Original bug ID: 564
Reporter: administrator
Status: closed
Resolution: fixed
Priority: normal
Severity: minor
Category: ~DO NOT USE (was: OCaml general)

Bug description

En utilisant libcamlrund.a pour déboguer un truc, j'ai eu ça (bien
avant que mon bogue n'arrive):
/usr/local/adm/src/caml/bin/cash tflush.ml

O'Caml runtime: debug mode

Initial minor heap size: 128k bytes
Initial major heap size: 248k bytes
Initial space overhead: 42%
Initial max overhead: 1000000%
Initial heap increment: 248k bytes
Initial stack limit: 1024k bytes
Starting new major GC cycle

O'Caml runtime: heap check

!<>$<Growing heap to 496k bytes
Growing page table to 390 entries

$file alloc.c; line 36 ### Assertion failed: wosize > 0

Avec un breakpoint dans caml_failed_assert, la trace de la pile est:
Breakpoint 1, caml_failed_assert (expr=0x80728e9 "wosize > 0",
file=0x80728e1 "alloc.c", line=36) at misc.c:27
gdb:bt
#0 caml_failed_assert (expr=0x80728e9 "wosize > 0", file=0x80728e1 "alloc.c",
line=36) at misc.c:27
#1 0x80587e1 in alloc (wosize=0, tag=0) at alloc.c:36
#2 0x8058b24 in alloc_tuple (n=0) at alloc.c:64
#3 0x8069c82 in dynlink_get_current_libs (unit=1) at dynlink.c:215
#4 0x806bc85 in interprete (prog=0x401dc008, prog_size=1080172)
at interp.c:830
#5 0x806d3c5 in caml_main (argv=0xbffff174) at startup.c:390
#6 0x8057c02 in main (argc=2, argv=0xbffff174) at main.c:41

et dans dynlink_get_current_libs, on fait effectivement un alloc (0),
car:
#3 0x8069c82 in dynlink_get_current_libs (unit=1) at dynlink.c:215
gdb:p shared_libs
$1 = {size = 0, capacity = 8, contents = 0x807a158}
---------^^^^^^^^

alloc.c compilé sans -DDEBUG ne semble pas s'en porter plus mal, par
contre.

Bruno.