Re: [groovy-user] opening pipe to external program

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
19 messages Options
12
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] opening pipe to external program

Alan-19
* Guillaume Laforge <[hidden email]> [2005-10-21 05:13]:
> On 21/10/05, Hein Meling <[hidden email]> wrote:
> > These are nice; thanks.  These are really worth a page on the wiki.
>
> Feel free to add it ;-))))

    Still inerested?

    It might be worth while to write a Groovy as Script article.

--
Alan Gutierrez - [hidden email] - http://engrm.com/blogometer/
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] opening pipe to external program

Alan-19
* Alan <[hidden email]> [2005-10-29 22:54]:
> * Guillaume Laforge <[hidden email]> [2005-10-21 05:13]:
> > On 21/10/05, Hein Meling <[hidden email]> wrote:
> > > These are nice; thanks.  These are really worth a page on the wiki.
> >
> > Feel free to add it ;-))))
>
>     Still inerested?
>
>     It might be worth while to write a Groovy as Script article.
   
    Er, Groovy as system scripting language.

--
Alan Gutierrez - [hidden email] - http://engrm.com/blogometer/
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

[groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Russel Winder
On Sat, 2005-10-29 at 22:55 -0400, Alan wrote:

> >     It might be worth while to write a Groovy as Script article.
>    
>     Er, Groovy as system scripting language.

I keep trying to write a short article on using Groovy for publishing in
the ACCU journal.  From my limited perspective (i.e. I do not work on
large Java-based (server-side) systems) I keep coming across the problem
that Python and Ruby start faster (humanly obviously so) and have better
library support for things like GTK+ and unbuffered I/O.  Groovy wins
hands down in making Ant easier to work with but for general command
line scripting I keep being drawn back to Python and Ruby.

I suspect though that Groovy and Java have far better support for XML
processing compared to Python and Ruby.

Overall, it seems that Ruby has found it's niche replacing PHP, Groovy
has found a niche for scripting in large Java systems and Python has
found it's niche replacing Perl and being the general systems scripting
language.

Unless I am missing something.

Footnote:  I have been told that JDK now always has the client JVM and
the server JVM and that the server JVM should always be used for server
and client work, i.e. always.  Yet I cannot imagine Sun keeping two JVMs
going if one is superior in all cases.
--
Russel.
====================================================
Dr Russel Winder                +44 20 7585 2200
41 Buckmaster Road              +44 7770 465 077
London SW11 1EN, UK             [hidden email]

signature.asc (196 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Alan-19
* Russel Winder <[hidden email]> [2005-10-30 03:36]:
> On Sat, 2005-10-29 at 22:55 -0400, Alan wrote:
>
> > >     It might be worth while to write a Groovy as Script article.
> >    
> >     Er, Groovy as system scripting language.
>
> I suspect though that Groovy and Java have far better support for XML
> processing compared to Python and Ruby.

    One reason I don't want to walk away from Java, XML seems to
    start there. XML requires many small objects, and Java is pretty
    good at that, and provides garbage collection, and it's hard
    to implement an XML DOM without that.

> Overall, it seems that Ruby has found it's niche replacing PHP,
> Groovy has found a niche for scripting in large Java systems and
> Python has found it's niche replacing Perl and being the general
> systems scripting language.

> Unless I am missing something.

> Footnote:  I have been told that JDK now always has the client JVM
> and the server JVM and that the server JVM should always be used
> for server and client work, i.e. always.  Yet I cannot imagine Sun
> keeping two JVMs going if one is superior in all cases.

    This is news. I'm on OS X, and I think stuck with whatever
    Apple gives me.

--
Alan Gutierrez - [hidden email] - http://engrm.com/blogometer/
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Simon Chappell
In reply to this post by Russel Winder
Russel,

On 10/30/05, Russel Winder <[hidden email]> wrote:
*snip*
> large Java-based (server-side) systems) I keep coming across the problem
> that Python and Ruby start faster (humanly obviously so) and have better
> library support for things like GTK+ and unbuffered I/O.  Groovy wins
> hands down in making Ant easier to work with but for general command
> line scripting I keep being drawn back to Python and Ruby.
*snip*

I also notice the JVM start time issue when working with Groovy; it's
a very real factor to consider when scripting. It seems to me that
there are two things that could be done, one by Sun and the other by
the community:

1. Create a faster starting JVM. Now, I've never looked at the JVM
code, but perhaps it would be possible for Sun to create a command
line switch that would start the JVM with as much as possible
(services or other internal JVM processes) either left-out or
postponed until after the JVM has been up for a while (5, maybe 10,
seconds?) and then it could be started in the background on an "as
needed" basis.

