Hi..About Gsoc 2009 project ideas

Hi,

I’m an Computer Science Undergraduate who is currently undergoing
training
and engaged n the field of Open Source SOA products n development.
I’ve
recently done project work in the area of WS Security (WS [Web Services]
Policy derivation from SOAP messages n some OSGI component[WSO2 Carbon
platform] implementatons around that…) and have some knowlege on WS
security implementations.

                  Also i've recently been working on deploying Ruby

Services on Apache Axis2 Engine using JRuby frame work…I Would really
like
to participate in a Ruby project for GSoc 2009 n would like to have some
project ideas to be heard from Ruby people…One of the things i’m really
interested in doing is to develop a Ruby Annotation framework
(DSL/parser)for a given spec as ruby currently doesn’t support this
behavior
very well. For my opinion this will be really important for WebServices
and
related stuff as i my self has come across difficulties in not being
able
to use a proper annotation framework as available in Java or C++. I’m
still
pretty much new to ruby n it’s aspects and hence doesn’t know how
feasible
this is or the scope of this n really appreciate if you all could
enlighten
me on this . Also i’m open to other project ideas as well and again
appreciate your kind reponse. Thnx in advance…

Thnx,
Udayanga Shaminda Wickramasinghe

Hi,
Just thought of asking about how the things are going on with 2009 GSoc
project proposals for Ruby .However i’m not too sure whether this is the
exact mail list ,where new ideas should be discussed regarding GSoc 09.
If
someone has an idea please provide me with your kind feed back. I would
really appreciate your thoughts regarding my previous mail as well , if
possible. Thnx in advance…

Udayanga Wickramsinghe

Just thought of asking about how the things are going on with 2009 GSoc
project proposals for Ruby .However i’m not too sure whether this is the
exact mail list ,where new ideas should be discussed regarding GSoc 09.
If
someone has an idea please provide me with your kind feed back. I would
really appreciate your thoughts regarding my previous mail as well , if
possible. Thnx in advance…

Utah Ruby U. group is good–I think that’s where Pat Eyler hangs out,
who has some connection or other–see the recent post about it.
-=r

Actually, instead of posting off on the urug list, I’d still recommend
discussing it here (or in a ruby brigade list/meeting closer to your
home).

On Fri, Mar 6, 2009 at 11:29 PM, usw wickramasinghe

Hi,
Thnx Roger for your kind suggestion. However Pat Eyler suggested me to
post
my idea here when i first contacted him. Neverthless i’ll post my
request
there as well. Thnx again for your feedback…
Regards ,
Udayanga

me on this . Also i’m open to other project ideas as well and again
appreciate your kind reponse. Thnx in advance…

If you just want ideas here are a few [or anybody else can contribute
ideas if you’ve thought of something once that you would put a bounty
our for].

  1. mod_rails: integrate jruby with it somehow. Beef it up with a few of
    its most starred wish list items [1]. My own favorite would be Make
    RailsMaxPoolSize Vhost than Global :slight_smile:

  2. experiment with the MRI garbage collector, see what settings are good
    with what, create getters/setters for it [2]. Could profile it with the
    Ruby Benchmark Suite [3].

  3. Try to optimize rails. I.e. look for bottlenecks and try and clear
    them up with quality LRU cacheing or prepared statements or something :stuck_out_tongue:

  4. ruby 1.9 quality debugger. ruby-debug is currently only 1.8.x compat,
    I believe. I’m actually not sure if this is totally possible, but some
    parts of it are probably 1.9 compatible.

  5. “pie in the sky” psyco for ruby implement an LLVM JIT for Ruby.

Speaking of which I wish that the Google Summer of Code people somehow
were more transparent–like they had to post a result of their work to
at least ruby talk or something.

