|
Boost : |
From: Toon Knapen (toon.knapen_at_[hidden])
Date: 2002-02-07 10:43:21
Steve M. Robbins wrote:
> On Thu, Feb 07, 2002 at 03:40:54PM +0100, Toon Knapen wrote:
>
>>John Maddock wrote:
>>
>>
>>>>I'm sorry, my previous msg was not correct, I've been a bit too hasty.
>>>>
>>>>I experienced the problem on an SGI but not with CC but with gcc 2.95.2.
>>>>
>>>>
>>>Does that mean the fix is required or not? Can you predicate it on sgi
>>>version? I assume later version have conformant iostreams or don't they?
>>>It would seem perverse if not, given that they have an open source version
>>>available.
>>>
>>
>>well, the only gcc 2.95.2 I have access to is on this SGI. Using
>>gcc-2.95.2 there is no <sstream> but there is a <strstream>.
>>
>>I just read the comments in /boost/config/stdlib/sgi.hpp where the
>>BOOST_NO_STRINGSTREAM is set and it is indeed documented that it needs
>>to be set for gcc-2.95.2 and not for gcc-2.95.3 but that there is no way
>>to detect them. So I guess it's a known problem without a proper solution.
>>
>
> On the contrary, there is a well-known solution: detect at build
> time, e.g. using autoconfigure.
>
> c.f. http://ac-archive.sourceforge.net/C++_Support/ac_cxx_have_sstream.html
that is also mentioned in the comment in /boost/config/stdlib/sgi.hpp
;-) but boosters generally do not use configure.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk