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

Using timezone: Central European Time
* zzorn leaves03:16
* CIA-7 leaves07:52
* CIA-7 joins08:04
* CIA-7 leaves08:07
* CIA-7 joins08:10
* zzorn joins09:41
* zzorn_ joins09:55
* zzorn leaves09:59
* zzorn joins10:00
* zzorn_ leaves
* arjanb joins10:15
* zzorn leaves11:18
* zzorn joins11:20
* zzorn_ joins11:43
* zzorn leaves11:49
* zzorn joins11:50
* zzorn_ leaves
<CIA-7>03arjanb * 10Nice/src/bossa/ (18 files in 3 dirs): Converted the last piece of NiceClass.12:12
* zzorn leaves12:15
* bangoola leaves
* zzorn joins12:20
* bangoola joins
* zzorn_ joins16:10
* zzorn leaves16:26
<CIA-7>03arjanb * 10Nice/src/bossa/syntax/ (27 files): Conversion cleanup part 1.17:42
* bonniot joins18:04
<arjanb>hi18:07
<bonniot>hi18:09
<arjanb>what do you think of the suggestion of bug #1096734 to add the expression in assertionfailed message?
<bonniot>i'm not sure. the expression will often be useless if you don't look at the source code anyway18:11
<arjanb>otoh a expression is at least as usefull as an empty message18:15
i noticed location are still missing sometimes so i'm thinking about making location a required parameter for construtors of Expression, Definition and Statement18:19
<bonniot>you can add testcases when you find missing locations18:20
<arjanb>that's not possible for (nullness) warnings18:21
<bonniot>right, but then it should be possible to change the code so that it's an error with the same expression18:22
<arjanb>btw what does it mean if the name of a VarSymbol is null?18:24
<bonniot>it's anonymous18:25
<arjanb>how can that happen?18:26
*away for dinner18:33
<bonniot>in method implementations, if you specify the type but not the name of a parameter18:38
f(@Foo) in the old syntax
not sure if it's possible in the new one18:39
but maybe with foo(null) = ...
or foo(1) = ...
<arjanb>i see19:08
the nullness annotations in the compiler need some more work..19:09
<CIA-7>03arjanb * 10Nice/src/ (7 files in 4 dirs): Removed some dead code.19:54
<arjanb>any reason that overriding a final java field isn't possible?21:32
<bonniot>overriding the type?21:38
<arjanb>yes
<bonniot>that field might be set by java code21:39
<arjanb>right though it can be convenient21:41
<bonniot>do you have a concrete example?21:54
* CIA-7 leaves21:55
* CIA-7 joins22:07
<arjanb>well i wanted to make 'name' notnull in PolySymbol but some notNull's work too..22:20
<CIA-7>03arjanb * 10Nice/src/bossa/syntax/ (27 files): Conversion cleanup part 2.23:47
* arjanb is tired of chasing null's23:48
<bonniot>that should be easier once the conversion is over, right?00:24
<arjanb>conversion is almost complete but finding where null could come from is the problem00:27
<bonniot>if you change the type, compilation errors should give you the answer00:50
<arjanb>true but then it turns out it depends on something else and changing that type gives another error and ...00:51
it's a little annoying finding after 3 or 4 steps that the type must be different00:53
<bonniot>normally it should also be chosen by design01:01
whether it makes sense for that value to be null or not
of course sometimes it can be null for technical reasons only
but then, if this is in an exceptional case, it can also make sense to declare the type as not-null, and use a cast in the exceptional condition01:02
<arjanb>i try only to use cast for thing that are local01:04
the types that seems logical to me are mostly to precise for all uses but then i haven't written most of the code01:06
<bonniot>with the conversion, you're getting close ;-)
<arjanb>40k java and 520k nice in bossa.syntax atm01:07
<bonniot>cool01:09
i have to go. goodnight
<arjanb>g'night01:10
* bonniot leaves01:15
* CIA-7 leaves01:31
* CIA-7 joins01:43