Digital Mars - c++ - STLport and Boost Progress

↑ ↓ ← → Christof Meerwald <cmeerw web.de> writes:

Let's start with the good news:

DM 8.32.9 is able to compile the STLport library with namespace support
enabled.

But not all STLport regression tests can be compiled with the current DM
beta (but I hope it's just one namespace bug I reported earlier).

The patch is available from http://cmeerw.org/prog/dm/stl.html
(http://cmeerw.org/files/dm/STLport-4.5.3-dm.diff). In addition to enabling
namespace support, I have renamed the STLport libraries and I am setting
default library information in config/stl_dm.h.

Of course, DM still doesn't support explicit template instantiation, so
unresolved external symbols are still likely when using STLport's iostream
library.


Now for the bad (or at least not so good) news:

Getting Boost to work with Digital Mars C++ is a bit more difficult - some
of the Boost libraries depend on the Boost preprocessor library which
currently doesn't work (and tracking down bugs in the preprocessor is a bit
difficult).

But some small parts of Boost are already working (smart pointers and parts
of boost.utility and boost.threads). And I got jam/bjam compiling with DM.

See http://cmeerw.org/prog/dm/boost.html for my latest patch.


And yes, I am also trying to get omniORB 4.0 working with DM. I got omniidl
compiling with DM, but it doesn't work yet (no, it's not that FILE * passing
problem). I have successfully compiled the omnithread library, but the ORB
core doesn't compile yet (but I guess it soon will, it doesn't look that
bad).


bye, Christof

-- 
http://cmeerw.org                                 JID: cmeerw jabber.at
mailto cmeerw at web.de

...and what have you contributed to the Net?

Dec 09 2002

↑ ↓ ← → "Rajiv Bhagwat" <dataflow vsnl.com> writes:

Both of you are doing an excellent job.. Just watching you at this speed
makes us happy.
- Rajiv

"Walter" <walter digitalmars.com> wrote in message
news:at3ioc$1f05$1 digitaldaemon.com...
 You are doing a fantastic job. I can't thank you enough. -Walter

Dec 09 2002

↑ ↓ ← → Richard <fractal clark.net> writes:

In article <at39o4$10eb$1 digitaldaemon.com>, Christof Meerwald says...
Let's start with the good news:

DM 8.32.9 is able to compile the STLport library with namespace support
enabled.


You are amazing! Congrats.

Richard

Dec 10 2002

↑ ↓ ← → Richard <fractal clark.net> writes:

The patch is available from http://cmeerw.org/prog/dm/stl.html


typo.. should be cd STLport-4.5.3\src not cd STLport-4.5.3\lib

As a general note, beware of dm\bin\sc.ini settings if you use environment
settings on windows for things like INCLUDE and LIB. LIB environment setting is
apparently ignored and INCLUDE can cause conflict.

Stlport libs and dll compiled just fine after applying the patch.

Namespace support is MUCH better. All my test cases compiled, and I had to
actually specify using std::cout; - very exciting stuff.

wow.

Richard

Dec 10 2002

↑ ↓ ← → Richard <fractal clark.net> writes:

slist appears to be dead due to undefined symbol.. list, algo, func and others
work fine though :)) You mentioned that some things might not work due to this..
but this seems a little odd. Here is the error from beta 10:

#include <slist>

using namespace std;

void main()
{
slist<int> s;
}

Error 42: Symbol Undefined ?__splice_after ?$_Sl_global std _N std  SAXPAU_Slis
t_node_base 2 00 Z (void cdecl std::_N::d::_Sl_global<>::__splice_after(_Slist_n
ode_base::d *,_Slist_node_base::d *,_Slist_node_base::d *))
SMAKE fatal error: command "LINK" returned with error code 1
Stopping.

Richard

Dec 13 2002

↑ ↓ ← → Richard <fractal clark.net> writes:

I'm having a little trouble with deque.

This code used to work, but perhaps it was wrong all along. Anyway:

#include <list>
#include <deque>

using namespace std;

void main()
{
typedef list<int> Li;
typedef Li::iterator Liter;
typedef deque<Liter> Di;
typedef Di::iterator Diter;
Li l;
Liter lit = l.insert(l.begin(), 10);
Di d;
Diter dit = d.insert(d.begin(), lit);
}

c:\dm\stlport\stl/_deque.h(93) : Error: integer constant expression expected

Richard

Dec 13 2002

↑ ↓ ← → Christof Meerwald <cmeerw web.de> writes:

On Fri, 13 Dec 2002 13:44:30 +0000 (UTC), Richard wrote:
 I'm having a little trouble with deque.


Here is a self-contained test-case:

template<class T>
struct A
{
  enum E
  {
    val = sizeof(T) + 1
    // Error: integer constant expression expected
  };
};

template<class T>
struct B
{ };

int main()
{
  typedef A<B<int> > C;
  C::E e;

  return 0;
}


bye, Christof

-- 
http://cmeerw.org                                 JID: cmeerw jabber.at
mailto cmeerw at web.de

...and what have you contributed to the Net?

Dec 13 2002

↑ ↓ ← → Christof Meerwald <cmeerw web.de> writes:

On Mon, 9 Dec 2002 23:41:57 +0000 (UTC), Christof Meerwald wrote:
 And yes, I am also trying to get omniORB 4.0 working with DM. I got omniidl
 compiling with DM, but it doesn't work yet (no, it's not that FILE * passing
 problem). I have successfully compiled the omnithread library, but the ORB
 core doesn't compile yet (but I guess it soon will, it doesn't look that
 bad).


I have been able to get omniORB 4.0.0 working with DMC 8.32.14 (I haven't
done much testing, but at least omniNames seems to work...)

The patch is currently available at
http://svn.cmeerw.net/src/dmc/omniORB-4.0.0-dm.diff


bye, Christof

-- 
http://cmeerw.org                                 JID: cmeerw jabber.at
mailto cmeerw at web.de

...and what have you contributed to the Net?

Dec 31 2002