2. Start looking at Java based shells. Now that Groovy has proven that
a great scripting language can be written using Java, perhaps we need
a great shell written using Java to compliment our scripting language
(GASH - the Groovy Again SHell?). A Java-based shell would enable the
JVM startup time hit to be taken once and then individual Groovy
scripts would be able to run with almost immediate startup times.

Just a thought. Am I onto something or just loopier than usual? :-)

Simon

--
www.simonpeter.com
uab.blogspot.com
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Martin C. Martin
AFAIK, most of the start up time isn't the JVM startup, it's other
stuff.  I think this is on the to-do list, but not for 1.0.  We'd rather
have things like closures, multi-D arrays and the synchronized keyword
working first.

- Martin

Simon Chappell wrote:

> Russel,
>
> On 10/30/05, Russel Winder <[hidden email]> wrote:
> *snip*
>
>>large Java-based (server-side) systems) I keep coming across the problem
>>that Python and Ruby start faster (humanly obviously so) and have better
>>library support for things like GTK+ and unbuffered I/O.  Groovy wins
>>hands down in making Ant easier to work with but for general command
>>line scripting I keep being drawn back to Python and Ruby.
>
> *snip*
>
> I also notice the JVM start time issue when working with Groovy; it's
> a very real factor to consider when scripting. It seems to me that
> there are two things that could be done, one by Sun and the other by
> the community:
>
> 1. Create a faster starting JVM. Now, I've never looked at the JVM
> code, but perhaps it would be possible for Sun to create a command
> line switch that would start the JVM with as much as possible
> (services or other internal JVM processes) either left-out or
> postponed until after the JVM has been up for a while (5, maybe 10,
> seconds?) and then it could be started in the background on an "as
> needed" basis.
>
> 2. Start looking at Java based shells. Now that Groovy has proven that
> a great scripting language can be written using Java, perhaps we need
> a great shell written using Java to compliment our scripting language
> (GASH - the Groovy Again SHell?). A Java-based shell would enable the
> JVM startup time hit to be taken once and then individual Groovy
> scripts would be able to run with almost immediate startup times.
>
> Just a thought. Am I onto something or just loopier than usual? :-)
>
> Simon
>
> --
> www.simonpeter.com
> uab.blogspot.com
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Jochen Theodorou
In reply to this post by Simon Chappell
Simon Chappell schrieb:
[...]
> I also notice the JVM start time issue when working with Groovy; it's
> a very real factor to consider when scripting.

besides the things groovy does, the java startup times are so bad due to
the need to load hundreds of classes where each has to be verified first.
[...]
> 2. Start looking at Java based shells. Now that Groovy has proven that
> a great scripting language can be written using Java, perhaps we need
> a great shell written using Java to compliment our scripting language
> (GASH - the Groovy Again SHell?). A Java-based shell would enable the
> JVM startup time hit to be taken once and then individual Groovy
> scripts would be able to run with almost immediate startup times.

you mean domething like jsh, http://gerard.collin3.free.fr/ ?

bye blackdrag
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Simon Chappell
In reply to this post by Martin C. Martin
Are you sure? Does anyone have timings to confirm/deny this?

