Squeak like environment for Ruby

[Logan Capaldo]

Lately I've been playing around with Squeak (http://www.squeak.org/),
and I was wondering wouldn't be cool if ruby could have a similiar
environment. Would anyone else be interested in something like this?

 

[Caio Tiago Oliveira]

Logan Capaldo, 8/2/2005 00:45:

Lately I've been playing around with Squeak (http://www.squeak.org/),
and I was wondering wouldn't be cool if ruby could have a similiar
environment. Would anyone else be interested in something like this?

Yes. But Ruby is yet a lot imature for this.
At least we can try to make something like the worksspace (that's cool).

 

[Alexander Kellett]

Logan Capaldo, 8/2/2005 00:45:

Yes. But Ruby is yet a lot imature for this.
At least we can try to make something like the worksspace (that's
cool).

immature? in what sense?
thats troll talk. you should take the time to rephrase

Alex

 

[gabriele renzi]

Alexander Kellett ha scritto:

immature? in what sense?
thats troll talk. you should take the time to rephrase

maybe he meant that we do not have, yet, a rubyish standard gui[1] or
audio engine, and that our engine is much slower than Squeak's.
Maybe it could be more interesting to hack a ruby interface *for* squeak
instead of reinventing the wheel


[1]
well, I think TkRuby should be that, but people seem to dislike it no
matter how many cool things hidetoshi nagai put in it (html support,
graph support, theme engines..). Anyway we can hope that ruby/wise get
some more work or that MorphR appear

 

[Alexander Kellett]

maybe he meant that we do not have, yet, a rubyish standard gui[1] or
audio engine, and that our engine is much slower than Squeak's.
Maybe it could be more interesting to hack a ruby interface *for*
squeak instead of reinventing the wheel

i can't stand squeak personally
though it has some useful tools.

i'll try and do a rpa/gem for qtruby4
whenever its out. maybe then we'll have
an easy to install cross platform gui
then we can actually make a good class
browser

[1]
well, I think TkRuby should be that, but people seem to dislike it no
matter how many cool things hidetoshi nagai put in it (html support,
graph support, theme engines..). Anyway we can hope that ruby/wise get
some more work or that MorphR appear

does it look pretty now? if so. screenies?

Alex

 

[Robert Feldt]

maybe he meant that we do not have, yet, a rubyish standard gui[1] or
audio engine, and that our engine is much slower than Squeak's.
Maybe it could be more interesting to hack a ruby interface *for*
squeak instead of reinventing the wheel

i can't stand squeak personally
though it has some useful tools.

i'll try and do a rpa/gem for qtruby4
whenever its out. maybe then we'll have
an easy to install cross platform gui
then we can actually make a good class
browser

With Trolltech releasing qt4 for windows it would be interesting to
know the state of qtruby4 for windows? And Mac OS? Is it truly
cross-platform?

On the squeak issue: Depends on what you mean with squeak, there are
very many aspects to it. Related to the VM issues I've resurrected my
old writing from my last employer (Thanks Ryan!) and have a few
squeak-related rubyvm material up on this page

http://www.pronovomundo.com/projects/ruby/rubyvm/

although I think ruby2c is where similar kind of action has happening today.

On MorphR: I'm actually trying to find a student for finishing that
off this spring. We'll see what happens. I still think having a GUI
fully implemented in and thus controllable from/with pure Ruby would
be very useful.

Best,

Robert

 

[Alexander Kellett]

With Trolltech releasing qt4 for windows it would be interesting to
know the state of qtruby4 for windows? And Mac OS? Is it truly
cross-platform?

yup. works fine on mac os. however i've not yet
tried it on windows. i'll try and hijack a windows box
and get it compiled with mingw32. it should just work out
of the box. however neither me nor richard have begun
work on a qt4 port of qtruby as i'm busy with various other
projects and qt4 isn't actually used in kde at all upto
now. hopefully in the coming weeks i'll take a look.
qt4 and thusly windows qt4 will be second quarter this
year so we should have quite a bit of time

On the squeak issue: Depends on what you mean with squeak, there are
very many aspects to it. Related to the VM issues I've resurrected my
old writing from my last employer (Thanks Ryan!) and have a few
squeak-related rubyvm material up on this page

maybe you'd be interested in:
www.lypanov.net/rubydium7.pdf

btw, i tried contacting you a while back wrt ruth
and was wondering if the spam filters caught the
email? summary - would you like patches?

