summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--PC/readme.txt16
1 files changed, 3 insertions, 13 deletions
diff --git a/PC/readme.txt b/PC/readme.txt
index c626208..62a2ae9 100644
--- a/PC/readme.txt
+++ b/PC/readme.txt
@@ -51,6 +51,9 @@ src A subdirectory used only for VC++ version 1.5 Python
Microsoft Visual C++ Version 4.0 (32-bit Windows)
=================================================
+(While we have done all our testing with VC version 4.0, we are
+confident that the same procedure will work for versions 4.1 and 4.2.)
+
The distributed Makefile is vc40.mak. This file is distributed with
CRLF line separators, otherwise Developer Studio won't like it. It
will NOT work from this PC directory. To use it, first copy it to the
@@ -72,9 +75,6 @@ python The Python main program, named python.exe. This should
_tkinter The optional _tkinter extension, _tkinter.dll; see below.
-NumPy, multiarray, fast_umath
- Optional projects to build Numerical Python.
-
ALl end products of the compilation are placed in the subdirectory
vc40 (which Developer Studio creates); object files are placed in
vc40/tmp. There are no separate Release and Debug project variants.
@@ -93,16 +93,6 @@ those files in the PC subdirectory. In order to use _tkinter, the
Tkinter.py module must be on PYTHONPATH. It is found in the
Lib\tkinter subdirectory.
-*** How to build Numerical Python ***
-
-This has been tested only with NumPy 1.0alpha2. Unpack the NumPy
-tarfile so that it creates a NumPy subdirectory in the Python
-distribution directory. You need to create exports files
-NumPy/multiarray.def and NumPy/fast_umath.def (see PC/_tkinter.def for
-an example). Then build the targets NumPy, multiarray and fast_umath,
-in that order. To use Numerical Python you have to append NumPy/Lib
-to sys.path.
-
Additional files and subdirectories for 32-bit Windows
======================================================