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

Using timezone: Central European Time
* arjanb leaves01:16
* cm_ joins02:29
* cm_ leaves
* CIA-3 leaves08:34
* CIA-3 joins08:35
* bonniot joins09:43
* arjanb joins12:28
<arjanb>on friday you said "I just found a bug that I would like to see fixed before 0.9.9"12:35
which bug would that be?
<bonniot>the dispatch bug i commited testcases for12:37
<arjanb>i see
i think there is not much difference anymore between resolving method declarations and comparing specificity of patterns12:44
<arjanb>maybe that bug can be solved by reusing pattern code12:48
<bonniot>you want to look at it?12:55
<arjanb>yes though i haven't time for that today12:57
*away till evening14:20
* arjanb leaves
* fcb joins16:11
hello Daniel
* fcb leaves16:46
* CIA-3 leaves16:58
* CIA-5 joins
* arjanb joins17:24
* CIA-5 leaves17:52
* CIA-5 joins17:53
* cm_ joins18:36
hi bonniot, arjanb18:37
Have you read about Nemerle? A language for .NET
The blog has some info: http://nemerle.org/blog/18:38
yes, i read about it some time ago
<cm_>Is it interesting?18:45
About the type system and the relation to Java. This is good introduction to generics in Java:18:46
Good example of the invarient "problem"18:47
Do nice have to syntax similar to <? extends Foo> apart from <T <: Foo>?18:51
And don't which one I would prefer: f(List<? extends Foo> or <T <: Foo> f(List<T>)18:52
don't know
<bonniot>yes, it's the same idea, only the syntax differ18:53
<cm_>There are obvious advantages having a syntax similar to Java but on the other hand is Java often tedious (much to write)18:55
<bonniot>yes. we'll have to go through the 1.5 syntax to see what we should support, just to make things simpler when possible (as it has been the philosophy in Nice since the beginning)18:58
<cm_>Could rip some of the examples in the pdf to a nice tutorial19:01
<bonniot>a tutorial could surely be inspired by these examples. copy-pasting them would not be a good idea though19:03
<cm_>About naming constraied types. It is called "bounded wildcards" in Java.
No, but the essence could be copied19:04
<bonniot>about nemerle, i just checked again. the main interesting part seems to be the macros
isn't wildcard only for types with "?" ?
<cm_>Some macrosystems are really powerfull, eg in lisp
Bounded wildcard <? extends Foo>19:05
<arjanb>the java syntax for constraints is quite limiting19:07
<cm_>arjanb: Any example?
<arjanb>you can't have an upper and lower bound at the same time19:08
and contraints between type variables aren't possible19:09
bonniet: Youe right, wildcards are only for types with "?". 19:12
arjanb: This is possible in java: f(List<T> l1, List<? extends T> l2). Isn't there a constraint between these type variables?19:14
Dependency might be more adequate19:15
<arjanb>it's half a year ago i looked at it but i think extends a type variable is not allowed19:16
<cm_>Like: <T, S extends T> f(...)?
<cm_>It is possible
Wonder if you could write, <C extends B, B extends A, A>?19:18
<arjanb>hmm i should read again about the details of java generics19:25
<cm_>A friend on mine told me that there is something similar to the stop problem in Eiffels type system such that the types can't be inferred. This could consequently result in run-time errors. Anything you are familar with?19:27
Something that you can't deside whether a specific member is going to be present, eg methods.19:28
<bonniot>isn't this about their "patch" to the unsafety of their typing of methods (where they use covariance) ?19:29
<cm_>It might be, he mentioned covariance. I have to ask him again. Anyway, nothing similar in nice then I guess.19:30
<bonniot>no. Eiffel's rule is wrong in the first place. It's a good intention to try to make it safe, but it's an afterthought, and it just does not work well19:46
in nice there is no problem: covariance works naturally with multi-methods
<arjanb>one thing i like from nemerle is macros on custom attributes20:36
daniel: the ambiguity bug is for you to fix22:57
it's caused by Pattern.inDomain not handling null patterns correctly22:58
<bonniot>ok, i'll look at it when I find the time23:38
could you document the extended version of get/set syntax?23:40
<CIA-5>03arjanb * 10Nice/NEWS: Multi-argument indexing.00:07
<arjanb>good night01:07
* arjanb leaves

Generated by Sualtam