Alex

 

[Robert Feldt]

On the squeak issue: Depends on what you mean with squeak, there are

maybe you'd be interested in:
www.lypanov.net/rubydium7.pdf

I am, but the little time I have for ruby vm/interpreters has gone
into compilers lately...

I'm especially interested in your use of pyggy though. Does this mean
you have a GLR grammar for Ruby lying around? I would like to try that
with rockit so if it is available / open-sourced please pass on; that
could save time. I have the starts of GLR grammars for Ruby but
nothing complete. Maybe we can join forces?

btw, i tried contacting you a while back wrt ruth
and was wondering if the spam filters caught the
email? summary - would you like patches?

Sure, although I tend to think that ParseTree would be a more
maintained/modern alternative?

Best,

Robert

 

[Richard Dale]

yup. works fine on mac os. however i've not yet
tried it on windows. i'll try and hijack a windows box
and get it compiled with mingw32. it should just work out
of the box. however neither me nor richard have begun
work on a qt4 port of qtruby as i'm busy with various other
projects and qt4 isn't actually used in kde at all upto
now. hopefully in the coming weeks i'll take a look.
qt4 and thusly windows qt4 will be second quarter this
year so we should have quite a bit of time

All I've done is download Qt 4.0 and have a look at how slots/signals are
implemented. Instead of QUObjects, it uses arrays of 'void *'s to pass the
arguments to a slot. That actually seemed more similar to how the Smoke
library expects its args, than the old way. Just that some things were
pointers to pointers, instead of just pointers.

But it would be really good to do some sort of prototype as soon as possible
with Qt 4, to get any implementation uncertainties out of the way.

maybe you'd be interested in:
www.lypanov.net/rubydium7.pdf

btw, i tried contacting you a while back wrt ruth
and was wondering if the spam filters caught the
email? summary - would you like patches?

I'd love to do Squeak bindings for Qt/KDE, so we could integrate Croquet
with KDE. I did do some KDE Objective-C bindings, and solved the problem of
how to derive Smalltalk style method names from a C++ api - Squeak should
be very similar. Just need to change the Smoke library to keep tables of
arg names, like it keeps arg types at present.

-- Richard

 

[Phil Tomson]

I am, but the little time I have for ruby vm/interpreters has gone
into compilers lately...

Of course, this begs the question:
So how is your Ruby compiler work going?

Phil

 

[Alexander Kellett]

I am, but the little time I have for ruby vm/interpreters has gone
into compilers lately...

hehe. ruby compilers? how goes with that?

I'm especially interested in your use of pyggy though. Does this mean
you have a GLR grammar for Ruby lying around? I would like to try that
with rockit so if it is available / open-sourced please pass on; that
could save time. I have the starts of GLR grammars for Ruby but
nothing complete. Maybe we can join forces?

its also just a beginning. just fiddling around with
parsing of heredocs, #{}, %w{}, etc. really a very tiny
syntax but it was fun. what sort of parser does rockit
have? you have a working 1.8.x release now i guess
will that be released anytime? i seem to recall that
you went part c part ruby, which means its out for
use with rubydium unfortunately. as soon as i'm finished
with the current rubydium optimisations i'll do some more
work on the glr ruby grammar.

Sure, although I tend to think that ParseTree would be a more
maintained/modern alternative?

ruth does everything i've needed up to now
just a few minor patches to the ast and some random
stuff like multiassigns aren't supported, but i've
got plenty more to implement in rubydium before that

was going to switch to a pyggy based parsetree
whenever i find the time, but that'll be several
months ago at a guess.

Alex

 

[Robert Feldt]

I am, but the little time I have for ruby vm/interpreters has gone

Of course, this begs the question:
So how is your Ruby compiler work going?

Still work left to do...

/R

 

[Robert Feldt]

I'm especially interested in your use of pyggy though. Does this mean

its also just a beginning. just fiddling around with
parsing of heredocs, #{}, %w{}, etc. really a very tiny
syntax but it was fun. what sort of parser does rockit
have? you have a working 1.8.x release now i guess
will that be released anytime?

Yes, it is also GLR with embedded regexps. It works with 1.8 and a
release is long overdue. I really want to get a full Ruby grammar in
there...

i seem to recall that
you went part c part ruby, which means its out for
use with rubydium unfortunately. as soon as i'm finished
with the current rubydium optimisations i'll do some more
work on the glr ruby grammar.