Cheers!
-=r

[1] Google Code Archive - Long-term storage for Google Code Project Hosting.
[2] http://redmine.ruby-lang.org/issues/show/1047
[3] http://groups.google.com/group/ruby-benchmark-suite?pli=1

Roger P. wrote:

me on this . Also i’m open to other project ideas as well and again
appreciate your kind reponse. Thnx in advance…

If you just want ideas here are a few [or anybody else can contribute
ideas if you’ve thought of something once that you would put a bounty
our for].

  1. mod_rails: integrate jruby with it somehow. Beef it up with a few of
    its most starred wish list items [1]. My own favorite would be Make
    RailsMaxPoolSize Vhost than Global :slight_smile:

This actually could be fun and easier than the Passenger to wire up for
someone who knows Apache well. People have been running fast proxies to
JVM-based servers for a decade using plugins like mod_jk and friends. It
wouldn’t take much effort to wire in the process management + something
like GlassFish gem and have it as drop-in-easy as Passenger is now.

  1. ruby 1.9 quality debugger. ruby-debug is currently only 1.8.x compat,
    I believe. I’m actually not sure if this is totally possible, but some
    parts of it are probably 1.9 compatible.

I’d like to see someone work on a debugger that takes advantage of JRuby
compiling to JVM bytecode. The basic pieces are all there; all that’s
really needed is a nice UI:

~/projects/jruby âž” cat test.rb
puts 1
puts 2
puts 3

~/projects/jruby âž” jruby --jdb test.rb
Initializing jdb …

stop at test:1
Deferring breakpoint test:1.
It will be set after the class is loaded.
run
run org.jruby.Main test.rb
Set uncaught java.lang.Throwable
Set deferred uncaught java.lang.Throwable

VM Started: Set deferred breakpoint test:1

Breakpoint hit: “thread=main”, test.file(), line=1 bci=12
1 puts 1

main[1] list
1 => puts 1
2 puts 2
3 puts 3
main[1]

  • Charlie

Roger P. wrote:

  1. “pie in the sky” psyco for ruby implement an LLVM JIT for Ruby.

I wonder if you could do this for Rubinius, too.

Actually right now the closest thing to a fully functional JIT for Ruby
is JRuby. We run almost all Ruby code and tests out there, and after
compiling to Java bytecode the JVM takes that and turns it into native
code.

So the lesson is: if you want a native-jitting Ruby impl to run your
apps…use JRuby :slight_smile:

  • Charlie
  1. “pie in the sky” psyco for ruby implement an LLVM JIT for Ruby.

I wonder if you could do this for Rubinius, too.
-=r

Hi,

Firstly i would like to thank Pat, Roger and Oliver for your kind
thoughts
on Gsoc Ideas.
Regarding the ideas,

i like the ideas of integrating jruby with mod_rails , and ruby 1.9
debugger, but i need to go into more detail of the problem as i do not
have
clear cut understanding of these scenarios(especially about Phusion
Passenger). If anyone has a JIRA link or blog post or any other link pls
be
kind enough to post here…Obviously i’ll have to do some research as
well as
i’m still pretty much new to ruby and it’s frameworks…

Also i would like to know any suggestions/views on ruby annotation
framework that i have previously mentioned.
I don’t know how exactly Gsoc ideas are being put up for Ruby… whether
do
we have to contact Ruby Central or they’ll be selecting prospective
ideas
given by ruby community. I would really appreciate if anyone could
enlighten
me on the above . Thnx again for the thoughts you have rendered.

Best Regards
-Udayanga

project ideas to be heard from Ruby people…One of the things i’m really
interested in doing is to develop a Ruby Annotation framework
(DSL/parser)for a given spec as ruby currently doesn’t support this
behavior

Another few ideas would be:

porting java’s testng to ruby [or would that be useful–I suppose
test/unit and minitest are pretty good] :slight_smile:

porting python’s process module to ruby. Perhaps using the python apr
to get cross platform goodness.

