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

Using timezone: Central European Time
<Bluelive>http://blue.student.utwente.nl/unittest.html 00:16
only for compiled stuff though00:17
* CIA-4 leaves01:02
* CIA-4 joins01:07
* rickjames joins01:33
does anyone know what a compilation unit is?
* rickjames leaves01:39
* XCute4God joins01:57
hi all!! =]
any java programmer here?
* XCute4God leaves01:59
* CIA-4 leaves04:05
* CIA-4 joins
* arjanb joins13:12
* CIA-4 leaves16:08
* bonniot joins17:13
* CIA-3 joins17:16
<bonniot>what are the news?17:17
<arjanb>i have the the generic arrays almost working but i had no time to look at the bitvector bug yet17:21
friday someone came here for sualtam questions
send the generic array things by mail17:42
<bonniot>is it working now?17:47
<bonniot>ok, got the email17:48
<Bluelive>ive started working unittesting alpha classes
seems pretty handy, added a cgi to the webserver for showwing it17:52
<bonniot>arjanb: it's not very practical to have a bunch of files like this17:57
it would be easier to have a patch (output of diff) or a zip/tar file with directory information17:58
do all the files go into stdlib/nice/lang/inline?
<arjanb>no all GenericArray go into nice/lang and SpecialArray in nice/tools/code18:04
*away cooking*18:05
<Bluelive>there was a discussion btw about multimethods on #java on this network, ricky_clarckson seemed very intrested, maybe you can convert him :)
<bonniot>what is "this" network?18:07
<bonniot>were you part of this discussion?18:18
<Bluelive>i made a few remarks, pointed them to your implementation18:37
and who is ricky_clarckson
<Bluelive>not sure, he seems very familiar to java18:38
first hits on google seem to be him18:41
<bonniot>clarkson ?18:42
<Bluelive>ah i made a typo ? :)18:44
daniel should i make a patch of it?19:58
* kdeuser joins20:01
<bonniot>it's ok now, i applied it by hand
<kdeuser>uhh can i ask a question about sualtam?
<kdeuser>i've been having problems...
the whole thing runs...20:02
but no logs come out
i see the bot and everything in the channel
<bonniot>are you sure you are looking at the right place?
<kdeuser>but no matter how long i wait and no matter how much i type
no logs come out...
<bonniot>i think the logs are going in the curent directory
<kdeuser>right place?20:03
<bonniot>how do you start sualtam?
<kdeuser>in my /var/www/html/log/ directory
$JAVACMD -jar /var/www/html/log/sualtam.jar
<bonniot>you cd into /var/www/html/log/ before?20:04
<bonniot>what about writing rights?
<kdeuser>i was running sualtam as a service....20:05
wouldn't that give sualtam root rights?
or do i have to add su -c "$JAVCMD etc.etc."
<bonniot>i gues it depends on your system
you can see that with ps -ux
let's try this again...20:09
i'm in the room...20:10
and nothing's working
i mean...
nothing's logging
<bonniot>you could check what files are opened, with lsof20:11
too many files
let's |grep it
<bonniot>yes :-)
you couldgrep for .html
wait a moment
oh my!
the logs are generating in my home directory!
* ChanServ leaves20:14
<bonniot>isn't that the current dir?
it is........
* ChanServ joins
<kdeuser>i thought it generates the logs from where i run it20:15
<bonniot>i suspected something like this. that's why i was asking if you cd to /var/www...
<kdeuser>so i should add a cd /var/www/html/logs/ to my service script?
<bonniot>you mean where the jar is?
<bonniot>i'm not sure if it's even possible to find out
but programs usually write in the current dir, not in /usr/bin, no?20:16
<kdeuser>yeah...nothing writes to /bin/
oh man!
i wonder if all my logs are there20:17
thanks alot
<bonniot>no pb
what channel are you logging?
