Changeset 7138


Ignore:
Timestamp:
Sep 5, 2007, 1:21:10 AM (12 years ago)
Author:
gb
Message:
 
File:
1 edited

Legend:

Unmodified
Added
Removed
  • trunk/ccl/ChangeLog

    r7130 r7138  
     12007-09-04 19:13  gb
     2
     3        * lisp-kernel/: gc.h, memory.c, pmcl-kernel.c, ppc-exceptions.c,
     4        ppc-gc.c, x86-exceptions.c, x86-gc.c: Lose "HeapHighWaterMark",
     5        which was supposed to allow us to track the high end of a range of
     6        pages that'd been written to and not freed (so we'd zero them
     7        lazily when they were reallocated.)  Such a scheme would really
     8        need to track both ends of such a range, and the old scheme wound
     9        up being overly zealous and often zeroed pages that were already
     10        zeroed.  Zero the range between the old free pointer and the new
     11        one after each GC, instead.
     12       
     13        At least partly address ticket:101, by doing GROW-DYNAMIC-AREA more
     14        carefully.
     15
     162007-09-04 19:12  gb
     17
     18        * level-0/: l0-numbers.lisp, PPC/ppc-numbers.lisp: Fix to 64-bit
     19        %NEXT-RANDOM-SEED, random algorithm comments from Bernd Beuster.
     20
     212007-09-04 19:11  gb
     22
     23        * lib/: ffi-darwinppc32.lisp, ffi-darwinppc64.lisp,
     24        ffi-linuxppc32.lisp, ffi-linuxppc64.lisp, macros.lisp: Don't coerce
     25        single-float results to double-floats in DEFCALLBACK-BODY.  Do do
     26        this in the PPC backends.
     27       
     28        (This needs testing ...)
     29
     302007-09-04 19:10  gb
     31
     32        * compiler/X86/X8664/x8664-backend.lisp: Don't assume that
     33        :single-float callback results have been coerced to :double-float
     34        on x8664.
     35
    1362007-09-03 18:10  gb
    237
Note: See TracChangeset for help on using the changeset viewer.