develooper Front page | perl.perl5.porters | Postings from April 2010

Re: -Duse64bitint pros and cons

Thread Previous | Thread Next
From:
Sisyphus
Date:
April 15, 2010 04:05
Subject:
Re: -Duse64bitint pros and cons
Message ID:
1959DCC8FDE74D6ABB77C83272B90040@desktop2

----- Original Message ----- 
From: "Niko Tyni" <ntyni@debian.org>

> I'd be glad to hear any opinions and recommendations on this.

Since you asked, it still bothers me that, on a perl built 
*with* -Duse64bitint but *without* -Duselongdouble, the expression (and 
similar):

(2**58)+1 == (2**58)+3

returns true. (That will always bother me.)

On a perl built with *both* -Duse64bitint and -Duselongdouble, the same 
expression returns false .... with which I am quite content.

Admittedly this is all documented, and workarounds exist. But the above 
expression consists of very fundamental operators - and I expect better 
DWIMmery from fundamental operators. Ok, that said ... I'm not complaining 
... the solution is obvious. Those, such as myself, who are unhappy with 
that behaviour should simply build with both -Duse64bitint 
and -Duselongdouble (or make use of the -Dusemorebits abbreviation) ... and 
that's what I do ... and that's what I'm suggesting you consider doing.

I must also admit that I haven't yet checked out this particular behaviour 
wrt 5.12, and I'm merely assuming that it hasn't changed from perl-5.10 and 
earlier. Someone please correct me if it *has* changed .... and then also 
explain to me why it wasn't changed earlier ;-)

Cheers,
Rob 


Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About