<kdeuser>my own IRCD20:20
<bonniot>so you want to log several channels?20:21
<kdeuser>that would be nice...
i was thinking about just starting the program a half dozen times
<bonniot>yes, i think you'd have to do that at the moment
<kdeuser>okie doke
so why the wierd name?
<bonniot>if you feel like working on it, you're welcome :-)
ah, i still did not write that part of the web page...
it seemed like you didn't want to tell
* gamsl joins
<bonniot>it's from a story/legend
<gamsl>hi daniel , hi arjan20:23
it's been a long time!
<gamsl>it's been a long time
how are you all doing?
<bonniot>kdeuser: Sualtam is the name of a being which has a complete memory of the world in that story
<arjanb>hello martin20:24
<bonniot>gamsl: doing well
<bonniot>gamsl: and you? how was morroco?
<kdeuser>i did see that Sualtam was taken from a story...
<gamsl>very nice!! i love it :-)20:25
<kdeuser>unfortunately...i couldn't tell which story cuz it was in another language...
<bonniot>in french?
<bonniot>you found it on google?
<bonniot>gamsl: what did you do there?
<kdeuser>i don't know my foreign languages too well...
<gamsl>i'm doing fine now home again also ... university is mostly about programming this term
oh i just stayed in thagazout relaxing on the beach .. went to marrakesh a few days and to essaouira ... but most of the time, well, the beach :-)20:26
<bonniot>the water is quite cold there, no?20:27
<gamsl>we were living with a maroccan family (i was there together with friends ..)
about 18 centigrade i should guess
it's ok -)20:28
<bonniot>and you had such a long uni break?
<kdeuser>if i put -genIDX it'll generate an index every day wouldn't it?20:29
<gamsl>i had 4 weeks holidays and took the first week of the term also :-) it's only about registering for lectures and i did that online from morocco
<kdeuser>whoo hoo!20:30
it's all there!
THANKS a whole bunch Bonniot!!!!!!20:31
Thank you for writing the program and all...
<gamsl>daniel when i read messages from you like gamsl: some message ... is that some kind of private message?
<bonniot>no, everybody gets it20:32
<bonniot>it's just to make clear who i'm speaking to :-)
it does
but i can never seem to do that...
i keep forgetting people's names
takes to long to type it...
<bonniot>kdeuser: i'm not sure about -genIDX, i did not write that part
okie doke
<bonniot>my client has auto completion :-)
so i can write kd:
xchat does that?
<kdeuser>bo: test
<bonniot>i have xchat 2.0.7
<gamsl>oh i can do that also :-) i mean bo..
i have chatzilla ..20:34
<kdeuser>bo, test
<bonniot>ke: what did you send me?
kdeuser: what version do you have?
i didn't send you anything20:35
<bonniot>maybe it's an option
<kdeuser>i just checked to see what client you were running
<bonniot>'Received a CTCP VERSION from kdeuser'
right click on my name...
hit CTCP
then version
<bonniot>'Indent nicks' in Preferences?20:36
no: 'Automatic nick completion'
in 'Input box'
bonniot: test20:37
there we go
<gamsl>daniel i think i feel motivated to work for nice again!20:39
<bonniot>sounds great!20:40
we need you!
<gamsl>not that i was not motivated all the time but there was little time :-) it should be more this term ..
do you have plans about what to work on?20:41
<gamsl>what needs to be done? what about your enhancements to niceswing?
<kdeuser>i would like to work for nice...but i haven't the slightest clue how to program in that language!
<arjanb>gamsl: isaac started with a swt lib
<gamsl>i think i would really like to work on a (powerful in the long term) nice editor for eclipse
<bonniot>kdeuser: yes, a first step would be to read the tutorial ;-)
gamsl: that's very much needed20:42
<gamsl>i thought so :-)
but it's a big project .. it will take time ... since i also cannot only concentrate on that, i have to quite some coding for university too20:43
<bonniot>now we have sombody working on nicedoc
<gamsl>hi arjan! how far is the swt lib?
