Re: What to do with modules now ...
"J. Kean Johnston" <jkj(a)sco.com> writes:
All,
Once the module loading technology is in place and people are happier
with it, I would like to start work on ripping out major parts of XEmacs
and putting them in modules. I havem in my prototype tree, a mechanism
in place where a given source file can either be compiled in as part
of the dumped Emacs, or as a module. If module loading is availible,
the makefile is constructed in such a way that it compiles things as
modules rather than part of the dumped Emacs. This is a good thing.
It is not only possible, but preferable (I believe) to make large parts of
XEmacs modules. It will also make distribution a lot easier. I am working
on making the TTY and X11 code loadable, so that you only ever load in
those parts that you care about.
Doesn't this (demand paging?) happen on virtually every operating system
out there? (given some sort of assumption about locality in some way)
I guess that you could argue "no" because some sort of initialization
that has to occur. (could this be done lazily to get some of the same
effect?)
This gives users choices they didnt have before (eg, I think I'll
try
loading the Lucid widgets instead of the mOtif ones)
This, I buy. This would be very nice to have.
and it reduces
the
size of the in-memory footprint of Emacs considerably.
See above. Does this _really_ matter so much? The fact is that it _is_
all shared, and it _is_ demand paged, so this is sort of a possibly
shaky win. Would it _really_ be worth the work?
Soren