Log of the #nice channel on irc.freenode.net

Using timezone: Central European Time
* arjanb leaves01:55
* bonniot leaves02:03
* zzorn leaves03:46
* arjanb joins06:25
* zdk joins11:21
hello11:24
<arjanb>hi11:25
<zdk>Can I ask about building compiler here?
<arjanb>yes
<zdk>I've been doing on it.11:26
arjanb, do use lex/yacc?11:38
arjanb, do u use lex/yacc?
arjanb, ...?12:11
* zdk leaves12:38
* bonniot joins12:51
* ArtemGr joins15:22
* ArtemGr leaves15:28
* ArtemGr joins15:35
* zzorn joins16:50
* ArtemGr leaves
* zzorn leaves17:21
* zzorn joins17:27
<bonniot>i'm looking at #109757117:36
<arjanb>*away for meal17:52
i think it doesn't make much sense for long to extend float19:19
because of loss of precision the conversion should be explicit19:21
and with 2 groups of numeric classes one could have BigNumber as superclass and the other BigDecimal19:22
<bonniot>how is it in Java19:27
?
* zzorn_ joins
<arjanb>in java is long a subtype of float19:33
but i think that implicit conversion is rarely used
* zzorn leaves19:41
<bonniot>but int cannot fit on a float either in general...19:44
<arjanb>so i would like to seperate the integers from the floats19:45
<bonniot>i guess short fits on float and int on double19:47
if you did that you couldn't do 'float f = 0;'
0 is a byte
<arjanb>that could be written as float f = 0.0;19:48
<bonniot>sure, but that's a pain20:21
<arjanb>yes, overloading on returntype like in haskell could be useful here20:25
<bonniot>well, 0 is not a method, so it would need special handling21:30
i think it might make sense to have short <: float, int <: double21:31
* zzorn_ leaves22:25

Generated by Sualtam