[H-GEN] g++/egcs peculiarity

[Craig Eldershaw]

Hi All,
 
OK, here's a short C++ program:
 
=========================================
class myclass
{
public:
  myclass(int x, int y)
  {}
 
};
 
 
int main()
{
  int *first=new int(5);       // create an int
  int *second=new int[10](5);  // create 10 ints
 
  myclass *one=new myclass(1,2);      // create a myclass
  myclass *two=new myclass[10](1,2);  // create 10 myclasses
}
===========================================
 
Now this should be fairly straight forward:
We create a int, initialise it to 5, and point first to it.
We create an array of ints, initialise all to 5, and point second to it.
We create a myclass, with initialisation (1,2), and point one to it.
We create an array of myclasses, initialise all to (1,2), and point two to it.
 
Now, if I compile this using gcc (ver. 2.7.2.[12]) then all is OK.
Also, to prove that it's doing what I want, chucking in some printf's
shows that the initialisation is being handled correctly for both int's
and myclass's.
 
I recently upgraded gcc/g++ on my Debian/Linux machine to egcs.  On
it, compilation now fails with:
 
>g++ crash.cc
crash.cc: In function `int main(...)':
crash.cc:16: no matching function for call to `myclass::myclass (int)'
crash.cc:7: candidates are: myclass::myclass(const myclass &)
crash.cc:5:                 myclass::myclass(int, int)
>g++ -v
Reading specs from /usr/lib/gcc-lib/i486-linux/egcs-2.90.28/specs
gcc version egcs-2.90.28 980423 (egcs-1.0.3 prerelease)
>
 
It compiles cleanly with "-Wall -pedantic -ansi" on the older gcc's
(ignoring the "unused variable" comments).  However on the same problem
machine, egcs reports:
 
>g++ -Wall -pedantic -ansi crash.cc
crash.cc: In function `int main()':
crash.cc:13: warning: initialization in array new
crash.cc:16: warning: initialization in array new
crash.cc:16: no matching function for call to `myclass::myclass (int)'
crash.cc:7: candidates are: myclass::myclass(const myclass &)
crash.cc:5:                 myclass::myclass(int, int)
crash.cc:16: warning: unused variable `class myclass * two'
crash.cc:15: warning: unused variable `class myclass * one'
crash.cc:13: warning: unused variable `int * second'
crash.cc:12: warning: unused variable `int * first'
>
 
Have they changed the standard recently concerning initialising arrays ?
Egcs whinges, but still seems to accept it - if the myclass lines are
ommitted, then it compiles and runs correctly (well adding in some
printf's shows that it is doing the initialisation of second
correctly).
 
Or am I just missing something obvious ?
 
Any theories appreciated.
 
Cheers,
        Craig.
 

----------------------- HUMBUG General List --------------------------------
echo "unsubscribe general" | mail majordomo at humbug.org.au # To Unsubscribe