From: Andrew Crabtree <andrewc AT typhoon DOT rose DOT hp DOT com>
Message-Id: <199707100338.AA215105932@typhoon.rose.hp.com>
Subject: Re: ctor ?
To: billc AT blackmagic DOT tait DOT co DOT nz
Date: Wed, 09 Jul 1997 20:38:51 PDT
Cc: djgpp AT delorie DOT com
In-Reply-To: <33C4F7AB.248E@blackmagic.tait.co.nz>; from "Bill Currie" at Jul 10, 97 2:54 pm
Precedence: bulk

> I'll try (and it DOES work correctly).

As seems to be typical for me, immediately after I hit the send button
for a djlist message I come up with a new idea to try that
usually (this time anyway) straightens things out.
 
> static void __attribute__((constructor)) set_x(void)
Ahh, i didn't think it looked like valid C code.  Leave it to gnu
to have an attribute to do it.

> It should be working for C++ (I've never had any problems).
Turns out my libc was broken (again...).  I think I must have left
it compiled by my unreleased pgcc 0621 snapshot, which was really lousy.  
I pulled _main.c into my test directory and compiled it into my 
test application and everything worked as it should.

>  But I have
> no idea what gcc does with what you were doing in C.
Doesn't compile.

> (actually, gcc places a call to _main() at the beginning of main(), but
> libc calls _main itself (don't know why (dj, can you say why?)) and so
> _main() also has a test to see if it has done the .ctor calls yet).
I noticed this too while debugging, strange...

> As to why you can't break on f(), I don't know.  If it was with the C
I guess I just don't understand the label names.  I tried to type 
break _GLOBAL_$I$f__Fv 
which was exactly what I saw in the assembly file.  GDB just wants
the f__Fv part though.  I examined the first pointer from djgpp_first_ctor
and then told gdb to break on that absolute address and it told me it was
line 4 in function f__Fv which clued me in.

Now that I've got this fixed I can go back to trying to figure out
why my pgcc isn't emitting a .section ctor at the end of the 
file when it should..

Thanks

Andrew