🚀 go-pugleaf

Hi...
	I'm trying to put gcc up on a concurrent computer, and
I've run into a few questions that i was hoping one of you might
be able to help me out with...

1) gas - i've been told that this is a "trivial port"... does this mean
that the vanilla distribution from you guys would come up without any
necessary modifications from concurrent?  i know this one may not be for you
to answer...

2) is there a compiler option or a make variable or something to get the 
compiler to pack structures?  for example, the following structure

struct{
	short;
	char *;
} a;

{	char, char;
	struct a;
	int	x[2];
}
is represented in the code as

----|----|----|----|
char|char|xxxx|xxxx|
----|----|----|----|
  short  |xxxx|xxxx|
----|----|----|----|
   char pointer    |
----|----|----|----|
    integer [0]    | 
----|----|----|----|
    integer [1]    | 
----|----|----|----|

what i would like is a way to make it come out like this:

----|----|----|----|
char|char|  short  |
----|----|----|----|
   char pointer    |
----|----|----|----|
    integer [0]    | 
----|----|----|----|
    integer [1]    | 
----|----|----|----|

	the VAX C compiler does this, as do various other C compilers
i've worked with.  I looked through the optimization flags, but i couldnt
find anything about it... i've also tried fiddling various flags in the
configuration of gcc to no avail...

3) what are the references fltused and gcc_compiled. used for?  my first
little test program for gcc produces more efficient code than the concurrent
c compiler, but those two references, though unused in the program cause
an extra 2k of library routines to be brought in, making the resulting image
larger.  Is there any way to get rid of them?

4) with all of those optimization options, it's hard for a human to know what
to do... which -f options have sizable effects?

	i'll appreciate any and all responses to these questions a great deal!
we're getting fed up with the concurrent c compiler and would like to work with
gcc... our other option is to compile all of our programs on a Phillips Delta
series computer to the assembler level, ship them over to the concurrent 
(filtering the files for assembler imcompatibilities) and assembling them
there... yeccch....

thanks

mark modrall
mwm@greystone.com

"an equal opportunity individual"