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

Using timezone: Central European Time
* bonniot leaves03:15
* ChanServ leaves06:46
* CIA-9 leaves
* arjanb leaves
* ChanServ joins06:47
* arjanb joins
* CIA-9 joins
* arjanb leaves06:49
* bonniot joins08:58
* arjanb joins12:36
<bonniot>hi arjanb13:24
what happened with notNull ?
<arjanb>i haven't tried yet13:26
<bonniot>another thing before I forget:13:27
in this testcase:
/// FAIL no-location
/// package p1
/// Toplevel
interface I {}
interface I1 extends I {}
interface I2 extends I {}
private String foo(I);
foo(I1 i1) = "1";
foo(I2 i2) = "2";
/// package p2 import p1
println(foo(new A()));13:28
/// Toplevel
public class A implements I1,I2 {}
/foo(A a) = super(I2);
the ambiguity error message still uses the old @ syntax for patterns
that should be changed
<arjanb>is this one in the testsuite?13:33
the problem is that no syntax for nameless type patterns exist anymore13:51
<CIA-9>03arjanb * 10Nice/src/bossa/syntax/pattern.nice: Avoid using '@' in error messages.14:47
<bonniot>no, it's not in the testsuite. I was just checking rohan's claim that some ambiguities are not detected15:08
what is the problem?15:14
<arjanb>that the name of a pattern isn't always known15:15
<CIA-9>03arjanb * 10Nice/src/bossa/syntax/ (31 files): Nullness cleanup.17:40
<arjanb>it didn't help much17:43
<bonniot>do you have an example about patterns?22:33
it didn't help much?? something like 100 notNull removed, it's great!22:53
<arjanb>still over 500 left23:01
<bonniot>16% improvement is not neglectable ;-)23:09
i just don't understand the change to scope.nice. how is that related to notNull ?
<arjanb>i removed a few superflous nullness tests too23:11
<bonniot>but in that case, i think res can be null23:19
it's just the default retyping that hides the fact23:21
lookup and globalLookup should be retyped23:22
this code should be put back, but that will allow getting rid of the cast(null) just afterwards

Generated by Sualtam