[BUG] DSA failure in LargePositiveInteger in #new:<primitive: 71>
JArchibald at aol.com
JArchibald at aol.com
Wed May 31 06:57:01 UTC 2000
=> 5/30/00 5:23:25 PM EDT, sr at evolgo.de =>
<< I think the problem is clear: "To normalize or not to normalize!">>
Cute. However, it appears with new LargeIntegers package that normalize is
essential (in some cases), where with old LargePositiveIntegers it was
optional (in some cases).
<<I'm not motivated to make a full rework of the DSA stuff and would like to
hear John's opinion about this topic... >>
It sounds like John Maloney understands the problem (see full text of
following E-mail), and was taking certain liberties with LargePositiveInteger
for performance reasons.
=> 5/30/00 7:18:42 PM EDT, John.Maloney at disney.com =>
<< Incidentally, I'm about to vanish for a six week vacation in Europe. I'm
afraid I won't be able to do much about this myself before I go. >>
Stephan and John,
John's description seems quite consistent with Stephan's. So unless someone
is willing to submit a temporary fix as a standard fileIn for the DSA code
(perhaps a performance hit by a factor of 2 is good enough for six weeks), I
would say that it's broken and should not be used until John gets back. Other
opinions?
One more fox out of one more rabbit hole :-),
Jerry.
____________________________
Jerry L. Archibald
systemObjectivesIncorporated
____________________________
More information about the Squeak-dev
mailing list
|