Ok, we could exchange grammar sketches then; pls keep contact when you
move on that front.

ruth does everything i've needed up to now
just a few minor patches to the ast and some random
stuff like multiassigns aren't supported, but i've
got plenty more to implement in rubydium before that

Ok, please pass them along.

was going to switch to a pyggy based parsetree
whenever i find the time, but that'll be several
months ago at a guess.

I'm not sure I understand "pyggy-based parsetree" but it sounds fun...

/R

 

[Alexander Kellett]

Ok, we could exchange grammar sketches then; pls keep contact when you
move on that front.

shall do so

Ok, please pass them along.

new tarball -
http://www.lypanov.net/ruth.tgz
i'll do a patch sometime

I'm not sure I understand "pyggy-based parsetree" but it sounds fun...

like ruby2c's sexp style parsetree subproject.
but then generated by pyggy.

Alex

 

[Caio Tiago Oliveira]

Alexander Kellett, 8/2/2005 06:03:

immature? in what sense?
thats troll talk. you should take the time to rephrase

IRB, which is the thing more likely to squeak workspace doesn't work
very well on windows. The IDEs for Ruby I tried are almost only
shortcuts for run the program and syntax colouring. There is a lack of
great features. The Refactoring of Freeride is still in development
(almost anything worked for me) and I couldn't find it in my version of
ArachnoRuby.

Yes, Ruby is very slow.

Cools projects like yarv and ruby2c yet are in development.

Ruby is cool and I will love when the AspectOrientedRuby proposal
becomes part of the language.

But it has a lack of good suport.

 

[Alexander Kellett]

IRB, which is the thing more likely to squeak workspace doesn't work
very well on windows. The IDEs for Ruby I tried are almost only
shortcuts for run the program and syntax colouring. There is a lack of
great features. The Refactoring of Freeride is still in development
(almost anything worked for me) and I couldn't find it in my version
of ArachnoRuby.

have you tried the mingw32 1.8.2 windows install that eban
uploaded to raa today? i found cygwin very unstable in the
past, maybe other versions work much better? if i remember
correctly there are some bugs in completion on linux also,
but not as severe as the ones you see. not many people use
windows here unfortunately so until the windows users help
out nothing much will change. (the people that do use windows
are severely time limited and there are very few of them )

ruvi also supports the rrb refactoring, and i also noticed
that while it works well on small examples it falls down
majorly on more complex examples. if you could file bug
reports with testcases i'm sure that someone out there
would be willing to fix them

Yes, Ruby is very slow.

wouldn't say its very slow
but yeah. it could be faster.

Cools projects like yarv and ruby2c yet are in development.

and my own. 'rubydium' - http://www.lypanov.net/rubydium7.pdf

Ruby is cool and I will love when the AspectOrientedRuby proposal
becomes part of the language.

i've not seen this. never really found any real uses for aop
maybe if i had more time to refactor rather than struggling
with ugly code i'd be happier

But it has a lack of good suport.

you mean these above problems or more?

Alex

 

[Caio Tiago Oliveira]

Alexander Kellett, 8/2/2005 12:59:

On Feb 8, 2005, at 4:50 PM, Caio Tiago Oliveira wrote:
have you tried the mingw32 1.8.2 windows install that eban
uploaded to raa today? i found cygwin very unstable in the
past, maybe other versions work much better? if i remember
correctly there are some bugs in completion on linux also,
but not as severe as the ones you see. not many people use
windows here unfortunately so until the windows users help
out nothing much will change. (the people that do use windows
are severely time limited and there are very few of them )

mingw32? I use the one-click installer the rc14.

ruvi also supports the rrb refactoring, and i also noticed
that while it works well on small examples it falls down
majorly on more complex examples. if you could file bug
reports with testcases i'm sure that someone out there
would be willing to fix them

I'll try it.

and my own. 'rubydium' - http://www.lypanov.net/rubydium7.pdf

"Ruby, but Fast", sounds good.

you mean these above problems or more?

These above and more too.

 

[Alexander Kellett]

mingw32? I use the one-click installer the rc14.

try the mingw32 download then. look at raa.

These above and more too.

umm.. ok...

Alex

 

[llothar]

Can we please add this to the FAQ.

It's a little bit boring to explain every two weeks again why an image
based language is completely different from a file based language. Same
for the class browser question.

 

[llothar]

Can we please add this to the FAQ.

It's a little bit boring to explain every two weeks again why an image
based language is completely different from a file based language. Same
for the class browser question.