[Ecls-list] ECL on L4

Julian Stecklina der_julian at web.de
Mon Aug 29 10:47:29 UTC 2005


On Mon, 29 Aug 2005 18:49:33 +0200
Juan Jose Garcia Ripoll <lisp at arrakis.es> wrote:

> On Mon, 2005-08-29 at 18:30 +0200, Julian Stecklina wrote:
> > I'll have to take a look at dietlibc to decide how "decent" it
> > actually is. :)
> 
> Long long ago, I managed to compile ECL with dietlibc on linux :-)

I take this as a positive sign. :)

> > Virtual memory magic is quite difficult in plain L4, as you have to
> > implement your pager yourself.
> 
> Hmm, I would rather begin without virtual memory, just ensure that you
> have a permanently allocated block of memory, of fixed size, and that
> ECL knows the location and size of this block. I would force ECL to
> signal an error if the memory is exhausted, so that it does not exceed
> the limits of this block.
>
> Once the garbage collector works in such a limited environment, you
> may begin worrying about doing more elaborated things with the memory
> and the garbage collect, like paging memory in and out to disk,
> implementing read/write barriers, improving the collector algorithm,
> or simply dynamically resizing memory as requirements increase.

Sounds like a reasonable approach. I hope I get some help from the L4
people at my university. :)

> BTW, sounds like an interesting task.

Indeed.

Regards,
-- 
Julian Stecklina

(Of course SML does have its weaknesses, but by comparison, a
 discussion of C++'s strengths and flaws always sounds like an
 argument about whether one should face north or east when one
 is sacrificing one's goat to the rain god.)         -- Thant Tessman




More information about the ecl-devel mailing list