Re: [help-GIFT] Gift and gcc 3

[Andreas Enge]

On Tue, 6 Nov 2001, Wolfgang Mueller wrote:

> > Actually, you are right, the problem is with the string classes and
> > persists with the newest version 3.0.2 of gcc and with the CVS version
> > of gift. Now it would be interesting to know whether this is a gcc bug
> > or whether strings are used in a way not conforming to the standards in
> > gift.
> >
> Dear Andreas, thanks for coming back to us. Could you check if the header is
> actually *there*? In the meanwhile I tried to install gcc on a sun (from
> sources), and had the usual compile error (on strings). I checked the
> installation to see what's wrong with the header, and in fact it was not
> there (in my buggy installation), and rightly the installation had complained
> about files being exsistant already and not being able to install.
>
> I would be interested in hearing more from you.

Dear Wolfgang,

sorry for not coming back to you. I had meant to track down the problem,
but was too busy during the past month. Not having used the standard string
class before turned out to be fatal to my motivation to follow the string
ramifications. Actually, I am equipped with too many string header files!
My system is a mixture between Mandrake 8.0, 8.1 and Cooker; gcc and the
standard c++ library in particular are from Cooker. The following header
files look most promising:

>From the package kernel-headers-2.4.3-20mdk:
/usr/include/asm/string.h
/usr/include/linux/string.h

>From the package glibc-devel-2.2.4-8mdk:
/usr/include/string.h
/usr/include/bits/string.h

>From the package libstdc++3.0-devel-3.0.2-1mdk:
/usr/include/g++-v3/bits/basic_string.h
/usr/include/g++-v3/bits/std_string.h
/usr/include/g++-v3/bits/std_cstring.h
/usr/include/g++-v3/string

I guess the very last file is the crucial one; unless _CPP_STRING is defined
(how would one know?), it includes bits/std_string.h, which itself
triggers a cascade of inclusions (unless other symbols are defined).

It would surprise me if these observations were of much help, but maybe
they are not completely useless...

Cheers,

Andreas


_______________________________________________________________________________

Andreas Enge
Laboratoire d'Informatique (LIX)
Ecole Polytechnique
91128 Palaiseau CEDEX
France

Phone:  +33 1 69 33 34 79
Fax:    +33 1 69 33 30 14
E-mail: address@hidden
URL:    http://www.math.uni-augsburg.de/~enge