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

Using timezone: Central European Time
* leblin joins00:59
arjan, did you see my problem on the help forum?01:06
<arjanb>hello01:39
<leblin>hi
<arjanb>yes
<leblin>i've reduced the problem to a single line
let StringBuffer v = new StringBuffer().append( "A" ).append( 'a' );01:40
<arjanb>it's the second append right?01:42
<leblin>yes... one append alone is okay.
<arjanb>StringBuffer has changed quite a lot in java 1.5 with the addition of the Appendable interface and the AbstractStringBuilder superclass01:47
<leblin>yes, they have a StringBuilder class now too01:49
<arjanb>nicec doesn't understand the covariant returntype correctly yet01:50
and there might be a difference in overloading resolution in case of boxed primitive
<leblin>but the primitive doesn't need to be boxed in this case01:52
<arjanb>but it could be01:54
first argument is more specific on one method and the second on the other01:55
<leblin>right however it will compile with no errors without the first append01:56
<arjanb>yes Stringbuffer has another append method defined that more specific than both you see in the error message01:59
writing out all chained stringbuffer calls could make nice bootstrap on 1.502:00
but it might be easier to wait for daniel to make nicec understand the new bytecode constructs02:02
<leblin>probably, i just installed jre 1.402:04
the makefile is not windows compatible.02:05
<arjanb>yeah i found it easier rewrite it to a batchfile than getting it work in cygwin02:07
<leblin>well the nicec and makefile uses : to delimit paths however you need to use ; in windows02:09
<arjanb>did try to build nicec yourself because you wanted to make changes to it?02:13
<leblin>yes, i would like to join the team.. i was going to try to write nicedoc last year however life got really complicated, and i had to forget about it.02:15
i have a sucessful build using 1.4 02:16
<arjanb>great02:17
<leblin>the eclipse plugin does not work correctly using eclipse 3.0.1.. i was going to see if i can find the problems with it to start02:18
<arjanb>who made a start with nicedoc has switched jobs and don't seem to have time for it
<leblin>is anyone maintaining that plugin?02:20
<arjanb>no it has been on pause for a year now02:21
<leblin>i will email the author about it02:28
<arjanb>you could mail the modified makefile to daniel02:36
<leblin>whats his address?02:37
<arjanb>Daniel>DOT<Bonniot>AT<inria>DOT<fr02:40
<leblin>ok thanks
do you have the email for the author of the eclipse plugin?02:41
<arjanb>daniel has it surely and he knows more about the plugin02:43
<leblin>i'll ask him thanks.
<arjanb>other questions for today? (it's getting late here)02:51
<leblin>no thanks for your help02:52
<arjanb>good night
* arjanb leaves
* leblin leaves03:28
* arjanb joins10:15
* cm_ joins12:12
hi arjanb12:13
<arjanb>hi cm_
<cm_>Heard anything from bonn?12:17
<arjanb>no12:18
he is responding to email though12:21
<cm_>ok12:23
Is the examperiod over?12:24
<arjanb>yes :)12:25
<cm_>Nice, did you pass?12:28
<arjanb>i have doubts about that..12:30
anyway i have time for coding now
do you know darcs?12:33
<cm_>Darcs? A human being in flesh and blood?12:49
<arjanb>no http://abridgegame.org/darcs/
<cm_>I read about it a long time ago. Written in haskell I recall12:53
Have you tried it out?12:54
<arjanb>yes and i'm surprised at how easy it is to use12:55
<cm_>Have you tried subversion?13:01
<arjanb>no13:03
<cm_>Very easy to use as well. I have minor CVS knowledge but the drawbacks seems to be very annoying 13:05
I was referring to the drawbacks of CVS and not subversion :-)13:07
<arjanb>darcs has the additional advantage of being distributed13:09
<cm_>How are conflicts resolved?13:10
<arjanb>darcs is almost 1.0 now and i think it could make nice development easier for others13:11
<cm_>No central repository then? (and no single point of failure)13:12
<arjanb>right13:13