<bonniot>the interesting part is that nicedoc reuses the core of the compiler for learning about source programs
<gamsl>i read that .. i unil february i quite regularly read the irc logs ... but from then while i stayed in morocco i didn't read ..20:44
<bonniot>i think a powerful editor (refactoring,alternate views on the code) should take a similar course
<gamsl>i was thinking that there will be need for support in the compiler .... i was browsing through eclipse documentation a bit ... and i was wondering how much also of the java editor part one could reuse or at least take as a guideline20:46
how far is the nicedoc project?20:47
* kdeuser leaves20:52
* kdeuser joins20:54
<arjanb>not so far because he is learning nice at the same time20:55
<gamsl>ah ok
<gamsl>and what parts for interfacing with the compiler are already written?
<bonniot>JavaDocs -> nicedoc generated pages
the doc comments are there yet, but Frank's working on that
<kdeuser>who wrote nice? 20:57
<bonniot>gamsl: there is an api to get the abstract syntax20:58
<gamsl>that's definitely needed .. fine :-)
<bonniot>of course you can manipulate this representation of the code20:59
<bonniot>kdeuser: i started it, and did a good part. alter arjan joined in, and several other people are helping with tools, libraries, ...
<gamsl>following Javadocs i only come to a page which lists all packages and type hierarchy .. no method or type documentation
<bonniot>gamsl: well, the compiler internally uses objects to represent everythin in programs, so you have access to those
<gamsl>that's good ... the jdt (eclipse java development tooling) does it like that also ... i guess it will be very interesting to study how they do it!21:02
<bonniot>so formally there is no special need to write more API (apart from adding 'public' if it's not there). in practice probably we can discover some useful methods to add to the interface
gamsl: most of nice.swt is just retyping at the moment, and that is not part of the output yet21:03
and as i said, doc comments are worked on at the moment21:04
(this output is actually not from the latest version)21:05
<kdeuser>bye bye21:10
* kdeuser leaves
<gamsl>i see21:13
daniel what about your enhancement ideas for niceswing?21:14
is this urgent?21:17
<arjanb>i think getting the eclipse plugin moving again is more important atm21:22
<gamsl>i agree21:25
arjanb, do you know about daniel's idea?21:26
would it mean much work? i could use it maybe as a starting point to familiarize myself with nice again :-)
<arjanb>i don't know about that21:27
what about the current eclipse plugin? is alex working on it?
* bonniot leaves21:34
<arjanb>alex is too busy with his job and another project so nothing happened with the eclipse plugin for a half year
<gamsl>hmm ok21:35
* bonniot joins
<gamsl>do you all agree to write the plugin against the current 3.0.0 M8 milestone build .. and not stay compatible with the current stable 2.1.3 ?21:36
<bonniot>did you get my messages about niceswing?21:37
<gamsl>i think they changed quite something in the PDE part
<bonniot>seems they are not in the log
<bonniot><bonniot> ok, so basically the idea is to get rid of the abstract interfaces
<bonniot> the XxxHandlers
<bonniot> i think some of them are implemented by only one class, so that's easy
<bonniot> for those that are implemented by several, there needs to be several implementations of the addXxxListener(NiceXxxLisdtener) anyway
<bonniot> so we could instead have several addXxxListener(anonymous function)
<gamsl>i saw you just signing in again
<bonniot><bonniot> does that make sense?
<gamsl>give me some time to think :-)21:39
<bonniot>sure :-)21:41
<CIA-3>03bonniot * 10Nice/regtest/java/Makefile: Use the current nice compiler, not the system one.21:42
* gamsl leaves21:47
* gamsl joins21:53
i'm having connection problems with my wlan ...
