Re: Undumpable object type : buffer, extent

"Stephen J. Turnbull" <[email protected]> wrote:

> /.../  For example, it's possible that if you killed all the
> extents (including text properties) used, the problem would be
> solved.  Note: I don't recommend you try it on my say-so, though, it
> would likely be a lot of effort, and not quite work in the end.

Someone who can try this out is welcome to do so and send me a patch.

> Well, to be precise, for this purpose, compile-time is included in
> run-time, AFAIK.  Ie if you can wrap in it (eval-when (compile eval)
> ...), Ilya wins, I bet.

Having (eval-when (eval) ...) can hardly make any difference. As for
compile-time evaluation, that would increase the byte code dependency
on the specific (X)Emacs version even more, so I'd rather avoid that.
(c-emacs-features test for stuff that has changed between minor
versions in some cases, and it might also be used to test for bugs in
very specific versions.)

> I don't know if it's worth rearranging things to make that possible,
> just so cc-mode can be dumped in an XEmacs configured --pdump, but I
> think it should be possible.

Well, afaics the problem with CC Mode is only a symptom of a bug in
the dumper, so adding a kludge to CC Mode doesn't really solve the
problem (although, as I said above, I can add a kludge if someone
gives me a patch and it doesn't have too severe side-effects). If CC
Mode had stored buffer or extent objects directly in constants then it
would be another matter, but in this case the result is just a simple
list of symbols.