Re: congrats! -- extended dynamic object loading from AMIGAOS to WIN3 2_NATIVE & POSIX

[Sam S. <sd...@gn...>]

> * In message <9F8582E37B2EE5498E76392AEDDCD3FE05F4BA26@G8PQD.blf01.telekom.de>
> * On the subject of "congrats! -- extended dynamic object loading from AMIGAOS to WIN3   2_NATIVE & POSIX"
> * Sent on Tue, 5 Aug 2003 16:20:21 +0200
> * Honorable "Hoehle, Joerg-Cyril" <Joe...@t-...> writes:
>
> Another case of worse is better! :)
I don't see why my approach is `worse'.

> a) being able to restart foreign functions from previous images, and
I tested it and it works.

> That'll certainly not work on MS-Windows or UNIX!!
> There's no guarantee I know of about the addresses returned by
> dlsym(), and even if simple tests (a la "let's just start it twice and
> see if the values differ") seem to indicate otherwise, I'll construct
> a situation where it'll crash (by using a different ordering of dlopen
> or different linking libraries, or using a custom DL_PRELOAD etc.). 
> I don't know about MS-Windows, but it's not likely to work either.

I see.

options:

1. keep library name in the Ffunction & Fvariable structs and call
   dlsym() lazily - when the object is accessed.

2. keep the list of objects loaded from the library in the
   O(foreign_libraries) alist and call dlsym() for each of them when
   the library is updated.

   opinions?

> Also, I haven't seen just from looking at your foreign1.diffs whether
> compiling a file with dynamic library call definitions still generates
> a .c file as a (then superfluous) side effect.

fixed that now.  thanks.
you just don't call NOTE-C-FUN and NOTE-C-VAR for dynamic objects.

> I didn't even know then that Sam had ported my Amiga code!
I did not look at your Amiga code.


-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html>
Why use Windows, when there are Doors?