Re: Lisp object that refers to a C struct

[Eli Zaretskii <eliz@gnu.org>]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: Juanma Barranquero <lekktu@gmail.com>,  eggert@cs.ucla.edu,  emacs-devel@gnu.org
> Date: Thu, 18 Oct 2012 08:33:59 -0400
> 
> >> In that case, wouldn't it be better to have the integer be an index
> >> into a table, like Stefan suggested?
> > It's easy to do that (of course, at a price of some slightly more
> > complicated memory management),
> 
> I don't understand why the memory management would care about which
> integer you use to represent file-watcher descriptors.

It doesn't.  I meant the need to manage the table itself, grow it when
needed, etc.

> > but I think Stefan still prefers the "new object" solution.
> 
> I generally like stronger types, so you're probably right, but an "index
> into a table" doesn't necessarily sound so terribly bad, if you have
> such a table anyway

I don't have a table.  What I have is an association list.  I used the
same method as Rüdiger in his inotify patch.

> (BTW, is this table usable by Lisp, e.g. can they get the list of
> current file-watchers?).

Yes, it's an alist.  But it currently isn't exposed to Lisp, neither
as a variable nor via an API.  It could be, of course.  But since
Rüdiger didn't, and since we didn't discuss in detail how this feature
will be used from Lisp, I didn't want to introduce APIs whose
necessity is not clear.

> >> It's faster to check that it is in range than compare it with an
> >> arbitrarily long list of known pointers.
> > That's not what the code does.  It calls Fassoc_quit to find the

I meant Fassoc, of course.

> > descriptor in the list of known live watches (which will still be
> > needed under the table suggestion),
> 
> Isn't Fassoc_quit going to "compare it with an arbitrarily long list"?
> To me this "list of known live watches" sounds exactly like the "table"
> I mention above.

Yes, but since it's an alist, I get its memory management for free.

> > and then validates the pointer itself, just in case (which is also
> > independent of the table suggestion).
> 
> What means "validate"?

Call w32_valid_pointer_p, and in addition verify that the struct
pointed to by it has the correct magic signature.