hmm you didn't get my answers, right?21:54
they're not in the log
so again:
i was thinking about the AIs more as a guideline for me to enforce the implementation of the methods to be complete. I don't think that they are really necessary ..21:55
i cannot think of any striking arguments for AIs in niceswing at the moment ... 22:02
<bonniot>if i remember, the idea of the AIs was also to share implementation, no?22:07
<gamsl>but, i don't understand "several addXXListener(anonymous function)" ..especially the anonymous function22:10
mhm you're right ... seems i have to look closer into niceswing again ... :-)
mhm you're right ... seems i have to look closer into niceswing again ... :-)
but, i don't understand "several addXXListener(anonymous function)" ..especially the anonymous function
arrr my wlan is so unstable ..22:11
<bonniot>i think the situation is that several classes can have an ActionListener, but there is no interface that represents this capability22:14
so one needs to have several methods in niceswing taking an anonymous function, wrapping it in a Java listener, and adding it to the object22:15
<gamsl>yes that's how it works ...22:18
<bonniot>i think getting rid of the AIs would simplify the design, greatly speed up compilation time, and allow repackaging if desired...22:20
<gamsl>you mean you would have more like this:22:21
public <ChangeListenerHandler H> void addChangeListener
(H this,
ChangeEvent -> void stateChanged
(new NiceChangeListener
(stateChanged: stateChanged
these are good points!
<bonniot>more like this???
what's the difference?22:22
<gamsl>not parameterize it with Ais but overload them ... or did i get you wrong? (it looks like ...)h22:24
<bonniot>ChangeListenerHandler is not an AI?
<gamsl>yes it is .. but I thought of getting rid of the current version using an Ai but replacing it by overloaded methods .. (for all the types that implement the now used AI)22:25
is that what you mean?22:26
<bonniot>yes it is
but then i don't understand why you said " you would have more like this"22:27
this looks to me like what we want to get rid of
<gamsl>yes that sounds weird i see now :-)22:28
but i meant it like i said in the end ...
<bonniot>ok ;-)22:29
<gamsl>and for the addNiceXXXListener methods the same procedure?22:30
<bonniot>that should become useless, no?
just call the original addXxxListener22:31
<gamsl>so it's basically just overloading ...22:33
we loose the coverage test but anyway ....
that should work
<bonniot>yes, but nothing guaranteed that you put all the abstract interface implementations before, so it's not much different, no?22:35
<gamsl>we lose the whole coverage test but we don't need that really do we?
<bonniot>so i'm not sure we lose anything22:36
<gamsl>that's true that was a weak point .-) once you have them nice leads you till its complete .. i remember especially liking this :-)
<bonniot>btw, you know that nicec now checks for implementation of java methods, which it was partially lacking before?22:37
<gamsl>you have an example?
<bonniot>implementing any Java interface22:38
now it is checked that all methods are implemented
<gamsl>ah ok22:40
i saw that a lot of new features entered the language
<bonniot>yep :-)22:41
<gamsl>what would u say are the most notable?22:42
and what is currently being worked on?
<bonniot>mpst notable: depends since when ;-)22:48
<gamsl>hehe well ... since niceswing :-) i read about syntax changes, custom constructors which are especially nice! ...
enums ..22:49
what about information hiding?
<bonniot>custom constructors, method and field overriding
that still needs to be implemented
recently i've been working on type inference
<gamsl>mhm i saw that in the commit logs ...22:50
i took a volunteer compiler design lecture last year ... besides my lectures without grading .. very very interesting :-)
<bonniot>arjanb: i can't bootstrap the generic arrays:22:51
JAVA="java" /home/daniel/projets/GenArrays/bin/nicec --exclude-runtime -d "/home/daniel/projets/GenArrays/classes" --sourcepath="/home/daniel/projets/GenArrays/stdlib:/home/daniel/projets/GenArrays/src" -R bossa.syntax
Exception in thread "main" java.lang.VerifyError: (class: nice/lang/fun, method: filter signature: (Lnice/lang/GenericArray;Lgnu/mapping/Procedure;)Ljava/util/Collection;) Expecting to find array on stack
at nice.lang.dispatch.notNull(dispatch.nice)
at nice.tools.compiler.console.fun.compile(fun.nice:20)
at nice.tools.compiler.console.dispatch.compile(dispatch.nice)
at nice.tools.compiler.console.fun.main(fun.nice:1)
seems like a bug in the generic arrays implementation
<gamsl>daniel, i will see to change niceswing in the next days! i think this should not take too much time ...23:02
do you know about the state of alex's eclipse plugin .... is this work to build on ?23:03
<bonniot>i don't know it in details23:07
it's in CVS, so you should see for yourself
there is the question of eclipse versions23:08
the current plugin is made for 2.1, but i made sure it also works with 3.0 M7
is there any problem with keeping this route?
