Python MSVC++ binaries considered evil

Cliff Wells logiplexsoftware at earthlink.net
Fri Jan 25 20:16:20 CET 2002


On Thu, 24 Jan 2002 18:46:25 -0500
Tim Peters wrote:

> [Cliff Wells]
> > And I imagine at least part of the reason for Windows volunteers
> > being rare is not many of them have a C/C++ compiler that they can
> > use to compile Python.
> 
> Python is written in standard C, so the notion that only MSVC is capable
of
> compiling it is absurd.  If you want the *Windows*-specific features to
> work, though, you have to play the Windows API game, and MSVC excels at
that
> part.

What I was trying to say is that for a Windows user to actually do the
compilation with a free compiler, it is nowhere near as simple as it is for
say Linux, where gcc is the standard.  Obviously Python itself is very
portable.  I also won't disagree that VC++ may in fact be the best Windows
compiler (probably for reasons best left to the antitrust courts).

> > Perhaps if the Windows port were decoupled a bit from VC++ (or at
> > least some Makefiles provided for FreeBCC) ...
> 
> As I said, PythonLabs hasn't the bandwidth to pursue Windows
alternatives.
> "Support for alternatives will have to originate from, be resolved
within,
> and implemented by the community", exactly the same story as for
Macintosh
> ports, Plan9 ports, BeOS ports, RiscOS ports, HP-UX ports, Cray T3E
ports,
> EPOC32 ports, etc etc etc.

Since I was replying to your post, maybe you misinterpreted my intention,
but I wasn't suggesting that PythonLabs should be responsible for it.  This
list _is_ "the community" so I was discussing it here.  
 
> > that situation would change for the better. Or maybe not.  It may be
> > more of a mind-set difference between Unix-types (who are used to
> > building from source and fixing/reporting the occasional platform-
> > dependency issue) and Windows-types who expect InstallShield to do
> > it all for them.
> 
> I expect that's more on target.  Also that most high-powered Windows
> developers (meaning developers who target Windows (like Mark Hammond),
not
> developers who just happen to use Windows (like me)) use MSVC anyway.

This is of course the classic problem with the Windows platform: everyone
uses (or believes everyone else does) Microsoft tools so everyone who
doesn't want to drop $$$ for them is left out (or at least left looking for
pre-built binaries).  I guess we can hold our breath for the day when MS
includes a basic C/C++ compiler and libraries with the OS.  What a
concept... this is one of the main reasons I've stuck with Linux despite
it's shortcomings and growing pains.  IMHO, an OS without a default
compiler is like an OS without built-in networking: next to worthless.

> > Hrm.  I may have to download FreeBCC and see for myself how much
> > trouble it would really be...
> 
> Read Python's NEWS file -- you wouldn't be the first.  The last person
who
> took a crack at it seemed to make swift progress, but didn't finish the
job
> (by his own account).  The greater pain is likely sorting out
> incompatibilities between the MS and Borland libraries (or declaring that
> yuu don't care about that, but then you've got a port with a small
> audience).

Yeah, and it would only add to the confusion to have two separate Windows
ports :P  Still, I think I'll look at it this weekend.


-- 
Cliff Wells
Software Engineer
Logiplex Corporation (www.logiplex.net)
(503) 978-6726 x308
(800) 735-0555 x308




More information about the Python-list mailing list