lazy initialization

[Han-Wen Nienhuys]

[CC to mailing list]

>> On 2nd thought, this isn't entirely true. The bitvectors for all
>> cards are set from the init routine, see
>> (scm_i_initialize_heap_segment_data), but that could be postponed to
>> (scm_i_sweep_some_cards() ).
>
>So, this would be a nice project, no?  We could increase the heap
>segment size (and thereby the number of heap segments) without
>increasing our real memory use.

Added this to CVS; now a new segment of memory is only touched when it
is needed.

BTW, check-guile nowadays reports 

  UNRESOLVED: eval.test: promises: basic promise behaviour: a forced promise 
does not reference its environment


-- 

Han-Wen Nienhuys   |   address@hidden   |   http://www.cs.uu.nl/~hanwen