Front page | perl.perl5.porters |
Postings from February 2000
Gurusamy Sarathy writes:
> for their hashes, but use Perl's internal hash table routines. This might
> make sense if you have a custom data set on which Perl's algorithm performs
> poorly (as it is known to on keys that are repeating sequences).
This remark is a year (or two) too late.
There are no known "natural" vulnerabilities of the current hashing
code. Note "hash quality" below, and the fill-histograms on the ARRAY
line. [Well, x 6 or x 26 give much worse results, but still not
catastrophic.]
Ilya
monk:~/perl/perl5.5.640->./perl -Ilib -MDevel::Peek -e '$h{$_ x 8}++ for "a" ..
"z"; Dump \%h'
SV = RV(0x19d810) at 0x184834
REFCNT = 1
FLAGS = (TEMP,ROK)
RV = 0x1a407c
SV = PVHV(0x194438) at 0x1a407c
REFCNT = 2
FLAGS = (SHAREKEYS)
IV = 26
NV = 0
ARRAY = 0x184230 (0:15, 1:8, 2:9)
hash quality = 107.1%
KEYS = 26
FILL = 17
MAX = 31
RITER = -1
EITER = 0x0
Elt "qqqqqqqq" HASH = 0x55f22a80
SV = NV(0x194098) at 0x1a7798
REFCNT = 1
FLAGS = (NOK,pNOK)
NV = 1
Elt "bbbbbbbb" HASH = 0xdb6302e0
SV = NV(0x194020) at 0x17f220
REFCNT = 1
FLAGS = (NOK,pNOK)
NV = 1
Elt "yyyyyyyy" HASH = 0x31fa3fa2
SV = NV(0x1940d8) at 0x18481c
REFCNT = 1
FLAGS = (NOK,pNOK)
NV = 1