with better template support.

www.digitalmars.com/dmc/dmcppDownload.html

Thanks, that's a step... er, a leap, in the right direction! - Rajiv

"Walter" <walter@digitalmars.com> wrote in message news:ag0l69$d1e$1@digitaldaemon.com...
> with better template support.
>
> www.digitalmars.com/dmc/dmcppDownload.html
>
>
>

On Wed, 3 Jul 2002 22:11:40 -0700, Walter wrote:
> with better template support.

Just tried to use STLport-4.5.3 with DMC 8.29.5, but DMC doesn't like this class template partial specialization:

  template <class A, class B>
  struct IsSame { enum { _Ret = 0 }; };

  template <class A>
  struct IsSame<A, A> { enum { _Ret = 1 }; };
  // Error: specialization arguments for template 'IsSame' match primary



bye, Christof

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

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

Here is another bug in 8.29.5:


template<class T>
struct A
{ };

struct B { };

template<> // it compiles if you comment out this line
struct A<B>
{
  A<B> *g()
  { return this; }
};


void f()
{
  A<B> a;
  // Internal error: template 2145
}



bye, Christof

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

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

These are good bug reports. Thanks!

"Christof Meerwald" <cmeerw@web.de> wrote in message news:ag269t$1t54$1@digitaldaemon.com...
> On Wed, 3 Jul 2002 22:11:40 -0700, Walter wrote:
> > with better template support.
>
> Just tried to use STLport-4.5.3 with DMC 8.29.5, but DMC doesn't like this class template partial specialization:
>
>   template <class A, class B>
>   struct IsSame { enum { _Ret = 0 }; };
>
>   template <class A>
>   struct IsSame<A, A> { enum { _Ret = 1 }; };
>   // Error: specialization arguments for template 'IsSame' match primary
>
>
>
> bye, Christof
>
> --
> http://cmeerw.org                                 JID: cmeerw@jabber.at mailto cmeerw at web.de
>
> ...and what have you contributed to the Net?

here is another bug:

template <class T1, class T2>
struct pair
{
  pair(const T1 &a, const T2 &b)
  { }

  template <class U1, class U2>
  // Error: no type for argument 'U1'
  pair(const pair<U1, U2> &p)
  { }
};

typedef pair<pair<int, int>, int> T;

void f(const T &a)
{ }


int main()
{
  f(T(pair<int, int>(1, 2), 3));

  return 0;
}



bye, Christof

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

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

Walter,

Thanks for the new template support.

However I'm having a problem :<. The compiler runs out of memory but does not give an indication as to why or where.

Is there anything I can turn on to find out where the compiler has got to so that I can attempt to create an RBT.

I am trying to build the very latest STLport from www.stlport.org, and thus pushing the new template code in the compiler rather hard. I don't mind if the output is large or slows the compiler down, just so that I can narrow the problem down to a few lines of code.

I know that you have listed the new template functionality, but could you please elaborate the actual acceptable syntax that you have implemented, as it is proving difficult to work out what may not have been implemented. I think I may have found a few problems but I am not totally sure that I am not expeecting too much for now. I will try to create some examples that demonstrate the problems (but not the crash one), but it may take me a while as I want to check against g++ (v3.03) and GreenHills (3.5) just to make sure it's not me!
Regards
Damian

I appreciate greatly the help you're giving. The way to find out where the compiler is possibly hanging is to compile with -e -l, and then a .lst file will be created up to the point where it failed. If that doesn't work, the only think left is to try the binary approach (delete the second half of the source, if it still fails, delete the second half of what's left, etc.).

"Damian" <damiandixon@netscape.net> wrote in message news:3D29A93E.4030603@netscape.net...
> Walter,
>
> Thanks for the new template support.
>
> However I'm having a problem :<. The compiler runs out of memory but does not give an indication as to why or where.
>
> Is there anything I can turn on to find out where the compiler has got to so that I can attempt to create an RBT.
>
> I am trying to build the very latest STLport from www.stlport.org, and thus pushing the new template code in the compiler rather hard. I don't mind if the output is large or slows the compiler down, just so that I can narrow the problem down to a few lines of code.
>
> I know that you have listed the new template functionality, but could
> you please elaborate the actual acceptable syntax that you have
> implemented, as it is proving difficult to work out what may not have
> been implemented. I think I may have found a few problems but I am not
> totally sure that I am not expeecting too much for now. I will try to
> create some examples that demonstrate the problems (but not the crash
> one), but it may take me a while as I want to check against g++ (v3.03)
> and GreenHills (3.5) just to make sure it's not me!
> Regards
> Damian
>
>

Maybe we should move this discussion to the c++ group - i've set a Followup-To...

On Mon, 08 Jul 2002 16:01:18 +0100, Damian wrote:
> However I'm having a problem :<. The compiler runs out of memory but does not give an indication as to why or where.

Maybe it's a recursive macro that DMC doesn't correctly handle yet (I think I have reported the bug a few months ago), try to compile/preprocess this test-case:

  #define A(x) x
  #define test A(test)
  test

And yes, STLport uses these kind of constructs...


> I am trying to build the very latest STLport from www.stlport.org, and

I have already made some progress trying to compile STLport-4.5.3. I am even able to compile simple programs like these (with an unofficial 8.29.7n beta):

#include <string>

int main()
{
  std::string a;

  return 0;
}


But as soon as I add some string operations (something like
a.assign("test")), I get an "Internal error: struct 3273".

Walter, can you give me a hint what I should be looking for to track it down?


BTW, I have put a diff against STLport-4.5.3 on my Web-page: http://cmeerw.org/files/dm/STLport-4.5.3-dm.diff (you need the Cygwin tools or a Linux/Unix machine to apply the patch with "patch -sp1 <STLport-4.5.3-dm.diff" inside your STLport-4.5.3 directory)


bye, Christof

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

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

On Mon, 8 Jul 2002 16:47:10 +0000 (UTC), Christof Meerwald wrote:
> But as soon as I add some string operations (something like
> a.assign("test")), I get an "Internal error: struct 3273".

I think I have tracked it down to this test-case:

template <class T>
struct A
{
  typedef A<T> B;

  template <class I> B &f(I a)
  {
    return *this;
  }
};

int main()
{
  A<char> a;
  a.f(5);

  return 0;
}


bye, Christof

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

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