Re: [boost] MSVC6/Intel C++ 5 with Dinkumware STL 3.x

From: Ken shaw (ken_at_[hidden])
Date: 2001-04-27 12:39:51

One initial problem you will be having is that the Dinkumware 3.08 library's
implementation of hash_(multi)set and hash_(multi)map are different than the
ones expected by the regression test and therefore I assume the entire boost
library. In particular the classes need a different number of template
arguments. So you will need to use the BOOST_NO_HASH macro at least.

    On this issue, why is the regression test and and config.hpp concerned
with classes outside of the library standard. While I hope that at some
point hash classes are in the standard until then should this matter? The
library requirements specify standard C++ so no boost code should be using
these classes.

Ken Shaw
Programming Manager
Computer Innovations
Chicago IL
[phone]
----- Original Message -----
From: "Kirk Klobe" <kklobe_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, April 27, 2001 11:35 AM
Subject: RE: [boost] MSVC6/Intel C++ 5 with Dinkumware STL 3.x

> >>In order that others can help you, please upload the full set of
> >>error message output from the regression tests to the boost files
> >>section (compressed, of course).
>
> I have uploaded regression_dinkumware.zip under the folder
> regression_results and included the output from regression.exe along
> with the config.hpp I'm using.  For the results, I stripped compiler.cfg
> down to support only MSVC and Intel C++.
>
> >>
> >>Reading boost/config.hpp, there's no point in changing the
> >>compiler-specific bug macros.  You may want to focus on these
> >>library-specific macros, some of which are already set
> >>differently for STLport, for example.
>
> <SNIP>
>
> Sorry, I wasn't clear in my message.  These are the BOOST_NO_x macros I
> was referring to.
>
> I'm currently only concerned with the Intel results, and simply included
> the MSVC results for comparison (although getting MSVC working with the
> supposedly more compliant Dinkumware 3.08 wouldn't hurt).
>
> Any feedback or help is much appreciated!
>
> - Kirk
>
>
> To unsubscribe, send email to: <mailto:boost-unsubscribe_at_[hidden]>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>