Anders J. Munch
andersjm at inbound.dk
Mon Feb 3 21:54:31 CET 2003
"Alex Martelli" <aleax at aleax.it> wrote:
> Anders J. Munch wrote:
> > "Chris Gonnerman" <chris.gonnerman at newcenturycomputers.net> wrote:
> >> You don't have a Microsoft C compiler. The distutils
> >> (invoked by the setup.py you mention) by default
> >> assume you are using MS Visual C.
> > Strange assumption.
> > 'in-the-face-of-ambiguity'-well-you-know-the-rest-ly y'rs, Anders
> If this is a criticism of the distutils' design choice, I
> want to express my disagreement with it. The distutils
> default to the same compiler that is used to build the
> standard Python distribution on a Windows platform: this
> seems to me to be exactly the right default.
That makes it the right choice when choosing between available
compilers. For a source distribution it's also reasonable to assume
that the compiler used to build the interpreter is available which
makes it a good default. For a binary distribution assuming that the
build compiler is available is - guessing.
The MSVC version is designed to be a binary distribution. That shows
through MSVC compiler options being hardwired into distutils, unlike
the unix variation, where as far as I can see options are remembered
from the build.
> Calling it an "assumption" may be misleading. But call it
> a "default" and it makes more sense, no?-)
I call it a default borne out of an assumption. Same difference :-)
Not a big deal, distutils is great and a good guesser;-). It just
struck me that the OPs problems were caused by distutils failing to
observe one of Tim's theses. Perhaps you feel that in this specific
case a little guesswork is warranted, if so fine with me, but it would
have been a perfectly fine alternative for distutils to print an
informative message instead.
More information about the Python-list