Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 31 Mar 2002 13:40:07 -0800
From:      Terry Lambert <tlambert2@mindspring.com>
To:        Alexander Leidinger <Alexander@Leidinger.net>
Cc:        nick-lists@netability.ie, hackers@FreeBSD.ORG, ports@FreeBSD.ORG, cy@FreeBSD.ORG
Subject:   Re: lang/icc doesn't compile c++ sources
Message-ID:  <3CA78237.7576200C@mindspring.com>
References:  <200203311752.g2VHqLFL006965@Magelan.Leidinger.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Alexander Leidinger wrote:
> ---snip---
> (7) netchild@ttyp1 % LD_PRELOAD= icc -c test.cc
> test.cc:
> "/usr/local/intel/compiler50/ia32/include/cwchar", line 16: error: the
> global scope has no "tm"
>   using ::mbstate_t; using ::size_t; using ::tm; using ::wint_t;

[ ... ]

These are complaining about missing "this" members, specifically
"tm", "btowc", "mbrlen", "mbrtowc", "mbsrtowcs".


The "tm" isn't obvious (maybe "to multibyte"?).  I'd like to see
a standards reference for "byte to wide character", "multibyte r(?)
length", "multibyte r(?) to wide character", "multibyte set r(?) to
wide character set".

Specifically, these member functions assume implementation that
is not required by the standards, which is that multibyte
characters are not 16 bit wide characters when they are
externalized, but are instead encoded some way.  Further, it's
not clear whether this is UTF-7, UTF-8, EUC, or some other
encoding that it's assuming.  I'll guess it's UTF-8, for the
necessary Windows compatability.

In general, this should not be an issue, in any case.

Specifically, it should not be an issue because I don't see
you actually *using* instances of a wchar_t class, and thus
it's really not believable that the it would be upset about
class membership for a class which is not instanced.

Probably this means you are still getting a mix of GNU and
ICC C++ header files, and need to carefully look at your
include paths.


> > (I'm waiting a little bit to see if I need to do something for the C++
> > problem).
> 
> I think this is not a problem which should get solved in the port, so I
> commit the new portrevision soon.

I'm not sure this is possible, without an additional "brand"
type; see other posting.

-- Terry

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3CA78237.7576200C>