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

Using timezone: Central European Time
* CIA-3 joins03:10
* CIA-3 leaves03:20
* CIA-3 joins03:21
* CIA-3 leaves03:23
* CIA-7 joins
* CIA-7 leaves03:27
* CIA-2 joins
* CIA-2 leaves03:33
* CIA-6 joins
* CIA-6 leaves03:34
* CIA-6 joins
* CIA-6 leaves
* CIA-3 joins03:35
* CIA-3 leaves03:36
* CIA-3 joins03:37
* CIA-3 leaves
* CIA-6 joins
* CIA-6 leaves03:43
* CIA-3 joins03:44
* CIA-3 leaves03:50
* CIA-7 joins
* CIA-7 leaves03:53
* CIA-3 joins
* CIA-3 leaves
* CIA-7 joins03:54
* CIA-7 leaves04:00
* CIA-3 joins04:01
* CIA-3 leaves04:17
* CIA-7 joins04:18
* CIA-7 leaves04:34
* CIA-1 joins
* bonniot joins09:52
* arjanb joins11:47
* CIA-1 leaves14:03
* CIA-5 joins
<bonniot>hello arjan14:55
<arjanb>hi
<bonniot>i'm working on improving the support for nullness markers on type variables:
class A<T>
{
?T x;
}
<arjanb>so smart handling of duplicate nullness markers14:58
<bonniot>yes14:59
the present situation is not very intuitive15:03
<arjanb>another issue with nullness markers is that you can't use in the raw form15:05
<bonniot>example?15:06
<arjanb>for example 'option or' can't be expressed in a single definition
<bonniot>what about:15:09
<T,U | U <: ?T> U `||`(?T e1, U e2);
<arjanb>hmm why has option or 2 defitions then?15:13
<bonniot>maybe i didn't think about it
<arjanb>i'm wondering whether all cases can be expressed in this way15:15
or would a nullness T = nullness U constraint be useful
<bonniot>another option is to have a syntax for constructing a full type based on a raw type and a type constructor marker15:17
good night22:34
* bonniot leaves

Generated by Sualtam