sam on DEC OSF/1 Alpha?

sam on DEC OSF/1 Alpha?

[Chris Siebenmann]

 Has anyone got sam (especially samterm) running on this platform?
 sam itself seems to work fine for me, but samterm is core dumping in
odd places for apparently inexplicable reasons (dbx is being unhelpful).

	- cks

Re: sam on DEC OSF/1 Alpha?

[Chris Siebenmann]

 This is probably a naive question, but is there any reason the
protocol has to pass around the addresses instead of tokens, except
for convenience and to avoid another lookup table? As far as I can
tell from a quick scan, sam itself never uses the fact that the
magic tokens are addresses, it just passes them around; only samterm
cares. Indeed, it might be that only a relatively minor modification
to samterm/mesg.c would be needed.

 Am I missing something here?

	- cks

Re: sam on DEC OSF/1 Alpha?

[brendan]

>  Well, thanks to John Mackin, I now have the problem identified.
> Pointers in DEC Alpha OSF/1 are 64 bits (as is 'long'), and the sam to
> samterm protocol passes pointers around, lopped to 32 bits. Naturally
> this causes problems when samterm gets a lopped pointer back and
> attempts to use it.
> 
>  Unfortunately, I'm not quite sure of the best way to fix this.
> Time to go spelunkering the sources.

One of the best ways to handle this that I've found is to prototype
all of the functions, then find where integers are being passed as
pointers, and visa-versa.  (ptrs are 64 bits, ints are 32...if you
pass `0' where a pointer is expected, it'll be bogus) Then when you
compile with an ANSIish compiler (use gcc :) ), it'll tell you about
all of the violations.

Brendan

--
Brendan Kehoe                                               brendan@cygnus.com
Cygnus Support, Mountain View, CA                              +1 415 903 1400
                     ``In a cruel and imperfect world,'' says critic Rex Reed,
 ``[Audrey Hepburn] was living proof that God could still create perfection.''

Re: sam on DEC OSF/1 Alpha?

[Chris Siebenmann]

 Well, thanks to John Mackin, I now have the problem identified.
Pointers in DEC Alpha OSF/1 are 64 bits (as is 'long'), and the sam to
samterm protocol passes pointers around, lopped to 32 bits. Naturally
this causes problems when samterm gets a lopped pointer back and
attempts to use it.

 Unfortunately, I'm not quite sure of the best way to fix this.
Time to go spelunkering the sources.

	- cks