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

Using timezone: Central European Time
* arjanb leaves02:35
* zzorn joins08:08
* arjanb joins12:45
* CIA-2 leaves13:22
* CIA-2 joins13:40
* CIA-2 leaves13:44
* CIA-2 joins14:16
* CIA-2 leaves14:24
* CIA-2 joins15:02
* CIA-2 leaves15:09
* Atomixx joins15:58
<Atomixx>hi there16:01
<bonniot>how are you doing with Nice?16:04
<Atomixx>it is nice ;-)16:07
<bonniot>what have you been doing with it?
<Atomixx>i want to write an ftpclient which can be extended to use other plugins16:08
for tftp for example
bonniot, i just tried to compile nice.swt16:20
but it gives me a compiler error
<arjanb>which error?16:27
* bonniot leaves
* bonniot joins16:29
got disconnect for a sec
<arjanb>that's a bug in the compiler, what version do you use?16:30
<bonniot>hu, looks like I missed something...16:31
<arjanb>16:28 <Atomixx> http://paste.phpfi.com/53120
<Atomixx>0.9.10 16:32
<bonniot>i tried compiling nice.swt a few days ago and it worked
i'll try again with 0.9.10
compiled with maven?16:33
command: 'maven nice:compile'16:34
<bonniot>it's working here
did maven download nice-SNAPSHOT?
<Atomixx>yes it did
and it is there in .maven/nice/jars/...
<bonniot>and you are using nice.swt from cvs?
<bonniot>ls -l ~/.maven/repository/nice/jars/nice-SNAPSHOT.jar
that's where it is for me
<Atomixx>yes i meant that
which JDK?
<Atomixx>atm 1.5
but i could switch to 1.4 for the compilation
<bonniot>we'll see, trying with 1.516:37
you could try, always interesting
1.5 works too16:38
do you have stuff in target/ ?
(unix or windows, so I know which language to use ;-)
and which direction for / or \ ;-)16:39
* Atomixx is running gentoo linux
should have known ;-)
can you show me a `ls -l ~/.maven/repository/nice/jars/nice-SNAPSHOT.jar` ?
and ls -l target16:40
<Atomixx>-rw-r--r-- 1 nice/jars/nice-SNAPSHOT.jar16:41
<bonniot>i'm interested in the date and the size
<Atomixx>ah ok
-rw-r--r-- 1 1452896 Mar 3 16:37 nice/jars/nice-SNAPSHOT.jar
<bonniot>ok, same size as me16:42
<Atomixx>what did you mean with 'target'?
<bonniot>let's divide and conquer. could you make me a tgz of your tree?16:43
`ls -l target`
`tree target` is even better ;-)
<Atomixx>the nice.swt tree ?
<Atomixx>ok. just a moment
<bonniot>if I can reproduce, then I can track it down
otherwise it must be a subtle difference in your setup
cannot make any sense of it at the moment16:45
<Atomixx>how can i send it to you?16:47
<bonniot>Daniel.Bonniot at inria.fr
<Atomixx>should arrive in a moment ;-)16:50
<bonniot>ok, i get the same thing as you!16:52
let's go for recursive diff...16:58
<Atomixx>it also appears with 1.416:59
<bonniot>i see something strange in nice.swt.all17:02
you have that package import itself
this package is autogenerated
<Atomixx>how could that happen?17:03
i fetched from CVS and then ran maven
<bonniot>because of the autogeneration17:05
actually I have it too, only in a different order
i'll check if this is it17:06
this is the only source difference
you bet it is!17:07
<Atomixx>should i modifie this file or does it get overwritten again?17:08
<bonniot>it will17:09
i sent you a copy of maven.xml that will not overwrite, and a good version of all.nice
all.nice goes into src/nice/swt/all/17:10
<Atomixx>yes, i found this too
<bonniot>this sounds like a very, very weird bug
<Atomixx>i think so too17:11
<bonniot>but since i'm going to overhaul the visibility implementation aynway, this might very well disapear by magic ;-)
<Atomixx>this would be nice ;-)17:12
* CIA-2 joins17:13
<bonniot>it would still be useful to know how to reproduce this in a small testcase
this is the second "random" bug we get in a few days17:14
does that mean all the deterministic ones have been fixed already? ;-))
<Atomixx>i'm not sure17:15
<bonniot>just joking
<Atomixx>i don't get
it isn't working too
<bonniot>same error?
<Atomixx>same error as before17:16
<bonniot>maybe i sent you the wrong maven.xml
can you check its content?