<cm_>Let say A and B is modifying foo.c. When commiting only C and D are available and the file is commited without any conflicts. How do darcs cope with the conflict when everyone is connected and available?13:17
<arjanb>there no commiting13:19
everyone has their on copy/branch
*own
<cm_>Do you merge different repos. to make a release and cope with conflicts at that point?13:44
<arjanb>repos aren't merged, you can either pull from or send patches to another repo13:49
if there is a conflict you can resolve that and make it patch13:50
<cm_>Isn't difficult to get the global picture of the project then?14:05
<arjanb>the repo where releases are made from is the most important one14:11
everyone will try to get their patches into that repo so that's close a global picture14:12
<cm_>So it is preferable to have a central and always connected repository then I guess.15:08
<arjanb>indeed15:17
<cm_>Is it possible to run it at sourceforge?16:00
<arjanb>it has no server thus it might be possible16:04
the only thing you need is an html readable dir and some way to get patches there16:05
* Vurlix joins16:58
hi... discovered Nice yesterday, am extremely pleased with the language -- it's exactly what i've been trying to put together16:59
but it could use a compiler that produces native code, so i'll probably work on that :)17:00
<arjanb>hello17:05
you can use Nice in combination with gcj17:06
Vurlix do you have experience with creating compilers?17:12
<Vurlix>lots of experience with parsers (without lex/yacc et al), but not compilers per se17:23
though i've made a regular expression "parser/compiler/executer"17:24
<arjanb>i see17:26
<Vurlix>hmmm... how does a gcj natively compiled nice program compare with an equivalent C program?
<arjanb>i don't know any benchmark comparing C with gcj17:29
<Vurlix>because i'm really looking for a language powerful enough to do high level programming and fast enough to avoid low level languages in performance critical scenarios17:30
currently C++ is the only such language and it's quite pitiful17:31
<arjanb>for most applications there's no big difference between C/C++ and jvm nowadays17:32
<Vurlix>jvm is catching up in speed, but it's a monstrosity with memory usage17:37
<cm_>Java performs very well if you avoid some common pitfalls like memory management, eg. void getPosition(Vector v) instead of Vector getPosition(). Compare with stack allocated instances in c++, eg const Vector& getPosition().17:38
<Vurlix>heh
<arjanb>true but memory usage isn't that important anymore except for embedded stuff17:39
<Vurlix>uhhh
when a java app uses 300 MB of memory where a very similar app that does the same thing, written in C, uses 10 MB.... there's a problem
<arjanb>in that case the java version is badly written17:40
<cm_>As they often are :-)
<Vurlix>heh.. that app in particular is azureus17:41
still...17:45
<arjanb>never heard of azureus17:46
i think programmer productivity is more important17:49
most performance comes from algorithms and critical parts can still be written in C17:50
<cm_>Ordo - what? Asymptotic behaviour?18:03
<arjanb>heh wrong channel?18:04
<cm_>Nope, just a joke :-)18:05
<arjanb>*away for dinner18:07
<Vurlix>hmm18:24
i guess i should try writing a small app with Nice and see how it goes18:25
i never liked the gcc suite, though18:39
<arjanb>*back18:40
<Vurlix>i might go ahead and make a nice compiler anyway
<arjanb>what's the problem with gcc?
<Vurlix>it produces huge executables (i've even seen 20MB ones)18:41
it's quite difficult to make *anything* under 500k with it
i'm not comfortable with that18:42
* CIA-7 leaves18:44
* CIA-2 joins
<Vurlix>i think i'll base it off lcc and adapt it for nice18:48
that way it can produce code for practically any target
<arjanb>i think you will end up with another bytecode to native compiler..18:49
try to get more familiar with Nice first18:52
<Vurlix>yeah, i'm still reading docs :)19:00
and wiki
<arjanb>some wiki pages could outdated19:05
* cm_ leaves19:33
<arjanb>Vurlix: feel free to ask questions and we're interested in feedback too20:25
* cm_ joins20:29

Generated by Sualtam