3.0 M7 and later have a compatibility mode, which means that the older API is supported23:09
<gamsl>i was asking before but i guess i was offline then ... i think we should work with the current 3.0.0M8 stream stable build and not try to stay compatible to the 2.1.3 stable API
ah ok23:10
what do you think of keeping close to the design of the java editor? since at least all of it's core features are desirable for nice too ... 23:13
i should think this is currently the most powerful one and it could be used as a guideline how to do it right ..23:14
<bonniot>the pb is that many people still use 2.1
i have no idea how much the api has changed23:15
but why is it a problem, since you can develop for 2.1 and use it on 3.0?
<gamsl>i also don't know (yet) ... but this project will take time anyway ... and the few testers that there hopefully would be could manage working with M8 .. why code against an old API? it's only a matter of time until 3.0 is standard ... and the plugin will take time!!!23:16
<bonniot>there is already a working plugin, so it can be improved gradually23:17
if you want a complete rewrite, I agree targetting 3.0 makes more sense
if you improve the current plugin, i don't see a reason to modify it to break compatibility just to say that it's using the newer api23:18
<gamsl>i don't know the sources yet .... but as i was thinking to keep as close as possible to the design of the jdt (which is also only a first idea) .. i was thinking maybe to use parts but to integrate them in a new one ... but as i said, i haven't looked into alex's sources yet23:19
i was just thinking keeping close to jdt is maybe good because nice is in some way also close to java ... maybe we can use something23:21
* CIA-3 leaves23:26
* CIA-3 joins23:27
<gamsl>i just read that The 3.0 release is targeted for June 200423:33
<bonniot>about the model:23:47
i'm not familiar with the jdt, but i suppose it's a valid model
a sure thing is that there should be compatibility where it makes sense
for instance reusing the debugger23:48
for working on the source, there is an alternative model which would be very interesting to try:
instead of the file based view (or in addition to it), it would put the focus on the duality class/multi-method23:49
you could see the class hierarchy (as a tree)
when selecting a class, see the methods that have a parameter that can be of this class, and the relevant implementation23:50
when selecting a method, see all its implementations
basically the class view is the traditional one (as in Java)
the method view is also useful, when you focus on a particular operation23:51
<gamsl>sorry i was away00:02
<bonniot>arjanb: could you reproduce the pb?00:05
<gamsl>sounds good! we will definitely have to think about new ways of navigating in the source code, and also maybe about different ways of outlining than java00:08
i didn't exactly mean to look like jdt and offer the same navigation and stuff ... but to design class interaction and interaction with the framework close to the way they do it in jdt ... 00:09
so you mean the implementation should be close?
<gamsl>since it is the most advanced programming language editor for eclipse .. i think we should stick to the way they do ... this does not mean provide the same functionality ..00:11
since we have to deal with similar tasks .. we could learn from their structure00:12
<bonniot>ok, it's always a good idea to reuse previous work and ideas ;-)
ok, i'll have to leave for tonight
i'm very glad to see you back, martin00:13
<gamsl>ok good night .. i will come here tomorrow ...
glad also :-)
<bonniot>arjan, tell me if you have news about arrays
good night
<gamsl>now i have time again, that's what i was waiting for :-)
* bonniot leaves
<gamsl>good night arjan00:18
* gamsl leaves00:19
* arjanb leaves01:16

Generated by Sualtam