On 10/30/05, Martin C. Martin <[hidden email]> wrote:

> AFAIK, most of the start up time isn't the JVM startup, it's other
> stuff.  I think this is on the to-do list, but not for 1.0.  We'd rather
> have things like closures, multi-D arrays and the synchronized keyword
> working first.
>
> - Martin
>
> Simon Chappell wrote:
> > Russel,
> >
> > On 10/30/05, Russel Winder <[hidden email]> wrote:
> > *snip*
> >
> >>large Java-based (server-side) systems) I keep coming across the problem
> >>that Python and Ruby start faster (humanly obviously so) and have better
> >>library support for things like GTK+ and unbuffered I/O.  Groovy wins
> >>hands down in making Ant easier to work with but for general command
> >>line scripting I keep being drawn back to Python and Ruby.
> >
> > *snip*
> >
> > I also notice the JVM start time issue when working with Groovy; it's
> > a very real factor to consider when scripting. It seems to me that
> > there are two things that could be done, one by Sun and the other by
> > the community:
> >
> > 1. Create a faster starting JVM. Now, I've never looked at the JVM
> > code, but perhaps it would be possible for Sun to create a command
> > line switch that would start the JVM with as much as possible
> > (services or other internal JVM processes) either left-out or
> > postponed until after the JVM has been up for a while (5, maybe 10,
> > seconds?) and then it could be started in the background on an "as
> > needed" basis.
> >
> > 2. Start looking at Java based shells. Now that Groovy has proven that
> > a great scripting language can be written using Java, perhaps we need
> > a great shell written using Java to compliment our scripting language
> > (GASH - the Groovy Again SHell?). A Java-based shell would enable the
> > JVM startup time hit to be taken once and then individual Groovy
> > scripts would be able to run with almost immediate startup times.
> >
> > Just a thought. Am I onto something or just loopier than usual? :-)
> >
> > Simon
> >
> > --
> > www.simonpeter.com
> > uab.blogspot.com
>


--
www.simonpeter.com
uab.blogspot.com
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Simon Chappell
In reply to this post by Jochen Theodorou
On 10/30/05, Jochen Theodorou <[hidden email]> wrote:
> Simon Chappell schrieb:
> [...]
> > I also notice the JVM start time issue when working with Groovy; it's
> > a very real factor to consider when scripting.
>
> besides the things groovy does, the java startup times are so bad due to
> the need to load hundreds of classes where each has to be verified first.

Right. But could this list of classes be reduced and any background
in-JVM services be reduced at all? (Only when a requested using a
scripting commandline switch.)

> > 2. Start looking at Java based shells. Now that Groovy has proven that
> > a great scripting language can be written using Java, perhaps we need
> > a great shell written using Java to compliment our scripting language
> > (GASH - the Groovy Again SHell?). A Java-based shell would enable the
> > JVM startup time hit to be taken once and then individual Groovy
> > scripts would be able to run with almost immediate startup times.
>
> you mean domething like jsh, http://gerard.collin3.free.fr/ ?

Yes. As to which one would actually be used, that's another decision.
I noticed that the last updated version of the shell you referenced
was in 2002. I'd prefer to work with something more obviously under
active development and maintenance.

Simon

--
www.simonpeter.com
uab.blogspot.com
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [groovy-user] Groovy articles [ was Re: [groovy-user] opening pipe to external program ]

Alan-19
In reply to this post by Simon Chappell
* Simon Chappell <[hidden email]> [2005-10-30 18:31]:

> Russel,
>
> On 10/30/05, Russel Winder <[hidden email]> wrote:
> *snip*
> > large Java-based (server-side) systems) I keep coming across the problem
> > that Python and Ruby start faster (humanly obviously so) and have better
> > library support for things like GTK+ and unbuffered I/O.  Groovy wins
> > hands down in making Ant easier to work with but for general command
> > line scripting I keep being drawn back to Python and Ruby.
> *snip*
>
> 2. Start looking at Java based shells. Now that Groovy has proven that
> a great scripting language can be written using Java, perhaps we need
> a great shell written using Java to compliment our scripting language
> (GASH - the Groovy Again SHell?). A Java-based shell would enable the
> JVM startup time hit to be taken once and then individual Groovy
> scripts would be able to run with almost immediate startup times.

    In addition to what other said, I'd like to propose to, a Java
    based shell that is XML aware. You are able to navigate an XML
    document as if it were a filesystem.

    XML has found a home on the JVM. Groovy can make it easier to
    deal with XML than writing Java or XSLT code.

--
Alan Gutierrez - [hidden email] - http://engrm.com/blogometer/
12
Loading...