odd machine configs: mips-siemens, gec63, pfa50, att3b

odd machine configs: mips-siemens, gec63, pfa50, att3b

[Ken Raeburn]

Is anyone familiar enough with these machines to answer a couple
questions?


The mips-siemens config defines XSETINT in a way that looks broken; it
appears to retain the old type of the object being set as the new
type, instead of setting it to Lisp_Int.  Does Emacs even work on this
system?  I'm inclined to just delete the macro, but I'd prefer
confirmation.


In the gec63, pfa50, and att3b config files, some configurations play
some interesting games with VALBITS, VALMASK, and the like.  My guess
is that this code predates DATA_SEG_BITS, and it would be enough to
set DATA_SEG_BITS appropriately and let VALBITS and the rest be set by
lisp.h.  And I doubt the att3b definition of XINT is correct any more,
even if the comment is.  Can anyone confim my suspicions?  Maybe even
test a patch (or if they've got a little time, write it)?

Ken

Re: odd machine configs: mips-siemens, gec63, pfa50, att3b

[Ken Raeburn]

> The mips-siemens config defines XSETINT in a way that looks broken; it
> appears to retain the old type of the object being set as the new
> type, instead of setting it to Lisp_Int.  Does Emacs even work on this
> system?  I'm inclined to just delete the macro, but I'd prefer
> confirmation.

I just discovered this was a problem in mips.h too and was cleaned up
a while back.  I'll just go ahead and delete it.

Re: odd machine configs: mips-siemens, gec63, pfa50, att3b

[Richard Stallman]

    In the gec63, pfa50, and att3b config files, some configurations play
    some interesting games with VALBITS, VALMASK, and the like.  My guess
    is that this code predates DATA_SEG_BITS, and it would be enough to
    set DATA_SEG_BITS appropriately and let VALBITS and the rest be set by
    lisp.h.

You are probably right.  Those machines are quite old, and it could be
that nobody has one any more.  You may as well give this change a try
if it will help you proceed.