<Atomixx>yes, how?17:17
<bonniot>the inside of the <preGoal> </preGoal> tag must be commented out
<!-- ...... --->
if it wasn't then you'll need to recopy all.nice from my mail
<Atomixx>no it isn't but i'll do that
i guess the order of iteration on the dirs must depend on the inode number or something like that. highly random17:19
<Atomixx>[nicec] Compilation successful.
when I got disconnected, i was asking what tftp is
and is the GUI for your ftp client too?
<Atomixx>yes it will be17:21
tftp is trivial file transfer
it is without login and so on
<bonniot>i thought it was a subpart of ftp
so you wouldn't need a plugin for it
<Atomixx>haven't looked at it closer
i don't know
i also thought of scp and so on17:22
<bonniot>don't really know either, it's just a guess
yeah, that's definitely needed
so the goal is to have interactive remote file copies?
<Atomixx>in the commons-net.jar are different packages for ftp and tftp
yes this should be the goal
and later it should be multithreaded17:23
and several open connections to different servers
<bonniot>is there no such tool that you like? or it's more about learning the stuff?17:24
<Atomixx>i'll see how far i will get
<bonniot>(i don't know much about ftp guis)
<Atomixx>and i don't know if there is some tool like that
but maybe you know leechftp
i like its frontend
but it is windows only17:25
<bonniot>i use mostly /usr/bin/ftp (scripted from the Nice distrib Makefile ;-)
<Atomixx>there is some similar python program, which didn't work for me
i also have the standard /usr/bin/ftp17:26
but it is only singlethreaded
<bonniot>so you'll need to implement the multithreading yourself?
<Atomixx>it seems so17:27
<bonniot>then that would make sense to try to export that as a library, independently of the GUI
<bonniot>you could even make the GUI a widget, so it can also be reused in other projects17:28
that would be very cool
<Atomixx>i have an approach that truncates the GUI from the transfer things
<Atomixx>it is very simple
<bonniot>if you do it in nice, I think these could be perfect candidates for packages.sf.net17:29
<Atomixx>yes indeed
but i didn't work with widgets before
<bonniot>cool, you will be able to beta-test the infrastructure then ;-)
<Atomixx>so i have no real idea of them
<bonniot>i guess the multithreading part is the most likely to be reusable17:30
he, when it's all in place it should be great: continuous integration, automatic webpage generation, ...17:31
<Atomixx>yes indeed17:33
<bonniot>testing will be fun, because you would need to write a simple server too!17:35
<Atomixx>oh no ;-)17:36
here you can see the first ideas: http://www.stud.uni-karlsruhe.de/~utaue/TransferClient.pdf17:37
my first program in nice: http://www.stud.uni-karlsruhe.de/~utaue/programme/17:38
i rewrote the scissors example to be able to play over internet ;-) (small use of sockets)17:39
to sound pedantic, i would use a boolean instead of int,int to indicate who won17:41
that seems less redundant
<bonniot>(i thought the prg was long, until I saw there were sockets ;-)
<Atomixx>but it could be used for error messages ;-)
<bonniot>if you want to put programs on the web, you can use 'highlight'17:42
it can do .nice -> .html
with syntax highlighting and all
stylesheets, even ;-)
<bonniot>i'll be gone for a while...17:45
<Atomixx>ok bye
i wont be here for a longer time maybe
it is because i have only 56k at home17:46
<Atomixx>and i wont go to university if i have no courses ;-)
<bonniot>don't hesitate to email me or nice-info if you have questions
that's more connection-time efficient ;-)17:47
<Atomixx>i'll spam your mailox *g*
<bonniot>don't worry, it's big
<Atomixx>btw i was the one with the compiler crash bug with the inet4addr
<bonniot>but if the question is general, then use nice-info
<Atomixx>i'll do, thanks17:48
<bonniot>what's that about?
<Atomixx>what is what about?
nice compiler?
<bonniot>is this an open bug?
<Atomixx>no, it was fixed a few hours after i had reported it17:49
<bonniot>sorry for the delay ;-)
<Atomixx>it was because it hadn't a public constructor
<bonniot>ah, that comes back17:50
<bonniot>you have no sf account?
<Atomixx>yes i have
<bonniot>ok, I'm flying. see you soon
<Atomixx>ok bye
* furko joins18:03
* furko leaves
* Atomixx leaves18:48
* Bluelive joins20:00
* cm_ joins21:13
<cm_>Any mind blowing news/progress? ;-)22:07
<bonniot>since when? ;-)
0.9.10 released
implementation of visibility started
a potential third developer on the core compiler ;-)22:08
<cm_>Sounds great
<bonniot>oh, by the way, frank, the main dev on nicedoc, wanted to contact you I think22:09
<cm_>Is frank the guy from australia?22:10
could you write to him at fbarber at users.sourceforge.net
you spoke about nicedoc with him, right?
