Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Armando Blancas
Just to clarify, I meant the same result as I got on OSX with 64 bits, namely Infinity. Both with 1.4.0. I'm not getting the exception. On Tuesday, June 26, 2012 6:06:13 PM UTC-7, Sean Corfield wrote: > > Makes sense. I got the exception on my System76 Netbook which, > although technically 64-bi

Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Colin Jones
To address the Infinity exception that Sean & AJ are seeing, I think the issue there is that in REPLy / lein repl, we do a `read` of the input, and then `pr-str` the result and send it on to nREPL. So when nREPL goes to re-read the value back in, that's what triggers the compiler error. user=>

Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Sean Corfield
Makes sense. I got the exception on my System76 Netbook which, although technically 64-bit, is mostly running 32-bit software, including Java. On Tue, Jun 26, 2012 at 5:56 PM, Armando Blancas wrote: > Tried it on my old XP laptop and got the same result with 32 bits. > > > On Tuesday, June 26, 20

Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Armando Blancas
Tried it on my old XP laptop and got the same result with 32 bits. On Tuesday, June 26, 2012 2:46:07 PM UTC-7, tbc++ wrote: > > > On Tue, Jun 26, 2012 at 11:25 AM, Armando Blancas > wrote: > >> I don't get the exception on 1.4.0: > > > Is this a 64bit vs 32bit issue? > > Timothy > -- You r

Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Timothy Baldridge
> On Tue, Jun 26, 2012 at 11:25 AM, Armando Blancas wrote: >> I don't get the exception on 1.4.0: Is this a 64bit vs 32bit issue? Timothy -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com

Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Allen Johnson
When I run the repl from leiningen (v2.0.0-preview6) I get the Infinity exception. But when I run the repl straight from the clojure.jar it works without error. java -jar ~/.m2/repository/org/clojure/clojure/1.4.0/clojure-1.4.0.jar AJ On Tue, Jun 26, 2012 at 11:25 AM, Armando Blancas wrote: >

Re: Buggy FP behavior with Clojure 1.3

2012-06-26 Thread Armando Blancas
I don't get the exception on 1.4.0: ~ $ clj Clojure 1.4.0 user=> 1e309 Infinity user=> On Monday, June 25, 2012 11:09:14 PM UTC-7, Sean Corfield wrote: > > On Mon, Jun 25, 2012 at 10:30 PM, dennis zhuang > wrote: > > Added a postfix "M" to make the number as BigDecimal or "N" as a > BigInte

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread Sean Corfield
On Mon, Jun 25, 2012 at 10:30 PM, dennis zhuang wrote: > Added a postfix "M" to make the number as BigDecimal or "N" as a BigInteger: Yes... user=> 1e309M 1E+309M The "Infinity" exception seems wrong but clearly using BigDecimal makes it work. -- Sean A Corfield -- (904) 302-SEAN An Architect'

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread Cedric Greevey
On Tue, Jun 26, 2012 at 1:30 AM, dennis zhuang wrote: > Added a postfix "M" to make the number as BigDecimal or "N" as a BigInteger: That doesn't work here. It's not double literals that are the problem, but computed values that are coming out as doubles. Furthermore, I'd like the code to use Big

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread dennis zhuang
Added a postfix "M" to make the number as BigDecimal or "N" as a BigInteger: user=> 1e308M 1E+308M user=> (class 1e308M) java.math.BigDecimal user=> (* 10.0 1e108M) 1.0E109 2012/6/26 Cedric Greevey > On Tue, Jun 26, 2012 at 1:11 AM, Sean Corfield > wrote: > > On Mon, Jun 25, 2012 at 9:19 PM, C

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread Leif
user> Double/MAX_VALUE 1.7976931348623157E308 >From http://docs.oracle.com/javase/6/docs/api/java/lang/Double.html : "A constant holding the largest positive finite value of type double, (2-2^ -52)ยท2^1023. It is equal to the hexadecimal floating-point literal 0x1.fP+1023 and also equa

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread Cedric Greevey
On Tue, Jun 26, 2012 at 1:11 AM, Sean Corfield wrote: > On Mon, Jun 25, 2012 at 9:19 PM, Cedric Greevey wrote: >> user=> 1e309 >> Infinity > > FWIW, on 1.4.0 I get: > > user=> 1e309 > CompilerException java.lang.RuntimeException: Unable to resolve > symbol: Infinity in this context, compiling:(NO

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread dennis zhuang
user=> Double/MAX_VALUE 1.7976931348623157E308 2012/6/26 Sean Corfield > On Mon, Jun 25, 2012 at 9:19 PM, Cedric Greevey > wrote: > > user=> 1e309 > > Infinity > > FWIW, on 1.4.0 I get: > > user=> 1e309 > CompilerException java.lang.RuntimeException: Unable to resolve > symbol: Infinity in thi

Re: Buggy FP behavior with Clojure 1.3

2012-06-25 Thread Sean Corfield
On Mon, Jun 25, 2012 at 9:19 PM, Cedric Greevey wrote: > user=> 1e309 > Infinity FWIW, on 1.4.0 I get: user=> 1e309 CompilerException java.lang.RuntimeException: Unable to resolve symbol: Infinity in this context, compiling:(NO_SOURCE_PATH:1) -- Sean A Corfield -- (904) 302-SEAN An Architect's

Buggy FP behavior with Clojure 1.3

2012-06-25 Thread Cedric Greevey
user=> 1e308 1.0E308 user=> (* 10.0 1e308) Infinity user=> 1e309 Infinity Whuh? Double precision exponents should go up to 1023 and single precision shouldn't go above 127. There's no floating point format in commonplace use that maxes out its exponent at 308. What is going on here? -- You re