Re: PDUMP vs. UNION_TYPE + C++
>>>> "OG" == Olivier Galibert
<galibert(a)pobox.com> writes:
> 2. XEmacs crashes for me as follows:
OG> [...]
> What I don't understand is how the
lrecord_implementations_table can
> possibly be resurrected in the post-dump temacs. It seems like all
> the pointers in the lrecord_implementations_table need to be
> relocated, and perhaps the structures they point to need to be
> re-created. I don't see where this is being done. I must be missing
> something.
OG> The structures they point to are global variables and global variables
OG> aren't supposed to move around...
Problem solved:
All I had to do was
rm temacs xemacs.dmp; make
So what needs to happen is some makefile hacking. But that goes
without saying seeing as the Makefile still assumes there will be an
`xemacs' created. In general, the binary should make sure that the
dump file `belongs' to that particular binary. One obvious way would
be to have some unique identifier in each temacs that would get
written to the temacs.dmp file as well. Also, there should be a
Makefile dependency of the .dmp file on temacs. Maybe it's time to do
all the future Makefile hacking now to get everything working. We are
still doing stupid stuff like this:
Dumping under the name xemacs
Testing for Lisp shadows ...
make: ./xemacs: Command not found
You can use
#ifdef PDUMP
...
#endif
in Makefile.in.in
Otherwise more people will get confused about this...
Martin