Jim Ursetto scripsit:
> As per the Chicken README, symbol GC is disabled by default,
> but may be enabled globally at compiler build time, or per-app
> at runtime.
Thanks. How about a primitive procedure, then to set and retrieve the
value of C_enable_gcweak?
--
Knowledge studies others / Wisd
Alaric Snell-Pym scripsit:
> Is that necessarily the best solution to your problem, though?
> Anything with an identity (a cons cell, for example) will work as a
> 'token' for a-list lookups?
The idea is to speed up string lookups. Currently the a-list maps strings
to values, which means I must
On 22 Aug 2009, at 11:43 pm, John Cowan wrote:
If symbols aren't gc'd now, I'd like to argue for them to be. In an
application I'm writing, I want to generate lots of random symbols at
runtime so they can be looked up more quickly in a-lists, but I don't
necessarily want them hanging around fo
Hello,
Currently, the build aborts with the following error:
gcc -Wl,--enable-auto-import -shared -Wl,--out-implib,libchicken.dll.a \
-o libchicken.dll library.o eval.o data-structures.o ports.o files.o e
xtras.o lolevel.o utils.o tcp.o srfi-1.o srfi-4.o srfi-13.o srfi-14.o srfi-18.o
sr
On Sun, Aug 23, 2009 at 1:41 PM, Fadi Moukayed wrote:
> Out of curiosity, why is Chicken using [random/srandom] instead of the
> standard C rand/srand ? There shouldn't be any difference
> functionality-wise AFAIK.
rand(3) typically produces much less usable randomness than
random(3) due to histo
On Sun, Aug 23, 2009 at 05:38:10PM -0400, John Cowan wrote:
> Alaric Snell-Pym scripsit:
>
> > Is that necessarily the best solution to your problem, though?
> > Anything with an identity (a cons cell, for example) will work as a
> > 'token' for a-list lookups?
>
> The idea is to speed up string
Andrew Reilly scripsit:
> If you're going to have a *lot* of these strings (and you are,
> or you wouldn't be worried about the GC angle), then you could
> be better off with a "real" hash table or r/b-tree or similar
> structure. Even with the assoc -> assq conversion your search
> is still O(N)
