broadcastArray was renamed to broadcast, shareArray was renamed to share,
  receiveArray was renamed to receive. The only function that retains
  a data type indication is receiveString.


2.0.1 --> 2.1.0
---------------

Additions:

- MPI interface in Scientific.MPI
2001-03-30 22:23:07 +00:00
jtb
03f862678a Fix typo which caused i386 bulk build problem.
mpich->=1.2.1 becomes mpich>=1.2.1 .
2001-03-29 01:00:10 +00:00
jtb
fbf3f55fab Changed the build process to work with the recently updated mpi-ch package
which installs its parts in locations different to the previous version.
2001-03-25 21:43:02 +00:00
wiz
a13ea108bb Update to new COMMENT style: COMMENT var in Makefile instead of pkg/COMMENT. 2001-02-17 17:52:59 +00:00
jtb
489b39e572 Initial import of new py-Scientific package:
Python modules for scientific computing
2001-01-30 02:01:43 +00:00