9:58 a.m.
Robert Pluim <rpluim(a)bigfoot.com> writes:
Here's a (possibly insane) suggestion: rather than dumping, why
don't
we compile the lisp files to C? We could then make select portions
dynamically loadable, and remove all that unportable dumping code.
This was Kyle's original plan. Olivier thinks it better to write the
raw data directly to a file that you can mmap in (i.e. skip the
compile stage).
A middle road would be to write to an elf .so object directly so you
can use the dynamic loader to get it back in.
The first thing you need for this are pointers that are truly
pointers, which is why minimal tag bits were invented. With the effects
of also increasing the range of integers and (with the current naive
allocator) a 50% memory overhead.
Jan
10:40 a.m.
New subject: XEmacs has just (more than) doubled in size
>>>> "Jan" == Jan Vroonhof
<vroonhof(a)math.ethz.ch> writes:
Jan> which is why minimal tag bits were invented. With the effects
Jan> of also increasing the range of integers and (with the
Jan> current naive allocator) a 50% memory overhead.
Well, now we know where half of Steve's bloat came from!
Seriously, what denominator did you have in mind for that 50%?
--
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
__________________________________________________________________________
__________________________________________________________________________
What are those two straight lines for? "Free software rules."
11:02 a.m.
New subject: XEmacs has just (more than) doubled in size
Jan Vroonhof writes:
Robert Pluim <rpluim(a)bigfoot.com> writes:
> Here's a (possibly insane) suggestion: rather than dumping, why don't
> we compile the lisp files to C? We could then make select portions
> dynamically loadable, and remove all that unportable dumping code.
This was Kyle's original plan.
I thought Kyle's plan was to write out a C _representation_ of the
parsed lisp, not to actually compile the lisp code to C? (I may be
misremembering here).
Olivier thinks it better to write the
raw data directly to a file that you can mmap in (i.e. skip the
compile stage).
A middle road would be to write to an elf .so object directly so you
can use the dynamic loader to get it back in.
Hmm, I have no idea which alternative would be better. As long as
dumping disappears, I'll be happy.
Robert
--
The above are my opinions,
and my opinions only.
10:30 p.m.
New subject: XEmacs has just (more than) doubled in size
Jan Vroonhof writes:
Robert Pluim <rpluim(a)bigfoot.com> writes:
> Here's a (possibly insane) suggestion: rather than dumping, why don't
> we compile the lisp files to C? We could then make select portions
> dynamically loadable, and remove all that unportable dumping code.
This was Kyle's original plan.
Sort of. I think Robert meant that we would translate Lisp code
to C code. My plan was to translate Lisp_Objects into C data
structures after temacs -l loadup does all the initialization.
Then we compile the new .c files that contain the structs and
link the new .o files with the other .o's to produce the final
XEmacs binary.
6:06 a.m.
New subject: XEmacs has just (more than) doubled in size
Kyle Jones writes:
Sort of. I think Robert meant that we would translate Lisp code
to C code. My plan was to translate Lisp_Objects into C data
structures after temacs -l loadup does all the initialization.
Then we compile the new .c files that contain the structs and
link the new .o files with the other .o's to produce the final
XEmacs binary.
Is the project on hold now or has it just been re-priorized?
--Marcus
7:44 a.m.
New subject: XEmacs has just (more than) doubled in size
Marcus Thiessel writes:
Kyle Jones writes:
>
> Sort of. I think Robert meant that we would translate Lisp code
> to C code. My plan was to translate Lisp_Objects into C data
> structures after temacs -l loadup does all the initialization.
> Then we compile the new .c files that contain the structs and
> link the new .o files with the other .o's to produce the final
> XEmacs binary.
Is the project on hold now or has it just been re-priorized?
I changed jobs and got swamped and didn't have the energy to
work on it. That's why I have not been deeply involved in
XEmacs development these last months although I still have a
keen interest in it. At last report Olivier Galibert has been
working on his portable dumper idea and hopefully he can get it
working before he gets swamped.
6:31 p.m.
New subject: XEmacs has just (more than) doubled in size
On Thu, Aug 19, 1999 at 11:44:55PM -0700, Kyle Jones wrote:
At last report Olivier Galibert has been
working on his portable dumper idea and hopefully he can get it
working before he gets swamped.
I hope so too. I've lost touch last month (5-10K lines of code
written for the project in two months ;-) but I hope to get back to it
quickly.
OG.
9227
days inactive
9228
days old
6 comments
6 participants
participants (6)
-
Jan Vroonhof -
Kyle Jones -
Marcus Thiessel -
Olivier Galibert -
Robert Pluim -
Stephen J. Turnbull