<bonniot>then that's it
what about you?22:11
how are you doing?
<cm_>Fine, any you?22:12
<cm_>Any new projects using nice?22:13
<bonniot>i learned a bit more about how Artem is using it for server pages, in a commercial project22:14
he wrote about it a bit on the list some time ago
<cm_>Is the compiler more embedded-friendly now?22:16
<bonniot>than when?
it's embeded in ant, eclipse plugin, maven plugin, nicetee, and the compiler console interface ;-)22:17
oh, and nicedoc
what would you want to do with it?
<cm_>I reckon a discussion on irc a couple of months ago but I'm not initiated.22:18
Yes, the eclipse plugin, any progress there.
<bonniot>i didn't hear from martin recently
i think he was going on a trip for some time22:19
<cm_>Ok, the mailing list have been rather silent.
<cm_>Any thought about the a macro type of system in nice?22:20
<bonniot>thoughts, yes, but it's not on the top of the stack yet...22:21
<cm_>A system like the dotnot-language (forgot the name again) would both be fun and powerful22:22
<cm_>heh, typo
<bonniot>dunno what you refer to
<cm_>The language arjan talk about some time ago22:23
<bonniot>arjanb ?
<bonniot>no, i'm asking if he's here ;-)
yeah, i read a bit about it22:25
* Bluelive leaves
<cm_>I like the power of lisp but I'm a type-safety kind of guy. 22:26
<bonniot>it defenitively seems worth studying
<cm_>A macro system could perhaps fill the gap
<bonniot>i believe so
do you have examples when you miss macros?22:27
<cm_>One is when you want to create a domain language.22:29
I'm especially interested in compile-time macros.
<bonniot>that's the ones i think are worth considering22:30
<cm_>Wonder when the industry is going to realise that plain hardcore c++/java/vb isn't the only way to go.22:31
<bonniot>can you give an example of a macro you would want to write?
* arjanb_ joins
<cm_>Had one idea some time ago. Consider a component system, eg COM. With compile-time macros/extensions would it be possible to generate proxy/stub code without the need of a specific tool.22:32
* zzorn_afk leaves22:34
<bonniot>in that case there is a file describing the interface?
<bonniot>only, if you do it with macros, you would have to regenerate the stubs at each compilation, right?22:36
that could be wasteful
<cm_>yes perhaps22:37
<bonniot>makes it interesting to decide whether a package is up to date...
* arjanb leaves
<cm_>Some things are not dependent on external resources. Consider the Visual c++ extension coclass. Such a construct could perhaps be impl. with a macrossystem.22:38
(coclass is used to create COM-components with c++-syntax)
Code transformation, often used in lisp, is also very powerful22:40
Btw, is generics in nice as powerful as in c++. The boost project is doing some impressive stuff with templates in c++.22:45
<bonniot>c++ templates are essentially a macro system22:48
so they can do more than generics22:49
<cm_>And turing complete :-)
<bonniot>but the way the do generics is kind of a hack
this means that you'll often get errors too late
at instantiation time instead of compile time
<cm_>And often very hard to interpret
so I don't think this is well designed
generics is somethings that is well understood in type systems, and that can be done completely safely, without syntactic hacks22:51
macros are a different issue, that should be handled separately. typically one step higher
<cm_>Ok. I have boost-python to generate python-bindings to c++. It is essential a domain language for creating bindings. It is really powerful but the errors are sometimes a bit tedious to decipher.22:55
Perhaps could something similar be impl. with a macro-system.22:56
<bonniot>arjanb: would you have time to try to reproduce atomix's bug?23:01
<arjanb>i don't have a similar setup23:12
<bonniot>i don't think it matters23:13
<cm_>bonniot: Have you read the "Template Meta-programming for Haskell" paper?23:20
<bonniot>yes. sounded like a very good approach23:21
what did you think?
<cm_>Very nice indeed.
Finally a printf :-) 23:22
<cm_>There is a printf function in ocaml. A *crazy* hack.
It manipiulates with internal things23:23
* cm_ leaves00:19
<CIA-2>03xoltar * 10Nice/ (3 files in 3 dirs): Some useful odds and ends00:50

Generated by Sualtam