Thoughts?
-=r

i like the ideas of integrating jruby with mod_rails , and ruby 1.9
debugger, but i need to go into more detail of the problem as i do not
have
clear cut understanding of these scenarios(especially about Phusion
Passenger). If anyone has a JIRA link or blog post or any other link pls
be
kind enough to post here…Obviously i’ll have to do some research as
well as
i’m still pretty much new to ruby and it’s frameworks…

here’s a link to some mod_rails stuffs:
http://code.google.com/p/phusion-passenger/issues/list

re: 1.9 debugger not sure. Looks like it might be hard.

what do you mean by annotation framework?

Cheers!
-=r

Also i would like to know any suggestions/views on ruby annotation
framework that i have previously mentioned.
I don’t know how exactly Gsoc ideas are being put up for Ruby… whether
do
we have to contact Ruby Central or they’ll be selecting prospective
ideas
given by ruby community. I would really appreciate if anyone could
enlighten
me on the above . Thnx again for the thoughts you have rendered.

Best Regards
-Udayanga

On Tue, Mar 10, 2009 at 9:16 PM, usw wickramasinghe
[email protected] wrote:

Hi,

I don’t know how exactly Gsoc ideas are being put up for Ruby… whether do
we have to contact Ruby Central or they’ll be selecting prospective ideas
given by ruby community. I would really appreciate if anyone could enlighten
me on the above . Thnx again for the thoughts you have rendered.

We’ll be posting a page of ideas, but encourage students to consider
their own ideas as well. Student proposals will be accepted through
the GSoC web site:
Home | Google Summer of Code
You might want to look at the FAQ:
http://socghop.appspot.com/document/show/program/google/gsoc2009/faqs

Best Regards
-Udayanga


thanks,
-pate

Don’t judge those who choose to sin differently than you do

Thnx Roger further enlightening us with your prospective ideas for Ruby
.And Thank you Pat for the information you have mentioned , i’ll
constantly
poll for new developments with the links you have provided .

And Roger about the annotation framework …I find it really useful for
web
services and stuff (like testng tools,etc) the presence of a proper
annotaion framework ,as it will provide ways and means to declare some
meta-data needed for these programmes (such as schema definitions ,test
criterias,etc) , in the programme itself rather than going for external
descriptors and stuff . In java we get java.lang.annotation framework
(and
various other adhoc annotation schemas ) where use of @ is extensively
used
as annotation syntax (ie:- @depricated ==> for javadoc stuff ,
@interface
=>custom type declerations, etc) .

In java you also get seperate annotation processing tools[Like Annogen]
,
varoious annotation API’s, etc to support these annotaion specifications
.
However in Ruby i didn come accross such a framework for
specifying/processing annoations and its purely my assumption that ruby
currently doesn’t support such behavior (if i’m wrong pls some one
enlighten
me on this ) . I’m no annotaion frameworks geek , bt would really like
suggest this as a Gsoc proposal for Ruby and willng to undertake such a
project with a mentor who knws this stuff well…
And thnx again for ur kind feedback…

-Udayanga

However in Ruby i didn come accross such a framework for
specifying/processing annoations and its purely my assumption that ruby
currently doesn’t support such behavior (if i’m wrong pls some one
enlighten
me on this ) . I’m no annotaion frameworks geek , bt would really like
suggest this as a Gsoc proposal for Ruby and willng to undertake such a
project with a mentor who knws this stuff well…
And thnx again for ur kind feedback…

'Right currently the only “blessed” annotation is that if you have
comments justified with your first line of a method those are used for
rdoc generation :slight_smile:

Typically to add annotations I believe you roll your own with
method_added [1 is an example].

Good luck!
-=r

[1] http://www.codecommit.com/blog/ruby/adding-type-checking-to-ruby

re: type specifying which this link contains

Now that I think about it, I wonder if there’s a way to do some type of
static ruby type hinting, a la duby, something like
module M
static_section do
typeSig a: Numeric, b: String
def go(a, b)
# this can be compiled down to C or have byte code emitted for it
or what not.
end
end
end

Then those spiffy ruby2c or JIT compilers would be helped along their
way, I guess. Just wishing for a ruby psyco still :slight_smile: