Re: [boost] Problems compiling boost_python under Borland C++ Builder 5

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2001-06-13 21:27:11

----- Original Message -----
From: "Vesselin Iliev Peev" <VesselinPeev_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Wednesday, June 13, 2001 9:58 PM
Subject: [boost] Problems compiling boost_python under Borland C++ Builder 5

> Dear Sirs,
>
> I've just tried to compile the boost_python library under Borland BCB 5
> (Update Pack #1) and got the following error (one of many similar):
>
> [C++ Error] classes.cpp(178): E2316 '__strcmp__' is not a member of 'std'

Boost.Python has never been tested with Borland C++. It appears from the
error message that Borland erroneously fails to define strcmp in std::. If
you would like to submit a workaround for this bug, we'd be happy to have
it. You would need to get the code to try to access strcmp in the global
namespace for BCC.

> It seems a coincidence that last message at the time on the mailing list
> stated the following:
>
> >Ralf Grosse-Kunstleve discovered a bug which will affect a wide class of
> >conforming compilers.
>
> Is this related?

Almost certainly not. The "wide class" of compilers only includes compilers
using a front-end from EDG as of this date (to our knowledge).

> I do need to map my C++ framework to Python *as soon as
> possible* and it does use a good deal of C++ so any help would be most
> appreciated.
> As I see things, these are serious problems.
> But if there appears to be a quick or nearly quick fix to the source code,
> perhaps you could instruct me how to do it (I'd be happy to supply more
> information about the errors). Of course you know that all this means that
the
> source fails to compile on the free Borland compiler (5.5x), too.
> The problem is that I do not have VC 6 at the moment, and even if I had
it,
> I would still need to use Borland's.
>
> Can you give me a solution/piece of advice? Something that just crossed my
> mind is using the MinGW compiler for Win32 since it should be compliant
with
> the Linux GCC you tested, wrap everything in a DLL with Python embedded,
then
> load it in Borland.
> Would this work

It might. It's worth a shot.

> i.e what I ask is can I use MinGW to do it and if I can't
> could you point me to a free compiler because I can't have VC 6
immediately.

MinGW and Cygwin are the only other known free compilers for Windows.

> Of course a fix to the Boost Python code would ease my life a lot.

In this case it's not our code that's broken, but your compiler's standard
library implementation.

> And one last question -- if the code can't be fixed for now, will there be
a
> fixed version (for Borland, since it's a popular good compiler) in the
near
> future i.e. 10 days for example?

Only if someone steps forward to provide it. Would you like to volunteer?

-Dave