Re: MPS codegen

["Gerd Möllmann" <gerd.moellmann@gmail.com>]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Gerd Möllmann <gerd.moellmann@gmail.com>
>> Cc: Eli Zaretskii <eliz@gnu.org>,  acorallo@gnu.org,  emacs-devel@gnu.org
>> Date: Sun, 16 Jun 2024 08:16:38 +0200
>> 
>> Gerd Möllmann <gerd.moellmann@gmail.com> writes:
>> 
>> >   pdumper.c:4937: Emacs fatal error: assertion failed: rc == 0
>> >   gmake[3]: *** [Makefile:335: emacs-lisp/byte-opt.elc] Abort trap: 6
>> >   gmake[3]: *** Waiting for unfinished jobs....
>> >
>> > which is soothing.
>> >
>> > I'm reluctant to invest more time though because this all si obsolete
>> > with your change.
>> 
>> Just had an inspiration what it might be, and it is indeed the case:
>> 
>> Forwarded symbols are created like this
>> 
>>   #define DEFVAR_LISP_NOPRO(lname, vname, doc)	\
>>     do {						\
>>       static struct Lisp_Objfwd const o_fwd	\
>>         = {Lisp_Fwd_Obj, &globals.f_##vname};	\
>>       defvar_lisp_nopro (&o_fwd, lname);		\
>>     } while (false)
>> 
>> The pdumper copies these static structs to the dump, and when the dump
>> is loaded, the symbol's forwarding pointer is changed from pointing to
>> the static struct to the copy to the dump. Both have the same contents.
>> 
>> IOW, the static structs are no longer used. Consequenlty, when accessing
>> the structs to get their type (XFWDTYPE) we are accessing something in
>> the dump.
>> 
>> Why pdumper is doing that escapes me ATM.
>
> Because the values recorded during dumping could be different from the
> default values, and so using the static struct might be wrong?

I don't know, it makes no sense to me.

The only values the static struct contains is the type (Lisp_Fwd_Obj in
the example above, and the pointer to the variable we forward to
(&globals.f##vname in the example). I don't understand how these could
be not be right in the static struct if the dumper just left them alone.
I'd even assume that things would go haywire if the copy in the dump
didn't have the exact same values that the static struct has. Confusing.

@Helmut: forgot to mention that explicitly, but I think that might be
relevant to what you are planning to do...