🚀 go-pugleaf

In article <9203312049.AA15699@ori.cadlab.de> joost@cadlab.uni-paderborn.de writes:
>Due to the fact that the creation of static (and some local) objects has
>changed greatly, the target 'install' fails in the directory 'test-install'.
>
>In gcc-2.0 (and cfront, I believe) the proper constructor (in this case
>Foo(char*)) was called with 'this' = <address of static object>. Now, the
>'proper' constructor is called for an local (stack) address *and* afterwards
>the real static object is created by calling the copy-constructor, taking
>'this' the real static address and as second arg the 'dummy' object created on
>the stack. Afterwards the destructor is called for the local object.
>
>So, it does not look wrong, but awkward. Is this intended?

I would echo the above and add the following wrt the rest of the
test-install process.  Two other progams are made, a and b, and both
run.  A produces the output "Hello world!"; b produces:

Hello  world!
Errno:  1

There is no indication of the expected value of errno - is 1 correct??
I would also add that the the current documentation with libg++-2.0
seems to be out of sync, ie. protions of it relate to earlier versions
of the library and as such can lead to confusion.  Having said this, I
now have what I believe to be an installed and working version of
libg++-2.0 installed - I will start using it in anger over the next
few days.  My machine specs are:

	Processor: sparc  Manfacturer: sun  OS: sunos4.1.1

Regards,
Allan Clearwaters
MCT Ltd., 62 Albert Rd. Nortn, Malvern, Worcs. WR14 2TL, England
allan@mct.uucp